Critical Complexity (Categories) [Digital original ed.] 1501510797, 9781501510793

Citation preview

Paul Cilliers Critical Complexity

Categories

Edited by Roberto Poli (Trento) Advisory Board John Bell (London, CA) Mark Bickhard (Lehigh) Heinrich Herre (Leipzig) David Weissman (New York)

Volume 6

Paul Cilliers

Critical Complexity Collected Essays Edited by Rika Preiser

ISBN 978-1-5015-1079-3 e-ISBN (PDF) 978-1-5015-0259-0 e-ISBN (EPUB) 978-1-5015-0261-3 ISSN 2198-1868 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.dnb.de. © 2016 Walter de Gruyter GmbH, Berlin/Boston Printing: CPI books GmbH, Leck ♾ Printed on acid-free paper Printed in Germany www.degruyter.com

In loving memory of Paul Cilliers beloved husband, father, son, friend whose passionate engagement with life and meaning has touched our lives forever

Photo: Paul Cilliers at STIAS, Stellenbosch, May 2011 

© Jana du Plessis

Someone with a heart even fuller than his head And a head even fuller than the world would allow him And things would come out of his heart and his head And how lucky for us that they did And things would come out of his heart and his head And yet somehow they were also inside him. And there he would turn them every which way To make sure that the right side was facing up And sometimes he wouldn’t know which was the right side And so he would have to keep turning And because things would come out, we knew things were turning But he didn’t mind that we knew Because turning the things in his head and his heart was his favourite thing to do. And because things came out of his head and his heart The things in our heads and our hearts would turn too. And how lucky for us that they did Ilana Cilliers, 4 August 2011

Table of Contents Acknowledgements | XIII  Jan-Hendrik Hofmeyr   Foreword | XV  Rika Preiser & Minka Woermann   Introduction | 1 

Part 1: Single-Authored Papers Theme 1: Characterising Complexity    Paul Cilliers    The brain, the mental apparatus and the text A post-structural neuropsychology | 23  Paul Cilliers    Rules and relations Some connectionist implications for cognitive science and language | 39  Paul Cilliers    Rules and complex systems | 55 Paul Cilliers    What can we learn from a theory of complexity? | 67  Paul Cilliers    Knowledge, complexity and understanding | 77  Paul Cilliers    Boundaries, hierarchies and networks in complex systems | 85  Paul Cilliers    Why we cannot know complex things completely | 97 

X | Table of Contents

Paul Cilliers    Knowledge, limits and boundaries | 105 

Part 1: Single-Authored Papers Theme 2: Complexity and Philosophy   Paul Cilliers   Postmodern knowledge and complexity (or why anything does not go) | 117 Paul Cilliers   Complexity, deconstruction and relativism | 139  Paul Cilliers   On Derrida and apartheid | 153 Paul Cilliers, Willie van der Merwe & Johan Degenaar   Justice, law and philosophy An interview with Jacques Derrida | 171 Paul Cilliers   Complexity, ethics and justice | 181

Part 1: Single-Authored Papers Theme 3: Implications of Complexity Thinking  Paul Cilliers   Difference, identity and complexity | 193 Paul Cilliers   Complexity and philosophy On the importance of a certain slowness | 211

Table of Contents | XI

Part 2: Posthumous after 2011 Theme 1: Critical Complexity  Rika Preiser, Paul Cilliers & Oliver Human   Deconstruction and complexity A critical economy | 225 Oliver Human & Paul Cilliers   Towards an economy of complexity Derrida, Morin and Bataille | 245 Minka Woermann & Paul Cilliers   The ethics of complexity and the complexity of ethics | 265

Author Index | 285 

Acknowledgements The editor would like to express sincere gratitude to the following people for their encouragement and support: to Roberto Poli for your heartfelt interest in this project and for opening the door so that we could publish this book; to Maik Bierwirth and Olena Gainulina at De Gruyter for your patience, excellent advice and support every step of the way; to all the various journal editors for their overwhelmingly positive responses and for permission to reproduce the articles that constitute this book; to Johannes Richter at SUN MeDIA for technical assistance in converting all the PDFs to text and for the meticulous way in which you formatted the text; to Reinette (Oonsie) Biggs for your generous support and encouragement; to Minka Woermann and Jannie Hofmeyr for your significant contributions; to Sandra Cilliers and family, for your vision to share Paul’s work with those who were close to him, but also so that future readers might be inspired. Last but not least, to Paul for touching our lives and for sharing your extraordinary humanity with us.

Jan-Hendrik Hofmeyr

Foreword

Jan-Hendrik Hofmeyr: Wissenschaftskolleg zu Berlin, Germany, February 2015; on sabbatical leave from the Department of Biochemistry and the Centre for Studies in Complexity of the University of Stellenbosch, South Africa.

In 1998 a new voice exploded onto the stage of complexity studies with the publication of Complexity and Postmodernism: Understanding Complex Systems. For Paul Cilliers this monograph, based on his PhD in Philosophy, launched his international career, and in the ensuing 13 years until his untimely death at the age of 54, he became one of his generation’s most influential complexity thinkers. As close friend and travelling companion in the world of ideas over three decades Paul had an incalculable influence on my own thinking about our world, how to understand it, and how to live a life of quality in it. In my obituary to Paul (South African Journal of Science (2012) 108: 14–15) I told the tale of how he made the transition from electronic engineer to philosopher and complexity thinker. How was it, though, that a philosopher and a biochemist found common ground? In my world of cells and biomolecules I had worried from the start about the extremely reductionistic way in which biochemistry approached the study of living organisms, and I started using mathematical modelling and computer simulation to develop an understanding of how the myriad of components of the living cell work together in a harmonious and integrated way. Paul, with his electronic engineering background and his research at the Institute for Maritime Technology in Simonstown, South Africa, became interested in the emergent properties of neural nets, and we spent many evenings trying to figure out how to bring these two approaches to understanding systemic behaviour together. We struggled, however, not yet having escaped the constraints and limitations of what has come to be known as the Newtonian paradigm. We both needed a jolt to propel us over the hill into the basin of attraction of systems and complexity thinking. For Paul it was the discovery of the work of Jacques Derrida and other poststructuralist thinkers; for me it was the discovery of the work of the theoretical biologist Robert Rosen. It was through the development of their ideas that Paul and I realised that we had within our grasp the stirrings of a common language, the beginnings of a new way of looking at the world, a view that is now called the relational worldview. The active development of the sciences of systems and complexity in the domains of both the natural and human sciences has now shown that there is a generally perceived need for

XVI 

 Jan-Hendrik Hofmeyr

such a relational view of the world and that there is an active striving towards it, despite our seeming predilection for atomistic thinking. Paul’s contribution to the development of a philosophy that could underlie the relational worldview is of fundamental importance. The papers collected in this book provide a deep insight into his view of what it is to be a human being in a complex and inherently unpredictable world, and will ensure that this work remains accessible to a new generation of complexity thinkers. These works chronicle his efforts to understand how it happened that humankind, especially the Western variant, got bogged down in a worldview that managed, in his words, to “reduce humanity to an instrumentalised, commodified, superficial thing, isolated in a very real sense from the rest of life on earth?” And all of this despite humanity’s wonderful accomplishments – science, art, music, literature. In their introduction, the editors of this collection, Rika Preiser and Minka Woermann, have done a sterling job of outlining how Paul engaged with this question, while at the same time situating Paul’s work in the broader context of complexity studies in the last two decades. Paul’s legacy continues through the work of the Centre for Studies in Complexity, which has in the years since his death attracted new young stars. As I  write our Centre is forming a cornerstone of a new initiative: the Centre for Complex Systems in Transition, which consolidates existing Stellenbosch expertise in complexity theory, sustainability studies, social-ecological systems and resilience, modelling of complex systems, and transdisciplinary research methodology. Our current and future students will continue to be fed a healthy helping of Paul’s insights and wisdom and this book will be required reading for many years to come. In retrospect, of all that Paul has meant to me, one thing towers above all else: he taught me to try and make each act in my daily life a quality act. This is a tough ideal to live up to and I probably fail more often than not, but it has enriched my life immeasurably. If you want to emulate this ideal, these pages will offer you ample assistance.

Reference Hofmeyr, J-H. 2012. Friedrich Paul Cilliers: philosopher (1956–2011). South African Journal of Science, 108(3–4): 14–15.

Rika Preiser & Minka Woermann

Introduction 1 Context

Any contemporary manuscript on the notion of ‘complexity’ will be incomplete without mentioning the work of Paul Cilliers. Together with a number of other publications that came out just before the turn of the century (Byrne 1998, Holland 1999, Thrift 1999), Cilliers’ book Complexity and Postmodernism: Understanding Complex Systems (Routledge 1998a) became a cornerstone publication in the canon of literature on complexity. Cilliers was one of the first authors to approach the understanding of complex systems from a philosophical perspective by approaching it from a post-structural position. He believed that taking cognisance of the insights gleaned from the field of post-structural philosophy would change the way that we practice science, as is evident by the following citation: Adopting a post-structural perspective on science will certainly be in conflict with much of what is accepted as canonical theory of science, but may have less radical effects on the practice of science than one expects. Unless one would want to call the opening up of new spaces for creative thought something radical (page 52).

Cilliers’ research certainly opened up a number of new spaces for creative thought, specifically in terms of how we characterise complex systems, the implications for studying such systems, and the critical and normative implications that a study of complex systems poses for the way in which we model these systems. Ultimately, the study of complex systems challenges scientists to re-think the role of science and how it relates to larger societal questions such as how we should live and what it means to be human in a complex world. Cilliers framed his understanding of complexity not in terms of a fully-fledged ‘theory of complexity’, but by highlighting what consequences a complex systems approach has for re-thinking our research and decision-making practices. Adopting an ‘attitude of complexity’ could amount in a paradigm change, as demonstrated by the ethical and philosophical implications that such a stance holds. These implications are unpacked in his work and feed into his larger project of questioning and re-examining what it means ‘to be human’ in the face of complex

2 

 Rika Preiser & Minka Woermann

phenomena. Thus, Cilliers took the implications of adopting a complexity attitude to a much deeper level of understanding than was previously the case, and – in so doing – he generated a number of important insights that can be generalised and applied to a wide range of disciplines and practices. Although Cilliers’ book remains an important reference in any study of complexity, his articles explore the general ideas put forward in his book in more depth, and with greater applicability. In some cases, these articles also contain the further conceptual development of certain core ideas that were introduced in his book. Before Cilliers’ unexpected passing in July 2011, he had planned to put together a second book that would focus on integrating the range of themes that he touched on in his articles to culminate in an interpretation of complexity that he would call ‘Critical Complexity’. Sadly, he never got the chance to do so. On the one hand, this collection of essays is partly an effort to realise his dream, but, on the other hand, it also represents an attempt to consolidate and integrate his work into a coherent framework of thought. In order to facilitate this latter goal, the selection of articles in this collection is not presented in chronological order, but clustered around a number of themes that distinctly mark the theoretical and conceptual positions that influenced Cilliers’ understanding and approach to complexity, and that serve to explicate and systematise his own contribution. This book is further divided into two parts: Part  I includes a selection of Cilliers’ single-authored articles, whereas Part II comprises co-authored articles that were published posthumously. The three themes included under Part I are ‘Characterising Complexity’, ‘Complexity and Philosophy’, and ‘Implications of Complexity Thinking’; whereas Part II pertains to the theme ‘Critical Complexity’. Cilliers’ unique contribution to the field of complexity studies lies in his sophisticated understanding of how post-structuralism enriches our understanding of complexity. Likewise, his contribution to the field of philosophy is that he developed a rigorous interpretation of the post-structural position by means of a nuanced reading of complex systems. This simultaneous mutual cross-fertilisation of both fields forms a bridge between the natural sciences and the humanities, and generates a creative and alternative space for new thinking to emerge. From this space, new possibilities for collaboration and transdisciplinary research arise, as is evident from the wide range of collaborative projects that form part of Cilliers’ list of publications.

Introduction 

 3

2 Part 1: Single-authored articles 2.1

Introduction to Cilliers’ single-authored articles

Before taking a closer look at the articles that fall under the three themes in Part I, it is useful to briefly introduce Cilliers’ central contribution, as well as the positioning of his work by means of some opening remarks. Cilliers initially developed his view on complexity by teasing out the similarities between how meaning arises in distributed networks (such as the brain) and how meaning emerges through systems of differentiation in language. In so doing, he developed an understanding of complexity that was focused on the dynamic interactions or relations between the components of a system, rather than on the components themselves. What makes Cilliers’ work on complexity unique, is that he not only translated the highly technical and often mathematical language in which complex systems are described in the natural sciences into a more accessible vocabulary for scholars in the humanities and social science, but he also simultaneously re-articulated the notion of complexity through the lens of structural and post-structural philosophy in general (drawing on the work of Freud and Saussure), and through a very deep and refined understanding of the work of Jacques Derrida in particular. Whilst both Cilliers’ unique interpretation of complexity and his post-structural engagement with the philosophy of science allows for innovative thought and collaboration, working in the margins of both the fields of philosophy and complexity studies poses some challenges when it comes to finding journals that would be willing to publish work that does not conform to the conventions of academic disciplines. With this in mind, Cilliers found very supportive platforms for publishing his new ideas in the South African Journal of Philosophy and in Emergence: Complexity and Organization. He published extensively in both of these journals in a time when writing about complexity was not as trendy as it has lately become, and it is largely due to their support of Cilliers’ work, that we are now in a position to publish this collection.

2.1.1

Theme 1: Characterising Complexity

The first theme in Part I, ‘Characterising Complexity’, consists of eight articles that form the core of Cilliers’ understanding of the structural and functional characteristics of complex phenomena.

4 

 Rika Preiser & Minka Woermann

In the first three articles, Cilliers establishes the foundation on which he develops his understanding of complexity. Based on his work as an engineer on pattern recognition, neural networks, and the functioning of distributed networks, Cilliers puts forward an argument in ‘The brain, the mental apparatus and the text’ (page  23) that favours a connectionist approach for explaining how the higher brain functions, such as perception, memory and consciousness, emerge. By combining key ideas from the work of Freud, Saussure, and Derrida with the neurophysical theories of the time, Cilliers manages to develop what he calls a ‘post-structural neuropsychology’ (page  23). Cilliers compares the manner in which neurons process information to generate higher brain functions with the ‘logic of différance’ (page  30) (see also discussion below), and concludes that ‘[c]onsciousness is the différance of perception and memory’ (page 32). This view departs from the traditional theories of the time, in which the brain is compared to a machine, and in which artificial intelligence is understood as being based on formal, rule-based (and therefore necessarily limited) simulations of neural networks. He argues that the formal symbolic approach is an inappropriate model for understanding intelligence, and instead draws on the connectionist paradigm, in order to overcome the shortcomings of the traditional rule-based approach. Connectionist models (or neural networks), which are inspired by biological neural networks, conceive of brain function or cognition in terms of the interactions and relationships between neurons (and clusters of neurons). From the study of neural networks, we learn that complex systems consist of large numbers of simple neurons (elements) that are richly connected. Complex patterns are generated by the network of interrelated components. Based on this understanding of neural networks, Cilliers puts forward the following two important characteristics of a connectionist understanding of brain function (page 36): –– Knowledge is not represented locally in an iconic fashion (as is the case in conventional computers and rule-based systems), but is rather distributed over the whole system. This is because knowledge is a function of the connection strength between units. –– If brain functioning is purely relational, the system cannot be rule-based on a first level, because there are only interactions (traces). Given his description of how neural networks are structured and function, Cilliers argues that the connectionist model poses a more general model of complexity than traditional rule-based models. Moreover, Cilliers argues that the connectionist model of the brain shows a strong correlation with the structural and post-structural models of language. In these models, meaning is constituted relationally, which implies that there is ‘no distinction between levels, no overar-

Introduction 

 5

ching algorithm, but everything [should be understood] in terms of relations […] not between positive entities, but always only relations of relations’ (page 37). In the next two articles ‘Rules and relations’ (page  39) and ‘Rules and complex systems’ (page  55), Cilliers presents a more detailed understanding of the connectionist model. The reason for this being that, at the time, this new model was heavily criticised for undermining the assumptions on which artificial intelligence research was based. In these articles, the links to a post-structural understanding of how meaning arises in language and the implications thereof are explained in more detail. Cilliers was specifically influenced by Ferdinand de Saussure’s structural model of language. In this model, meaning is not the product of the substantive identities of components, but of the differences between components, which give rise to a negative view of identity (i.e. I am X, in virtue of the fact that I am not Y, Z, A, or B). This view of identity reinforces the connectionist insight that the primary unit of analysis in a complex system is the relations between the components, rather than the components themselves. In other words, it is X’s relations to Y, Z, A, and B that give rise to X’s identity. Despite the influence of Saussure’s view of language on the work of Cilliers, he ultimately found the model too rigid to adequately describe complex systems, specifically the dynamic manner in which relationships interact in time. For this reason, he turned to the work of Derrida, whose critical reworking of Saussure’s structural model of language provided the basis for further developing his understanding of complexity. The influence of Derrida on Cilliers’ work will be addressed in the next section. At this juncture, however, it is important to note that, in practical terms, the shift to relationality as the mechanism that generates complexity, translates into the insight that ‘[t]here is no Programmer, no Scientist that can uncover the full Truth and the final significance of each element. There is, and was, always only the relationship of traces’ (page 37). The insight that complexity cannot be located (or isolated) and analysed so as to find the essential and central operating unit from where all information can be accessed and stored, is a very significant characteristic of complexity. For Cilliers this insight has substantial consequences in terms of how we approach or model complexity, how we study it, and what type of knowledge we can generate. This concern is captured in the following quote regarding the nature and status of rules: There is nothing mystical about the workings of a complex system. However, since the nature of the system is the result of countless, local, nonlinear, non-algorithmic, dynamic interaction, it cannot be described completely and accurately in terms of a set of rules (page 62).

6 

 Rika Preiser & Minka Woermann

The remainder of the articles in Part I mainly focus on questions concerning the implications that such a relational understanding of complexity (as a characteristic of systemic interaction) has for how one can model complexity and, subsequently, for knowledge generation and understanding. The next five articles – ‘What can we learn from a theory of complexity’ (page  67), ‘Knowledge, complexity and understanding’ (page  77), ‘Boundaries, hierarchies and networks in complex systems’ (page  85), ‘Why we cannot know complex thinks completely (page 97) and ‘Knowledge, limits and boundaries’ (page 105) – can therefore be read together. These articles form the core of his argument that theories of complexity have important implications for the knowledge claims we make when dealing with complex systems. The golden thread that binds these articles together is the claim that all knowledge of complex systems will be limited due to the incompressible nature of complex systems. Recognising this fact holds implications for the manner in which we think about our knowledge claims. Specifically, it requires of us to exercise modesty in our epistemological practices. This argument is explained in more detail below. Cilliers argues that ‘to fully understand a complex system, we need to understand it in all its complexity’ (page 143). Considering that complex systems are characterised as open systems that interact with their environment in a dynamic and nonlinear manner, modelling complexity in all its complexity would imply that one would have to ‘understand the system’s complete environment before we can understand the system, and, of course, the environment is complex in itself’ (page  143). As a consequence of this ‘incompressibility’ that characterises complex phenomena, Cilliers purports that there can be no perfect representation of the system that is simpler than the system itself. In order to extract meaning, and generate an understanding of the system under study, one has to construct a model. A model can be explained as the way in which we frame the system in order to build knowledge. In building representations of open systems, we necessarily have to reduce complexity. Our modelling strategies by definition cannot therefore capture all the variables of the system under study and its environment. Furthermore, since the effects of these omissions are nonlinear, we cannot predict their magnitude. By acknowledging that knowledge of complex systems can never be complete, one is confronted with the unavoidability of the limitations of human understanding. The study of complexity thus points to the fact that the character of our representations of reality is at most limited and partial. This insight corresponds with Mitchell’s (2007: 7) view when she argues, ‘that there will never be a single account that can do all the work of describing and explaining complex phenomena’. Recognising the nature of complexity therefore facilitates a shift in attitude. Limitations are acknowledged and not concealed. Reductions are made explicit

Introduction 

 7

and confrontation with emergence is not obscured or denied. Cilliers argues that this shift towards a more ‘modest’ attitude does not mean that we have to take a ‘weak’ approach or ‘cringe in false modesty’, but that we can still make clear, impermeable assertions (page 150). On the contrary, ‘[t]he fact that our knowledge is limited is not a disaster, it is a condition for knowledge. Limits enable knowledge’ (page 150). The challenge of being able to know complex systems and the difficulties it poses for knowledge generating practices is one of the distinguishing characteristics that mark the discourse on complexity (Zadeh & Polak 1969, Allen 2001, Georgiou 2007, Wolkenhauer & Ullah 2007). However, the idea that all knowledge of complexity will in principle always only result in partial knowledge (cf. Poli 2013) is, for Cilliers, not only a technical consideration bearing on knowledge generation, but also necessarily poses normative questions when studying complexity, a view that is discussed in more detail below.

2.1.2 Theme 2: Complexity and Philosophy As mentioned in the introduction, Cilliers saw his complexity project as resonating with the ideas of postmodernism in general, and he developed this thesis in his doctoral dissertation completed in 1993, titled ‘Modelling Complexity’. His reworked dissertation appeared in 1998 under the title ‘Complexity and Postmodernism’, and was – as stated in the introduction – published to critical acclaim. The central argument of this work also appears in condensed form in the article titled ‘Postmodern knowledge and complexity (or why anything does not go)’ (page  117). Herein, Cilliers presents his oft-quoted ‘ten characteristics of complex systems’, and applies these to the economic system specifically, and a Lyotardian understanding of social systems in general, in order to demonstrate how postmodern society and its sub-systems function as complex systems. As suggested by the title, this article also presents a defense of postmodernism, in that Cilliers argues against a relativist approach to knowledge, arguing instead that ‘postmodernism […] provides us with a strategy for coping with the complexities we have to deal with when we wish to talk about our world’ (page 117). Cilliers’ strong defense of complex, postmodern positions is most evident in his inaugural address for his appointment as full professor in philosophy in 2004. This address, titled ‘Do modest positions have to be weak?’, and published in 2005 under the title ‘Complexity, deconstruction and relativism’ (page  139), employs a number of Derrida’s arguments in order to demonstrate that complex positions are neither relativistic nor vague, but that such positions demand recognition of the fact that our knowledge claims are necessarily limited, for the reason

8 

 Rika Preiser & Minka Woermann

that – as argued above – we cannot understand complexity in all its complexity. This means that we shall sometimes find ourselves in a performative contradiction. However, he argues that far from landing us in a position of irrationality, conceding to the necessity of ‘a certain performative reflexivity’ (Wood 1990: 132), is a means by which we can ‘demonstrate the difficulties that we are in; also in the way we talk about them’ (page  147). This statement reinforces the view articulated above, namely that normative questions are inevitable when dealing with complexity. Indeed, in this article Cilliers explicates this view stating that ‘[n]ormative issues are … intertwined with our very understanding of complexity’ for the reason that our knowledge claims represent choices, rather than programmable calculations. Ethics ‘are always already part of what we do’ (page 150). Cilliers’ understanding of complexity represents not only a descriptive, but also a normative, position. Indeed, his deep concern for normative issues becomes progressively more evident in the course of his work. Apart from taking inspiration from Derrida’s view of meaning and language, Cilliers was also deeply influenced by Derrida’s understanding of, and engagement with, ethical matters, such as racism and apartheid and the relation between law and justice, and this influence is also very evident in the articles included under this theme. South Africa, with its history of apartheid and its peaceful transition to a democratic country, fascinated Derrida, and his work on forgiveness and memory was, in part, influenced by South Africa’s history of racism and post-apartheid attempts at reconciliation (of which he was critical). Derrida’s earliest engagement with South Africa’s political situation was in the early 1980s, in which time he produced the text ‘Racism’s last word’. Cilliers critically engages with this text, in his article titled ‘On Derrida and Apartheid’ (page  153), in which he interrogates Derrida’s understanding of apartheid as the ultimate form of racism in the world, arguing that, in this description, Derrida betrays his own non-totalising philosophy. The significance of this article – to our minds – is that it represents Cilliers’ most explicit academic engagement with ethical-political issues. A common point of criticism lodged against Cilliers’ work is that, although he purports to be concerned with ethical issues, his view on ethics is so non-prescriptive and substantively empty that it becomes impossible to make serious judgements or take a definitive stance on any matter (Kunneman 2010). Cilliers’ ethical position will be elaborated upon in more detail below, but at this juncture we wish to empathically note that Cilliers was no fence-sitter: neither in his private life nor in his academic life. Although, after the publication of ‘On Derrida and Apartheid’ he never again used the academic platform to explicitly put forward his political beliefs, Cilliers was uncompromising in both his personal and academic convictions. His integrity and his deep concern for others and for the future of the planet shine through in his work. Professor Carl Folke, Science Director of the Stockholm

Introduction 

 9

Resilience Centre, recently summed up this sentiment well, when – after reading some of Cilliers’ work – he remarked that it is obvious that his work stemmed from a deep emotional and moral conviction, which lends a certain gravitas to his philosophy. Cilliers’ unflagging faith in the importance of his project and philosophy in general is also evident in an interview conducted with Derrida during his visit to South Africa in 1998, titled ‘Justice, law and philosophy’ (page  171). In this interview, Cilliers interrogates Derrida on his views regarding the commonplace argument ‘that philosophy is a luxury, that philosophy, and the teaching of philosophy cannot be a priority’ (page 178). We shall leave it to the reader to look up Derrida’s response to this question, but wish to note that the question itself represents an opening for Derrida to counter such views; an opening which Cilliers also sought in his own teaching and research, often reminding his students that – in a world dominated by a technocratic mind-set – there exists a moral and political imperative to engage in philosophy and in the arts, in order to counteract the totalising views that drive forward man’s relentless search for growth and progress at the cost of all else. In the same interview, Cilliers also poses questions to Derrida related to justice, law, and the relation between these terms. Derrida’s understanding of justice and law deeply influenced Cilliers’ own ethical position, and in an article published six years after the interview, and titled ‘Complexity, ethics and justice’ (page  181), Cilliers again returns to these issues, framing them in terms of a complex understanding of the world. For Derrida, the fullness of justice can never be realised in the world, which means that justice remains conceptually impossible (and hence undeconstructable). Law always tries to embody justice, but necessarily fails to so, and is therefore always subject to deconstruction and revision. For Cilliers, the impossibility of justice hinges on complexity. Cilliers takes the argument that we are forced to model (and hence reduce) complex systems in order to understand them and applies it to social systems. When applied to these systems, ‘the violation of something or someone that is not (or cannot be) considered in terms of that description’ (page  187) necessarily bears ethical consequences. In terms of justice, Cilliers describes the inevitability of this ethical conundrum as follows: It is impossible to arrive at a complete and just description of society, not because we lack the intellectual resources, but because the demands made on such a description are contradictory. To provide justice for someone will mean that somebody else is treated unjustly. One cannot begin to think about the problem of justice is one does not accept its impossibility (page 187).

10 

 Rika Preiser & Minka Woermann

The logic that informs the impossibility of justice also explains why Cilliers balked at a prescriptive or substantive ethics: any ethical system constitutes a model that is also only able to account for limited interests and concerns, and thus does violence to those whose interests are not accounted for in terms of that system. The upshot of this position, as argued for in the conclusion of ‘Complexity, ethics and justice’, is that we cannot escape responsibility for the choices and the decisions that we make.

2.1.3 Theme 3: Implications of Complexity Thinking The fact that one cannot escape ethics when dealing with complexity also means that a serious engagement with complexity also holds important implications for how we understand the questions of individual and institutional identities, as well as how we orientate ourselves in the world. In his later works, Cilliers turned to these implications and two significant articles in this regard are ‘Difference, identity and complexity’ (page  193) and ‘On the importance of a certain slowness’ (page  211). These two articles are key works to the extent that they demonstrate the creative manner in which Cilliers extended his insights beyond the scope of complex systems as such, to include issues of identity and practice, amongst others. In ‘Difference, identity and complexity’, Cilliers argues that difference is a precondition for a system’s existence and its identity. Difference, however, should not only be thought of in terms of noticeable differences (the Saussiarian understanding of meaning), but also in terms of how the interconnectivities within complex systems give rise to minute traces of difference, which pervade every component, and which means that the identity of components are characterised by the deferral of a fixed meaning, and the constant transformation of their identities as such (a process denoted by the Derridean term différance). However, Cilliers further argues that differences are not enough to ensure identity: differences must be framed in a specific way, in order to be recognised as differences. As such, boundaries (which give rise to constrained meaning) are a necessary condition for identity, or in the words of Cilliers (page 203), ‘[t]here is a certain economy involved in the process whereby differences generate meaning in a complex system.’ Boundaries also imply that complex systems have a certain knowable identity, which – although malleable – can be repeated over time, meaning that complex systems also have a recognisable and repeatable identity. The interplay between difference and sameness is ultimately what gives rise to identity, since, as Cilliers explains, ‘[t]he element of identity inaugurates the play of difference on the one hand, while on the other, it is the result of that

Introduction 

 11

very process.’ Identity is therefore premised on both ‘constrained difference and repeatable identity’ (page  205). As always, Cilliers also identifies the ethical implications that this view of complex identities holds: if differences are a precondition for identity, then it stands to reason that we should celebrate and stimulate systemic diversity, because ‘[t]he more diversity there is involved in the construction of the identity, the richer it will be’ (page 206). Recognising this point when thinking about organisations and institutions is critical, since as Cilliers argues, ‘[t]he way in which we conceive of differences and structures will determine the nature of our institutions, and thus of the world we live in’ (page 210). As is clear from the above, Cilliers believed that in order to live in a good world, we need to consciously reflect on our systems and practices. This reflection also has implications for the temporal aspects of systems, which, Cilliers argued, should be considered more thoughtfully. In contemporary applications of complex systems thinking, a lot of attention is given to the notion that complex systems are adaptive, have the ability to change, and operate in a constant state of flux. This common understanding of complexity could give the impression that change happens continually and quickly, and that complexity should be equated with rapid change and flexibility. However – as previously argued – the identity of the system is recognisable over time. This implies that a system is not defined in terms of a random momentary state, but rather in terms of its structural components that remain relatively stable. Change is thus not something that always happens rapidly, and in many instances it is misleading to think that rapid change is an indicator for effectiveness. In his article titled ‘On the importance of a certain slowness’, Cilliers argues that through analysing the temporal nature of complex systems, one can demonstrate that a slower approach to change is vital for understanding how complex systems generate identity, meaning, and a sense of memory. Moreover, reflecting on our choices and decisions in a slow and deliberate manner may also mean that we are better able to anticipate the influence that certain interventions may have on the system in question. For Cilliers, the idea of slowness does not merely amount to a suggestion on how we may approach complexity, but instead has a clear moral imperative attached to it. Along with Wendy Parkins (2004), Cilliers argues that a serious engagement with temporality could constitute the ‘ethics of time’, and he explicitly links the idea of slowness with that of integrity. His views on slowness also have political implications, as is clear from the following quote: At this point the argument for slowness becomes a political one: We should put up some resistance to a culture in which being fast is a virtue in itself. We should say “no” with a little more regularity (page 221).

12 

 Rika Preiser & Minka Woermann

Through this unusual reflection on the temporal nature of complex systems, Cilliers demonstrates that an engagement with complexity allows for new ways of thinking that challenge us to explore alternative possibilities of knowing and being, which may very well influence the manner in which we act and live in the world.

3 Part 2: Posthumous after 2011 3.1 Introduction to Cilliers’ posthumous articles The articles in this section represents a continuation of Cilliers’ earlier interests in the further development of the productive linkages between post-structuralism and complexity. Specifically, Cilliers’ reading of Derrida’s deconstructive philosophy is employed to further our understanding of the notion of economy, critique, and ethics. Collaborations in this section demonstrate Cilliers’ influence amongst his students. All the articles in this section were co-authored with students who Cilliers had taught on undergraduate and postgraduate level, and whose doctoral dissertations he had supervised shortly before his death. Cilliers was a generous supervisor, keen to empower his students and share his knowledge. It is for this reason that he (together with Prof. Jan-Hendrik Hofmeyr) started the Centre for Studies in Complexity Colloquium at the beginning of 2009. Both Cilliers’ and Hofmeyr’s postgraduate students were invited to attend the colloquium, along with interested academics and practitioners. During these sessions, Cilliers and Hofmeyr encouraged lively debate centred on topical articles in complexity, and many of the insights gleaned at these meetings found their way into theses and dissertations. Two prominent theorists that were studied at these colloquia were the French sociologist and complexity theorist Edgar Morin, who had a huge influence on Cilliers’ later works, as well as on the work of his students; and Robert Rosen, a theoretical biologist whose views underpin Hofmeyr’s own research in complex systems biology. Cilliers’ passion for his work and his support of his students is what ultimately gave rise to the three articles included here, all of which were started in collaboration with Cilliers, but completed by his students after his death. These articles bear testimony to the legacy that Cilliers left behind, and that continues to be carried forward in the work of those whom he inspired.

Introduction 

 13

3.1.1 Theme 1: Critical complexity The call for a self-critical reflection on our theories and practices ultimately defines the complexity enterprise as a critical enterprise. This insight became increasingly evident in Cillier’s later work, and the importance of the critical attitude underscores the arguments presented in the set of collaborative articles presented under this last theme. The chapter titled ‘Deconstruction and complexity: a critical economy’ (co-authored with Rika Preiser and Oliver Human; page 225), probably presents the most in-depth exploration of the linkages between complexity and deconstruction in this collection. Herein, the work of Morin and Derrida is compared in order to tease out the contributions that both these thinkers make towards understanding complex phenomena. Specific themes that are explored are the economy of complex systems, and the role of critique. These themes are used to inform an understanding of the economy of critical complexity thinking, in which thinking itself is defined as requiring an interruption and recognition ‘of the limitations that each different orientation of thought has to offer’ (page  238). It is a mode of thinking in which limits are critically negotiated. Knowledge and practices are shown to emerge as the interplay between a simultaneous rupture and reconciliation of the old and the new, in which the order of the calculable rubs up against the order to the incalculable. For Derrida, this mode of thinking is best characterised by the notion of critique as stricture, which Preiser et al. interpret as ‘a restorative critical practice, which allows for new and alternative ways of negotiating complex realities’ (page 236). Given this understanding of critique, Preiser et al. question the type of ethics that the economy of critical complexity gives rise to. They steer away from a prescriptive account of ethics, arguing that ‘the ethical moment is born when we have reached the limits of our computing or equalling strategies’ (page  240). In other words, according to them, ethics is born in the moment of thinking ‘in which we take the leap from that which is known to that which is uncertain or unknown’ (page 240). The interplay between the known and the unknown is given further attention in the chapter titled ‘Towards an economy of complexity: Derrida, Morin and Bataille (co-authored with Oliver Human; page  245), in which Human and Cilliers explore the notion of economy at the hand of Derrida’s reading of Georges Bataille’s general economy. In this article, Derrida’s reflection on Bataille’s notion of a restricted economy is highlighted, and it is shown how such a utilitarian conceptualisation differs from a general (or excessive) economy. However, Derrida argues that instead of having two separate economies (as Bataille postulated), the general economy resides in the very heart of a restricted economy, thereby preventing systemic closure. Cilliers and Human employ this view of

14 

 Rika Preiser & Minka Woermann

economy to argue that complex systems are examples of economies, to the extent that they exist within certain sets of possibilities and constraints (page  248). They further argue that the term ‘economy’ is a useful placeholder to describe the ‘slippery’ nature of dealing with complexity (page  261), and support this statement by discussing a number of implications that the term ‘economy’ holds for the manner in which we approach complex systems. Both of the articles discussed above provide additional support for Cilliers’ earlier claim that complexity cannot be known or engaged with in all its fullness, and that, acknowledging this influences the manner in which we understand and employ our knowledge strategies and practices. This brings us to the final chapter, titled ‘The ethics of complexity and the complexity of ethics’ (co-authored with Minka Woermann; page 265), in which the ethical stance that informs all of Cilliers work is explicated and systematised in detail. As concerns the ethics of complexity, Woermann and Cilliers make the claim (by now familiar to the reader) that every engagement with complex systems necessary entails normative considerations. Ethics is thus defined as a structural element of complexity thinking, an argument which is elaborated upon in more detail in Woermann (2013). As concerns the complexity of ethics, Woermann and Cilliers argue that, since all conceptual systems are necessarily limited, we should be vigilant when subscribing to a given theoretical position. Vigilance is safeguarded by practicing (what Cilliers coined) the provisional imperative, which is ‘a strategy for remaining open to complexity at the same time that we reduce complexity through our decisions and actions’ (page  270). Woermann and Cilliers further elaborate on four operations or mechanisms – introduced in an earlier publication by Cilliers (Preiser & Cilliers 2010) – that can aid one in practicing the provisional imperative, and that ‘serve to reinforce and promote the critical attitude, namely provisionality, transgressivity, irony, and imagination’ (page  271). As previously argued, rather than diminishing over time, Cilliers’ concern for the normative implications of his positions continued to grow, and this article is testimony to the fact that the question of ethics was one of his central preoccupations at the time of his death.

4 Conclusion 4.1 Applications of complexity thinking After his death, Cilliers was often referred to as a modern-day Renaissance man, due to his diverse interests and talents. The many co-authored articles that he

Introduction 

 15

wrote with colleagues from a wide range of disciplines demonstrates Cilliers’ talent for transdisciplinary research and collaborations. Three co-authored articles were published posthumously. These articles are not included in this volume due to restricted space, and they are also readily available online and are listed in the reference here below. A short summary of the scope and main arguments will be discussed here. The collaboration with Basarab Nicolescu in the article titled, ‘Complexity and transdisciplinarity – Discontinuity, levels of Reality and the Hidden Third’ (in Futures, 2012), was written during Nicolescu’s visit in Stellenbosch as a STIAS (Stellenbosch Institute for Advanced Study) fellow. Cilliers and Nicolescu’s relationship spanned over many years and they shared lively interests in diverse topics ranging from physics, literature, and complexity to the art of enjoying wonderful food. The argument in favour of connecting complexity thinking with transdisciplinarity is based on the fact that all knowledge of complex phenomena will, in principle, always be partial. This limitation creates opportunities for collaboration with other fields of study and, even more importantly, with non-academic stakeholders, in the effort to explore different possibilities and models for dealing with complexity. The argument put forward in this article is informed by insights gleaned from contemporary theories of complexity, which are employed to elucidate several aspects of a general ‘Theory of Transdisciplinarity’. These ideas are then linked to notions of discontinuity, ‘levels of Reality’, and ‘the Hidden Third’, all of which are central components of Nicolescu’s transdisciplinary framework. By coupling Cilliers’ complexity insights with Nicolescu’s own research insights, further possibilities for research and collaboration are opened up. The central insights to emerge from the article is that the transdisciplinary project should be built on the attempt to capture relations of integration in our knowledge generating practices, and that these knowledge practices could be extended and applied to fields that, for example, deal with complexity in social, political, and economic matters. The last two collaborative publication projects on which Cilliers worked shortly before his death, were the articles on ‘Complexity, modeling, and natural resource management’ (Ecology & Society, 2013) and ‘Complicated, complex and compliant: Best practice in obstetrics’ (Cognition, Technology & Work, 2913). These articles illustrate how versatile the conceptual engagement with complexity can be when one is challenged to apply it to current trends and practices in the world that exists beyond the ivory tower. As previously mentioned, Cilliers’ list of publications with co-authors span over a number of very different themes, which begs the question as to what the connection is between these articles. We argue that each of these articles is proof of the potential for collaboration that the

16 

 Rika Preiser & Minka Woermann

engagement with complexity holds (as suggested by the argument on transdisciplinarity). The article on natural resource management appeared in Ecology and Society, which is the leading journal in social-ecological systems research. This article was the outcome of a networking forum on social-ecological sciences, established under the banner of the National Research Foundation funded ‘Akili project’. The scientists who partook in this forum were using complexity in their research, which afforded Cilliers and Hofmeyr the opportunity to write this article with leading South African researchers in the field of environmental management and systems thinking. Co-authors include Harry Biggs, Sonja Blignaut, Aiden Choles, Graham Jewitt, and, the transdisciplinary expert, Dirk Roux. In the article the authors explore a complexity orientation when considering management strategies in the field of natural resource management, in an effort to pioneer new, more sustainable, and meaningful avenues for decision-making and action. Whilst conventional organisational structures rely on the command-and-control styles of resource management, the authors argue that – when faced with complex management challenges – an adaptive approach is more appropriate. Such an approach is sensitive to complexity and can therefore provide productive alternative modeling and management strategies to deal with issues arising from uncertainty and change. The authors provide a rich conceptual framework for understanding complexity based on Cilliers’ characteristics of complex systems and the implications that complexity has for the study of such systems. They go further by unpacking these implications in practical terms in providing five suggestions of how a complexity thinking approach could inspire managers to proceed differently. These five practical suggestions are summarised in the following imperatives: 1) harness diversity; 2) acknowledge provisionality and keep revising; 3) build (mental) models in a systemic way; 4) measure, scan, and sense; and, 5) have reasonable expectations of appropriate design. These suggestions can act as the basis for any discussion in very nearly any field of application aimed at engaging with the positive and practical implications of adopting a complexity approach. The above challenge of transferring theoretical ideas to practical matters is also characteristic of the third article in this section, titled ‘Complicated, complex and compliant: Best practice in obstetrics’, and co-authored with Sydney Dekker, Johan Bergström and Isis Amer-Wåhlin. In this article the focus and methodology is however different, in that the authors draw on the distinction between ‘complicated’ and ‘complex’ matters to shed light on compliance with best practice guidelines in the field of patient safety management issues. The distinction between ‘complicated’ and ‘complex’ is discussed at length in the work of Cilliers (1990, 1991, 1998) and a number of other authors (cf. Dyke 1988; Morin 2007; Poli

Introduction 

 17

2013; Rosen 1985; Richardson 2001, 2002), all of whom dispel the notion that the distinction is superficial (i.e. merely a matter of perspective or subjective interpretation). Instead, they argue that the distinction hinges on an order difference between complex and complicated phenomena. Based on a case study that was conducted at the surgical theatres and labour wards at a Scandinavian hospital, the authors contend that obstetric practices and interventions deal with complexity par excellence. In light of this, the standard intervention procedures should be re-evaluated in terms of the fact that clinical obstetric practice belongs to the domain dealing with complex as opposed to complicated phenomena. Given that the norms for clinical intervention in complex systems are contextual and contingent, varying with time, technology, and social-clinical composition (Dekker et al. 2013), the authors conclude with a very convincing argument that the notions of ‘best practice’ and ‘compliance’ can only be employed in matters relating to complicated phenomena ‘whose functioning is, in principle, exhaustively knowable, closed to environmental contingency, and for which single best methods can be drawn up’ (Dekker et al 2013: 194). In contrast to dealing with complicated phenomena, complex phenomena call for a diversity of methods and interventions that are sensitive to the radical contextuality and openness of such systems. As a result, the study of complexity demands that we constantly and self-critically reflect on our conceptual framings and practical interventions when negotiating the hurdles of complexity in everyday life.

5 Significance of Cilliers’ work This introduction attests to both the breadth and depth of Cilliers’ work, which makes it difficult to definitively define the significance of this work. Nonetheless, we conclude with some remarks concerning his influence in developing an understanding of complexity, and the implications that this understanding holds for thinking about the human condition. Between 1990 and 2010 ideas around complexity were starting to emerge in a number of unrelated disciplines. Cilliers took cognisance of these developments and engaged critically with a number of complexity theorists including John Holland, Peter Allen, Kurt Richardson, and David Byrne. His in depth understanding of these new developments in complexity, and his ability to conceptually integrate a number of these emergent insights into a general understanding of complexity, led to a sophisticated and coherent account of the nature of complex phenomena.

18 

 Rika Preiser & Minka Woermann

Contrary to many other accounts of complexity that understand complexity in terms of an epistemological framework (cf. Luhmann 1995; Morin 2008), Cilliers’ central contribution lies in his ontological understanding of complex phenomena, which is born out of the conviction that reality and its organising processes are inherently complex. This is not to say that questions concerning epistemology and ethics did not interest Cilliers. To the contrary: as demonstrated above, a large part of work was dedicated to exploring the epistemological and ethical implications emerging from his ontological understanding of complexity. It must however be noted that Cilliers had reservations regarding the strict distinctions between these three levels of study, often arguing that, when engaging with complexity, the boundaries between ontology, epistemology, and ethics become fuzzy. This is because any serious epistemological engagement with complexity (understood as an ontological phenomenon) necessarily raises normative questions. Therefore, for Cilliers, engaging with complexity was not just an academic exercise, but concerned a deep reflection on what it means to be human. In this regard, he often reminded us that there are two types of people in the world: those who engaged with the rich and diverse wonders of our complex world, and those who thought that there were only two types of people in the world! To our minds, his reflection on the human condition characterises Cilliers’ unique contribution to the field of complexity studies, and also marks his engagement with complexity as deeply philosophical. At the time of his death, Cilliers’ earlier insights were maturing in that he had moved beyond establishing the grounds for understanding complex phenomena in general, to exploring and elaborating on the rich implications that the paradigm of complexity holds for rigorously and critically engaging with the big socio-political questions of our time. We believe that the world is poorer as a result of his untimely death, but we are thankful for the body of knowledge that he has left behind. This body of knowledge marks his unique legacy, and we hope that this collection honours this legacy, brings his insights to new audiences, and continues to inspire all who read his work. Rika Preiser and Minka Woermann 2 December 2014, Stellenbosch, South Africa.

References Allen, P. 2001. What is Complexity Science? Knowledge of the Limits to Knowledge. Emergence, 3(1): 24–42. Byrne, D. 1998. Complexity Theory and the Social Sciences. New York: Routledge.

Introduction 

 19

Cilliers, P. 1993. Modelling Complexity. PhD Dissertation, Department of Philosophy. University of Stellenbosch. Cilliers, P. 1998. Complexity and Postmodernism: Understanding Complex Systems. London: Routledge. Cilliers, P., Biggs, H.C., Blignaut, S., Choles, A.G., Hofmeyr, J.S., Jewitt, G.P.W., Roux, D.J. 2013. Complexity, modeling, and natural resource management. Ecology and Society, 18(3): 1. http://dx.doi.org/10.5751/ES-05382-180301. Cilliers, P., Dekker, S., Hofmeyr, J-H.S. 2011. The complexity of failure: Implications of complexity theory for safety investigations. Safety Science, 49: 939–945. Cilliers, P., Nicolescu, B. 2012. Complexity and Transdisciplinarity – Discontinuity, levels of Reality and the Hidden Third. Futures, 44: 711–718. Dekker, S.W.A., Bergström, J., Amer-Wåhlin, I., Cilliers, P. 2012. Complicated, complex and compliant: Best practice in obstetrics. Cognition, Technology and Work, 15:189–195. Dyke, C. 1988. The Evolutionary Dynamics of Complex Systems: A Study in Biosocial Complexity. New York: Oxford University Press. Georgiou, I. 2007. Thinking Through Systems Thinking. Oxon & New York: Routledge. Holland, J. H. 1999. Emergence: from chaos to order. Reading, Mass: Perseus Books. Kunneman, H. 2010. Ethical Complexity. In: Cilliers, P., Preiser, R. (eds). Complexity, Difference and Identity: An Ethical Perspective. Dordrecht: Springer, 131–164. Luhmann, N. 1995. Social Systems, trans. J. Bednarz Jr. (with D. Baecker). Stanford: Stanford University Press. Mitchell, S. 2007. Why Integrative Pluralism? In: Cilliers, P. (ed.). Thinking Complexity. Complexity & Philosophy – Volume 1. Mansfield, MA. ISCE Publishing, 7–24. Morin, E. 2007. Restricted Complexity, General Complexity. In: Gershenson, C., Aerts, D., and Edmonds, B. (eds.). Worldviews, Science and Us: philosophy and complexity. London: World Scientific Publishing, 5–29. Morin, E. 2008. On Complexity, trans. S.M. Kelly. Cresskill: Hampton Press. Parkins, W. 2004. “Out of time: Fast subjects and slow living,” Time and Society, ISSN 0961-463X,13(2): 363–382. Poli, R. (2013). A Note in the Difference between Complicated and Complex Social Systems. Cadmus, 2(1): 142–147. Preiser, R., Cilliers, P. 2010. Unpacking the ethics of complexity: concluding reflections. In: Cilliers, P., Preiser, R. (eds.). Complexity, Difference and Identity. Dordrecht: Springer, 265–287. Richardson, K.A. 2002. Methodological implications of complex systems approaches to sociality: some further remarks. Journal of Artificial Societies and Social Simulation, 5(2). Available online at: http://jasss.soc.surrey.ac.uk/5/2/6.html Richardson, K.A. 2001. On the status of national boundaries: a complex systems perspective. Proceedings of the Systems in Management 7th Annual ANZSYS Conference, Edith Cowan University, Churchlands, 27–28 November 2001, 229–238. Rosen, R. (1985) Anticipatory Systems: Philosophical, Mathematical and Methodological Foundations. Oxford: Pergamon Press. Thrift, N. 1999. The Place of Complexity. Theory, Culture & Society, 16(3): 31–69. Woermann, M. 2013. On the (Im)Possibility of Business Ethics: Critical Complexity, Deconstructoion, and Implications for Understanding the Ethics of Business. Springer: Dordrecht.

20 

 Rika Preiser & Minka Woermann

Wolkenhauer, O., Ullah, M. 2007. All Models are Wrong. In: Boogard, F., Bruggeman, F., Hofmeyr, J., Westerhoff, H. (eds.). Towards a Philosophy of Systems Biology. Amsterdam & Boston: Elsevier, 164–179. Wood, D. 1990. Philosophy at the Limit. London: Unwin Hyman. Zadeh, L., Polak, E. 1969. Systems Theory. New York: McGraw Hill (electronic series).

Part 1: Single-authored Papers Theme 1: Characterising Complexity

Paul Cilliers

The brain, the mental apparatus and the text A post-structural neuropsychology I sing the body electric (Walt Whitman).

1 Introduction The desire of this work is to provide descriptions of the brain, and of what is called the brain’s “higher” functions (perception, memory and consciousness), that are neither crudely reductionistic, nor based on abstract or dualistic postulates. These descriptions have to be materialist (in a non-metaphysical sense of the word) without losing the ability to say something about the higher functions. The route taken in the process was the following: In the first place, information about the present state of neuropsychological theories was gathered. This information was related to Freud’s model of the “mental apparatus” in his early Project for a scientific psychology (Freud 1950). Such a reading of Freud suggested that there may be similarities between the way the brain works, and structural theories of how language works, a suspicion that was confirmed by a reading of Saussure (1974). These descriptions are placed in a post-structural context through Derrida’s scrupulous reading of both these texts – of Saussure in Of grammatology (Derrida 1976) and of Freud’s project in Freud and the scene of writing (Derrida 1978: 196–231). An interweaving of these texts with neurological theory results in what can be called a post-structural neuropsychology. Let me underscore the importance of language in these descriptions. Language is not only important in post-structural descriptions of the brain, there are also firm links between the rule-based computational models of higher brain functions and formal theories of language. As a matter of fact, there is an exact mathematical equivalence between the formal grammars proposed by Chomsky (1957) as a description of the structure of language and the abstract mathematical descriptions of general models of computers provided by Alan Turing (1936). This makes post-structural descriptions of language significant on two levels, a negative and a positive one. In the first place it deconstructs formal theories of language and thereby all related formal systems based on production rules, including computational models of the brain. In the second place it provides a Originally published in the South African Journal of Philosophy, 1990, 9(1): 1–8. © South African Journal of Philosophy.

24 

 Paul Cilliers

model that can be used to generate alternative descriptions of the brain. If this model is useful, it would show that post-structuralism is not merely a parasitic form of discourse analysis, but a constructive component of post-modern culture.

2 The brain Functionally, the human nervous system consists only of neurons. Groups of neurons that perform specific tasks, usually closely related to some bodily function, are clustered together in specific architectural structures, like the cerebellum or the hypothalamus. The largest part of the brain, the cortex, is however fairly homogeneous, and is in fact nothing more than a large sheet of richly interconnected neurons that is crumpled up to fit inside the cranium. Whatever their function and location, all neurons operate on the same principles: On the dendrites of the neuron incoming information is integrated spatially (i.e. relative to the position on the complex structure of dendrites where the information arrives) as well as temporally (i.e. relative to the frequency of the repetition of a certain message). If this process of integration exceeds a certain level, determined at the foot of the cell body, an impulse is generated and propagated down the axon of the neuron to serve as input to the dendrites of other neurons (Stevens, 1979). Despite the immense amount of neurons in the cortex, they are interconnected so richly that the route between any two neurons seldom involves more than a few steps. This results in a vast amount of loops and feedback circuits (Mountcastle 1975, 1978). A neuron does not operate on its own, but in concert with a number of others. Such a group of neurons is often referred to as a “neural assembly” (Palm 1982). There is therefore no significance in the firing of a specific neuron, it is always a pattern of activity that signifies something. Another characteristic of the functioning of the neuron concerns the way in which there is dialectic between the discrete and the continuous, or the digital and the analogue. The integration of the graded potential on the dendrites is a continuous process that results in the generating (or not) of the action potential in the axon, a discrete process. The action potential, however, does nothing more than initiate activity on the dendrites of a number of other neurons, and is therefore subsumed under a continuous process. The pattern of activity in a neural assembly will therefore always include digital and analogue processes simultaneously.1

1 For a discussion of the digital and the analogue, and the role the two processes generally play in communication see Wilden (1980: 155–201).



The brain, the mental apparatus and the text 

 25

The cortex can be seen as a distributed system. Activity is not localised, but involves patterns of interconnections that cover large areas. This idea is developed by Karl Pribram (1971, 1979) in his holographic theory of cortical functioning. A hologram is a complex kind of photograph that can be taken of an object by means of coherent light, like a laser beam. There is, however, not a one-to-one correspondence between the object and its representation in the hologram. The hologram looks nothing like the object at all. The information concerning the represented object is smeared over the whole hologram. This implies that loss of a specific part of the hologram will not result in the loss of a specific part of the object. What is more, the whole represented object can be regenerated from a small part of the hologram, with loss only in quality and definition. According to Pribram, the cortex functions in similar ways. Loss of a specific part of the cortex does not result in the loss of a specific function or memory. The brain can actually sustain a lot of damage, as is demonstrated by the severing of the corpus callosum – the large nerve bundle that connects the two hemispheres of the brain as a treatment for severe epilepsy. This major lesion has virtually no effect on the performance or behaviour of the subject in question. This can only be the case if functions are not localised, but distributed over large areas. The question to be answered, however, is the following: how does a distributed system of neurons sustain the higher functions of memory, perception and consciousness? Let us take this question to Freud.

3 The mental apparatus Very early in Freud’s career (in 1895), when he was still primarily a neurologist, he developed a model for the functioning of what he calls “the mental apparatus”, posthumously published as the Project for a scientific psychology (1950). This astounding work covertly served as a basis for much of his later work, and is virtually completely compatible with modern neurology (see Pribram & Gill 1976) The model consists of two types of interconnected neurons; what the he calls the φ and the ψ systems. The first is the perceptual system, and the second the “psychological” system. The only difference between the two systems lies in the way they are interconnected. The neurons transport energy, or what he calls “quantity”. In the perceptual system, the flow of quantity is unimpeded. In the ψ system the neurons are separated by “contact barriers”, and pathways have to be breached by the flow of quantity. lf the connection between two neurons are regularly active, the resistance of the contact barrier is broken down, and the pathway

26 

 Paul Cilliers

becomes “facilitated”. The characteristics of the mental apparatus is determined by the pattern of facilitations in the ψ system. Perception is for Freud just a registration of stimuli. This does not mean that the world is merely copied inside the brain. In the first place, the stimuli have no ideational content, they are merely bits of quantity that will be channelled in certain ways by the mental apparatus. In the second place, there is for him no difference between stimuli from outside the body and from inside the body (one of the many post-structural moments foreshadowed in Freud). Stimuli from the outside and the inside are “perceived” by the mental apparatus in the same way by channelling the received energy into the routes already opened up, or by using it to breach open new routes. The main characteristic of nervous tissue for Freud, is “memory”. (Freud 1950:  299). Memory is the furrowed pathways in the ψ system, the patterns of facilitated contact barriers. If the contact barriers were all equally facilitated, however, the system would be homogeneous, and the property of memory would not emerge. Freud therefore makes a small but vital change to the definition of memory. It is now “represented by the differences in the facilitations of the ψ neurons” (300, my emphasis). Memory does not lie in the facilitated pathways themselves, but in the relationship between facilitations, and this relationship is one of differences. Freud’s discussion of memory and perception does not include any reference to consciousness. They are both unconscious processes. Not unconscious in the sense that they are part of some entity in the psyche, merely unconscious in the sense that they have not passed through the filter of consciousness because they are before consciousness. This is how he formulates it: Hitherto, nothing whatever bas been said of the fact that every psychological theory, apart from what it achieves from the point of view of natural science, must fulfil yet another major requirement. It should explain to us what we are aware of, in the most puzzling fashion, through our ‘consciousness’; and, since this consciousness knows nothing of what we have so far been assuming – quantities and neurones – it should explain this lack of knowledge to us as well (Freud 1950: 307, 308).

When he has to answer the question of consciousness, Freud provides an ad hoc solution. He postulates another system of neurons, the ω system, embedded in the ψ system that is impervious to quantity. Here we only have a play of “qualities”, and this is what we are conscious of. This ad hoc solution is unnecessary because, as we shall see, Freud has already said all there is to say about consciousness. That Freud is unhappy with this solution is apparent from the Project itself (311), and from the fact that all references to the ψ system is dropped when Freud returns to these matters 30 years later in A note on the ‘Mystic writing-pad’ (1925).



The brain, the mental apparatus and the text 

 27

In this fascinating little article Freud dispenses with all characteristics of the mental apparatus except memory and perception. The guiding metaphor of the whole article is that of “writing”. But let us give meaning to “writing” by following another route.

4 The model of language The characterisation of the brain as an open system where entities have no significance on their own, but derive significance from their relationships with the other components of the system, bears a striking resemblance to Saussure’s characterisation of language. For Saussure language consists of a system of signs that each consist of a signifier (say, the word) and a signified (its meaning). The relationship between signifier and signified is completely arbitrary, there is no “natural” link between the two. The meaning of a sign is however not determined by the user performing a speech act. The system of language (“langue”) is constituted by a set of relationships that transcends the individual user. Although the sign has an arbitrary nature that is not manifested anywhere else but in its use, man is always already caught up in the system of language. The relationship between signs that constitute their meaning is one of differences. The meaning of a sign is related to the way in which it differs from all the other signs in the system. The signifier “brown” does not have a meaning because it can be identified with a concept that unambiguously contains the essence of “brownness”, but because it can be differentiated from the signifiers “black”, “blue”, “grey”, and also from “squirrel”, “fly” and “impeccable”. “In language there are only differences without positive terms” (Saussure 1974: 120). Because all the signs are related, any changes or any additions will eventually reverberate through the whole system, and in the end change even the agent of change. Such change, however, is always incremental and therefore Saussure’s system will not tolerate the concept of an epistemological break. Despite his insistence that language is a system that transcends the individual user, the sign remains for Saussure a psychological entity (66). Although the sign has no essential nature, the speaker always somehow gets it right. Because the signifier and the signified are only fully united in the speaking subject, speaking is, for Saussure the true form of language. In writing, the close relationship between the speaker and his intention breaks down. It is exactly on this point that Derrida concentrates in his critique of Saussure. lf the sign is arbitrary, if it is a changing system that transcends the individual user, then the meaning of words cannot be pinned down in the subject of the

28 

 Paul Cilliers

user, and “writing” becomes the metaphor that best describes language, even spoken language. The mental component of the sign, the signified, falls away, and there is only an endless chain of signifiers that refer to each other. In all senses of the word, writing thus comprehends language [...] There the signified always already functions as a signifier. The secondarity that it seemed possible to ascribe to writing alone affects all signifieds in general, affects them always already, the moment they enter the game. There is not a single signified that escapes, even if recaptured, the play of signifying references that constitute language (Derrida 1976: 7).

There is another important difference between Derrida’s and Saussure’s understanding of language. Although Saussure gives a description of language that is not rule-based, but relational, these relationships are fairly fixed. Although the sign has no identity of its own, it is well constituted by its relationships. His linguistic system is what Harland calls a “total simultaneous system”: Simultaneous, in that the system only balances if words push against each other at exactly the same time; total, in that the system only balances if there are no internal gaps to give words room for falling, and no surrounding void to give words room for dispersing. For a state of perfect equilibrium, words need to be packed tightly up together within a closed space (Harland 1987: 136).

For Derrida, the relationships between signs are not that stable. They are not closely stacked at all, but well separated by a spacing. The meaning of the sign is generated by the playful interaction of signs in this space. Meaning is the effect of play, and not determined by relationships. Instead of pinning it down, the interactive nature of the sign allows meaning to proliferate, to be excessive. To describe the mechanism of interaction between the components of open systems, Derrida employs amongst others, the concepts of “trace” and “différance”. Both are complex, playful concepts with elusive meanings. They cannot be defined, but I will attempt a brief description of them in order to render them useful. The meaning of “trace” can be unfolded by starting at some of the word’s usual meanings. Because a sign is the effect of its relationships with all the other signs in the system, each sign bears traces of every other sign. Because the meaning of all signs is the result of traces, there are no fixed reference points that serve as origins of meaning. Meaning can be traced, but never tracked down. Whether in written or in spoken discourse, no element can function as a sign without relating to another element which itself is not simply present. This linkage means that each ‘element’ phoneme or grapheme is constituted with reference to the trace in it of the other elements of the sequence or system [...] Nothing, either in the elements or in the system, is



The brain, the mental apparatus and the text 

 29

anywhere simply present or absent. There are only, everywhere, differences and traces of traces (Derrida 1981: 26).

To describe the mode of interaction in an open system, Derrida introduces the concept of “différance”. This is a polysemic word that does not have a “usual meaning”. The first sense of the word results from Saussure’s description of language as a system of differences. Traces are traces of difference, even if the differences are not of equal strength. In the play of differences, meaning is generated. However, this play is always still in progress, the meaning is never produced finally, but continuously deferred. As soon as meaning is generated for a sign, it resonates with the system, and this disturbance in the traces is invariably reflected back, shifting the meaning of the sign in question, even if imperceptibly. Because no trace is in a privileged position, “defer” can be read not only in the temporal sense, but in the sense of “to submit to”. Each trace is not only delayed, but also subjugated by every other trace. The mechanism of interaction between traces has still more to it than to differ/defer. The mechanism is not a passive characteristic of the system, nor the result of positive act. Différance remains suspended somewhere between active and passive – the dynamics of a system that does not exist outside its use by human subjects, but none the less by subjects who are in turn embedded in the system. An element of différance that is not apparent from its polysemia, is that of “spacing”. In order for signs to interact by means of traces and différance, they cannot, as we have seen above, be stacked tightly against each other. There must be space, and this space is the site of action. The creation of space for playful interaction is therefore part of the function of différance, besides being a precondition for it. In this way différance has both spatial and temporal characteristics (Derrida 1982: 89). Trace and différance play vital roles in Derrida’s reading of Freud. By retranslating “facilitation” with “breaching” and by stressing Freud’s point that memory is the result of differences in breaches, Derrida provides a description of memory as trace. We then must not say that breaching without difference is insufficient for memory; it must be stipulated that there is no pure breaching without difference. Trace as memory is not pure breaching that might be reapprophated at any time as simple presence; it is rather the ungraspable and invisible difference between breaches. We thus already know that psychic life is neither the transparency of meaning nor the opacity of force but the difference within the exertion of forces (Derrida 1978: 200).

But the matter is more complicated than the intractability of the trace. Mere breaching, mere control of the flow of quantity is only sufficient for the imme-

30 

 Paul Cilliers

diate, reflexive primary processed. The higher functions that are necessary for survival, the secondary processed, depend on quantity that is not immediately divested, on a deferment of the force of quantity. “All there differences in the production of the trace may be reinterpreted as moments of deferring. In accordance with a motif which will continue to dominate Freud’s thinking, this movement is described as the effort of life to protect itself be deferring a dangerous cathexis, that is, by constituting a reserve” (202). Differences deferred, not by an active ego, or as the passive flow of quantity, but by the logic of différance. No doubt life protects itself by repetition, trace, différance (deferral). But we must be wary of this formulation: there is no life present at first which would then come to postpone, or reserve itself in différance. The latter constitutes the essence of life. Or rather: as différance [...] Life must be thought of as trace before Being may be determined as presence [...] It is this the delay which is in the beginning (203).

Part of the logic of différance is therefore the play of repetition. In the origin of the psyche, its “essence” is constituted by trace and différance, there is no first time. Ache-trace is différance, is always already trace of trace. This does not mean that traces are not real. Derrida insists on the materiality of the trace, both in the context of writing and of the psyche (an issue we will presently return to). This insistence allows him to underline the radical reflationary nature of the sign (whether linguistic or neural) without the introduction of idealistic or mystic elements. The sign is real, but has no essence. This has two important consequences for the description of the psyche. First, if both language and the machinery of the psyche have no origin, but are generated by the play of différance and trace, the psyche becomes text: “there is no domain of the psyche without text” (199). If the psyche can be treated as text, we can follow the same strategy used to interpret language when we deal with the psyche. In the second place, if the machinery of the psyche is constituted by traces only, there is no governing role to be played by a conscious ego. The play of traces is unconscious. This idea must once again be tracked carefully. The unconscious does not become a place where the true text of the psyche can be found, and from which consciousness is in some form “transcribed”. There is no “other” that escapes the disseminating supplementarity of the trace. There is always, and only the other, but the other us always already subjected to the breaching of the trace. The other is the trace (Derrida 1978: 211, 212). To claim such a pervasive role for the trace must not be rad as an ontological claim for the sense that the universe is constructed out of linguistic entities. It is a claim rather that our models and understanding of the world are always mediated linguistically, and that the model of language is an excellent one for describing other complex systems.



The brain, the mental apparatus and the text 

 31

The description of the psyche as writing is elaborated on in Derrida’s reading of Freud’s A note on the ‘Mystic writing-pad’ (221–231). He stresses the mechanical nature of Freud’s model for the working of memory and perception. The writing-pad is a machine (226–227). The unconscious substrate of the mental apparatus is represented by “natural wax”, and the production of traces by the force of the stylus happens in a purely physical way. In this way Derrida’s theory of writing becomes a materialist one. His materialism should not be confused with any kind of positivistic position. “Realism or sensualism  – ‘empiricism’  – are modifications of Iogocentrisms” (Derrida 1981: 64). In opting for a materialism by discarding mentalist concepts, Derrida also transforms the concept of matter. “If, and in the extent to which, matter in this economy designates [...] radical alterity [...] then what I write can be considered ‘materialist’” (64). [Derrida’s] kind of materialism is clearly very different to those neurological positivists’ kind of materialism. When neurological positivists study neurones and molecules and electrical charges in the brain, they are studying positive things and entities. The ultimate assumption behind such study is the old atomizing assumption that the scientist’s key for unlocking all secrets is simply to observe entities on a sufficiently small scale. But in the meanwhile, the question as to how such entities might signify is conveniently left to one side (Harland 1987:147).

Derrida, affirms the materiality of the brain, but insists that matter as such has no meaning, wears no meaning on its face that is not mediated by the discourse in which it appears. Meaning is a result of the relationships between the traces inscribed in matter. “The subject of writing is a system of relations between strata: the Mystic Pad, the psyche, society, the world. Within this scene, on that stage, the punctual simplicity of the classical subject is not to be found” (Derrida 1978: 227). With the decentring of the subject, the concept of “mind” as self-present consciousness is also dislodged, and linked with other examples constrained by the metaphysics of presence. According to Derrida, consciousness (at least in the ordinary sense) is an illusion that human beings have invented because they have feared the consequences of a materialist conception of the brain. In this respect, the modern secular notion of mind is really no improvement upon older religious notions of soul and spirit (Harland 1987: 146).

Is there then an answer to the question “What is consciousness?” I will defer the question until we have interwoven the texts read up to now. It is not possible to construct a comprehensive theory that will account for all the features of the brain, but the grafting of neuropsychology on a post-structural description of language produces illuminating descriptions by relatively sparse means.

32 

 Paul Cilliers

Like the sign, the neuron only has significance in terms of its relationships with others. It is therefore constituted by the play of traces. We have seen how the neuron has both digital and analogue components. The sign (as signifier) works in a similar way. To enable communication, there has to be a discrete sign. The meaning of the sign, however, is generated by the continuous interplay of traces, and the meaning can only be recovered by re-entering the discrete sign in this play that has no distinct borders. The cortex consists of assemblies of interconnected neurons forming loops. Not closed loops, but loops with many entrance and exit points. However changed, after a deferment, neural activity is reflected back on itself. The activity itself initiates the process that changes it. In this way, the working of the cortex is subject to the logic of différance. Memory is the breaching of pathways, the forming of connections. It is the basic characteristic of the mental apparatus. It is the trace. Just as there is nothing before the trace in language, there is nothing before memory in the brain. Certainly not consciousness. Which brings us back to the intertwined problem of perception and consciousness (and memory). Freud’s model of the mystic writing-pad explains perception and consciousness in terms of an interaction between stimuli from outside and the unconscious substrate of memory (Freud 1925). This is confirmed by Pribram’s experimental work (see Hampden-Turner 1981: 94) that shows that there is some “casting forward” of activity from the cortex that meets incoming stimuli just like Freud’s “feelers of the unconscious” that are stretched out and retracted. Perceptual stimuli enscribes the traces in memory, but they are not perceived as such. Perception is the delayed interaction of newly arrived traces with already present ones. It is not a processing of stimuli, but the différance of traces. What then, can be said about consciousness? This is a question that should be resisted, because it carries with it all that heavy baggage of “mind”, “intention”, “feeling” and “reason”, fooling us into thinking that these concepts are somehow pure and important. In writing what has been written about memory and perception, everything about consciousness has been said. The question has decomposed. Still the question “What is consciousness?” insists itself. Fortunately we can compose a clear answer to it. Consciousness is the différance of perception and memory. What has been written has been overwritten.



The brain, the mental apparatus and the text 

 33

5 The brain and machines The venture to create artificial intelligence – that is to program computers so that they behave in an intelligent fashion – is the culmination of logocentric thought. It is often thought that the computer plays such an important role in our culture because it is such a powerful tool, because it performs so many important functions that it insists itself on our thinking. In a way, it is the other way round. Computers are so important because they fit perfectly into a tradition of thought that one would be able to trace from Plato through Descartes, Leibniz, Kant, British empiricism and logical positivism to the instrumental thinking that pervades so much of modern scientific, economic and political thought. It is a tradition that believes in the existence of abstract truths, raw data and stands under the spell of the Rule. Computer technology in general, and Artificial Intelligence in particular, marches under the same banner. The attempts to create artificial intelligence on Turing machines consist of providing the system with the relevant data, and then to manipulate it by means of the appropriate rules. By transforming “raw data” through a finite set of rules it is hoped that some understanding of the data will emerge . Some examples may help to clarify this strategy. Computer processing of natural language normally starts by trying to determine the grammatical structure of the sentence to be processed, and a lot of effort in this field is put into finding good parsing routines. Once the structure of the sentence is determined in terms of some scheme (and there are a number of these), the various components are related to the contents of an internally stored database in order to assign meaning to them. In computer perception, “perceived” data is processed in order to extract certain features (such as edges, contrasts, contours and colour) and these features are then manipulated by means of rules to try and recognise and understand perceived objects. In the technology known as Expert Systems, the knowledge of an expert is reduced to a database and a number of rules that can be programmed in order to enable non-experts access to the expertise. These strategies have had a considerable success in domains that are sufficiently framed that formal descriptions of them can be given. There are computer systems that can do a geological survey, or provide a fair diagnosis of diseases of a specific kind, or even land a Boeing. In terms of everyday human activities though, artificial intelligence has been a resounding failure. There is no computer system that can hold a normal conversation, understand a joke, devise a new recipe or distinguish between a Burgundy and a claret. Because there is such a strong association between the formal systems of computers and the formal models of language, the reasons for the failure of computers to simulate natural intelligence is closely related to the failure of formal

34 

 Paul Cilliers

grammars to give an adequate description of language. The question that immediately imposes itself is the following: If formal descriptions of language can be replaced by a description of language as an open system in post-structural terms, is there an associated computing device that also works on these principles? The answer seems to be yes. There is a rapidly developing technology that relates to these principles, known as Parallel Distributed Processing or Connectionism or Neural Networks.2 It is a highly transdisciplinary field (neuroscience, psychology, cognitive science, mathematics, computer science and engineering all provide inputs) that appears to be very promising. One of the characteristics of what Haugeland (1985: 112) calls Good Old Fashioned Artificial Intelligence, is that a successful simulation of human intelligence need not be achieved by the same processes that produce it in humans. Rationality and intelligence are seen as characteristics that can be abstracted from their specific implementations. To simulate it on a computer you need only the abstracted descriptions. Good Old Fashioned AI took very little notice of actual processes in the human brain. The failure of this strategy has resulted in a reconsideration of the importance of the way the nervous system works. If it can perform all. these impossible tasks, it must have some rather special features. The impetus behind computational models that take the brain seriously has also not come from computer science, but from neuropsychology. The idea of using neuron-like structures for computation is not a new one. It can be traced back to the work of McCulloch and Pitts in the early forties (Anderson & Rosenfeld 1988: 15–41). The first implementations of neuron-like computation devices were the perceptrons of the early sixties (89–150), but their success was short­lived. Towards the end of the sixties Minsky and Papert (1969) provided a mathematical proof that the problems that can be addressed by perceptrons are severely limited, and this discouraged research in the field of neural networks for quite some time. Towards the eighties it was realised, however, that the limi-

2 Literature on this subject is rapidly expanding. An international society and journal for Neural Networks was also established in 1988. The central text, and perhaps the best introduction remains Rumelhart and McClelland (1986). For a collection of articles that provide a historical perspective, see Anderson and Rosenfeld (1988). For a more mathematical perspective, see Arbib (1987), for a more neurological perspective, see Palm (1982) and for a perspective from computer science, see Hillis (1985). For some implications of connectionism for philosophy of mind see Horgan and Tienson (1987) and for a defence of traditional symbol systems against distributed systems, see Pinker and Mehler (1988). Grossberg (1988) collects a number of papers dealing with more technical neurological applications.



The brain, the mental apparatus and the text 

 35

tations of perceptrons can be overcome by means of changing its structure,3 and “Neural Networks” was established as a viable field of research. Neural networks work in the following way: A network consists of a number of “neurons” or units that are richly interconnected. Some of the units may be taken as input units, and some as output units. This “status” of certain units is not fixed. There is no difference between normal units and input or output units, and if circumstances demand it, roles can be reversed. Each unit has a number of units connected to it, and is connected in turn to a number of other units that may include those that provided the input, and the unit itself. Feedback and loops are natural characteristics of these systems. A unit will integrate all the inputs it receives, and if it exceeds a certain threshold, become active and in turn influence the units to which it is connected. The connection between any two units is associated with a certain “weight” that determines the strength of the influence that the active unit will have on the units to which it is connected. As with real neurons, these influences can be excitory or inhibitory. The characteristics of the network is determined by the pattern of weights between the various units. A specific weight, or a specific unit has no significance, it is always the pattern of activity over the whole system that bears meaning. If you want a certain output-state, the weights in the network as a whole is changed in order to achieve this. The network can also be implemented in such a way that weights can be changed automatically. In this way, the network can learn. If a certain input and the desired output is provided, the network will shuffle the weights in such a way that the two are matched. Say for example, you provide the network with a number of verbs at its input, and each time provide the past tense of the current verb at the output. The network will then adjust its weights so that it “learns” the past tense of these verbs. In future it will be able to provide the past tense for each verb by itself, and if it encounters an unknown verb it will automatically come up with the best guess.4 The formal description of a neural network is actually fairly straightforward, but if one examines some of the characteristics of these systems, the significance of this approach to computation becomes more apparent.

3 Perceptrons are devices that have essentially only one layer of connections between “neurons” that can be changed. By adding more layers, the device becomes much more powerful. 4 Such a system has actually been implemented by Rumelhart and McClelland (1986, vol. 2: 216–271). It showed characteristics similar to those of children acquiring language. After a few learning-cycles it got most of the verbs right. After a few more, it got verbs that it previously had correct, wrong, usually because of over-generalisation (adding -ed to irregular verbs). Some more learning was required before it settled in a stable pattern of correct recognition.

36 

 Paul Cilliers

The first important characteristic of neural networks is that knowledge is not represented locally in an iconic fashion (as is the case in conventional computers and rule-based systems), but that it is distributed over the whole system as a result of the fact that what is actually stared is the connection strengths between units (Rumelhart & McClelland 1986 vol. I: 31). Information concerning a specific object is also not identified with the connection strengths associated with a specific unit but is distributed over the connections between many units (33). There is not even a controlling unit or central overseer that orchestrates activity in the network (47), every aspect of the functioning of a neural net is purely relational. “All the knowledge is in the connections” (75). The second characteristic follows from this fact. If the functioning is purely relational, the system cannot be rulebased on a first level, because there are only interactions (traces). The network may appear to follow rules, but these are emergent properties that are abstracted from the functioning of the network, and not principles that determine the functioning (24). If the system is not rule-based, it is of course also non-algorithmic. No external agent has to design a program that must be followed. Learning and computation is done merely by a shuffling of dispersed connection strengths (32). Under the right circumstances the network can perform this “shuffling” by itself, by means of a process that remains non-algorithmic in the sense of an algorithm being an external program. Rules are always of a different logical level than that which it describes. If the notion of rules is relativised, the notion of “levels” becomes problematic as well. All the activity in a neural network is on a simple level of interaction. The less structured the network is, the more powerful it becomes (60). In language processing by neural nets, for example, information from the “levels” of the letter, the word, the syntax as well as the semantic level are all considered simultaneously (6, 7). No single one is privileged above the others. It is therefore unnecessary to have a theory that relates syntax to semantics, because the two are folded into each other in an inextricable way. The simultaneous presence of many elements, and the fact that many things are happening in parallel are features that provide neural network systems with their power. As a result of the distributed and parallel nature of these systems, they have a characteristic that they share with holograms, language and the brain, which can be called “graceful degradation” (29). Damage to any specific part of the system does not result in the loss of a specific characteristic. Because information is distributed over the whole system, damage to a specific part will result in a slight deterioration of the performance of the system as a whole, but not in the loss of something specific. This is an important characteristic of all distributed systems. In an iconic representation damage to a certain part not only entails the loss of that specific component, but, if the component was important enough, it may



The brain, the mental apparatus and the text 

 37

result in a catastrophic failure of the whole system. Such a catastrophe is avoided by the basic characteristics of a distributed system. The method of calculation itself can be seen as a process of “relaxation”. Instead of following a series of logical steps, a solution is reached by finding the optimal compromise between a number of constraints. “The system should be thought of more as settling into a solution than calculating a solution” (135). This provides a strong reminder of Freud’s model of the mental apparatus where the neurons are perpetually striving to divest themselves of energy. Distributed systems obey that principle of thermodynamics that maintains that the state of lowest energy is the most stable. A final characteristic of distributed nets is that it is an extremely general form of “calculator”. Virtually any computing device, serial or parallel can be described in terms of the basic model of a neural network (74), including a Turing machine (119). There is therefore a mathematical argument for saying that a distributed system is always more general than a rule-based one. The kinship between neural networks and post-structural models should be clear: no distinction between levels, no overarching algorithm, but everything in terms of relations. Relations not between positive entities, but always only relations of relations or traces. Practical implementations of neural networks are at present of course still of limited size and complexity. They are therefore not really open systems yet, but framed systems that can be given a complete, formal description, and can be simulated on conventional computers. The mechanisms, by which they operate, however, are certainly not conventional, and their potential capabilities transcend all conventional computing techniques. Although they may in practice still have a closer resemblance to the structuralist descriptions of Saussure than to post-structural descriptions, their affinity with the spirit of post-structuralism is perhaps best exemplified by their non-algorithmic nature. There is no Programmer, no Scientist that can uncover the full Truth and the final significance of each element. There is, and was, always only the relationship of traces. To follow these developments without simply reclaiming them for the logocentric tradition, there lies a challenge.

References Anderson, J.A. & Rosenfeld, E. (eds.). 1988. Neurocomputing. Foundations of research. Cambridge, Massachusetts: MIT Press. Arbib, M.A. 1987. Brains, machines and mathematics. (Second edition). New York: Springer Verlag.

38 

 Paul Cilliers

Chomsky, N.A. 1957. Syntactic structures. The Hague: Mouton. Derrida, J. 1976. Of grammatology. Baltimore: The Johns Hopkins University Press. Derrida, J. 1978. Writing and difference. Chicago: Chicago University Press. Derrida, J. 1981. Positions. Chicago: Chicago Univ. Press. Derrida, J. 1982. Margins of philosophy. Sussex: Harvester Press. De Saussure, F. 1974. Course in general linguistics. London: Fontana. Freud, S. 1925. A note on the ‘Mystic writing-pad’. Standard Edition, vol. 19, 225–232. Freud, S. 1950 [1895]. Project for a scientific psychology. Standard Edition, vol. 1, 281–397. Grossberg, S. (ed.). 1988. Neural networks and natural intelligence. Cambridge, Massachusetts: MIT Press. Hampden-Turner, C. 1981. Maps of the mind. New York: Collier Books. Harland, R. 1987. Superstructuralism. The philosophy of structuralism and post-structuralism. London: Methuen. Haugeland, J. 1985. Artificial intelligence: The very idea. Cambridge, Massachusetts: MIT Press. Hillis, W.D. 1985. The connection machine. Cambridge, Massachusetts: MIT Press. Horgan, T. & Tienson, J. (eds.). 1987. Connectionism and the philosophy of mind. Supplement to Volume XXVI, The Southern Journal of Philosophy. Minsky, M. & Papert, S. 1969. Perceptrons. Cambridge, Massachusetts: MIT Press. Mountcastle, V.B. 1975. The view from within: Pathways to the study of perception. The John Hopkins Medical Journal, 136: 109–131. Mountcastle, V.B. 1978. An organizing principle for cerebral functions: The unit module and the distributed system, in G.M. Edelmann & V.B. Mountcastle (eds.). The mindful brain. Cambridge, Massachusetts: MIT Press, 7–50. Palm, G. 1982. Neural assemblies. An alternative approach to artificial intelligence. Berlin: Springer Verlag. Pinker, S. & Mehler, J. (eds.). 1988. Connections and symbols. Cambridge, Massachusetts: MIT Press. Pribram, K.H. 1971. Languages of the brain: Experimental paradoxes and principles in neuropsychology. Englewood Cliffs, New Jersey: Prentice-Hall Inc. Pribram, K.H. 1979. Holographic memory. Psychology today, February: 71–84. Pribram, K.H. & Gill, M.M. 1976. Freud’s project reassessed. London: Hutchinson. Rumelhart, D.E. & McClelland, J.L. 1986. Parallel distributed processing. Explorations in the microstructure of cognition. (2 volumes). Cambridge, Massachusetts: MIT Press. Stevens, C.F. 1979. The neuron. Scientific American, 241(3): 49–59. Turing, A. 1936. On computable numbers, with an application to the “Entscheidungsproblem”. Proceedings of the London mathematical society, Series 2(42): 230–265. Wilden, A. 1980. System and structure. Essays in communication and exchange (Second Edition). London: Tavistock Publications Ltd.

Paul Cilliers

Rules and relations Some connectionist implications for cognitive science and language In a previous article (Cilliers 1990) a model of the brain was developed that could be called “post-structural”. It was achieved by relating the model of the brain developed by the early Freud to Saussure’s structural linguistics. This relationship was amplified by examining Derrida’s reading of both Freud and Saussure, and it was found that words like “trace” and “différance” can be fruitfully employed in descriptions of the brain (Cilliers 1990: 4, 5). In the concluding section of the article it was shown that such a model of the brain can be linked to a recent technological development called neural networks (in an engineering context) or connectionism (in the context of cognitive science). Because connectionism is causing a disturbance in cognitive science that some (Horgan & Tienson 1987: 97) would interpret as signs of a Kuhnian crisis, it warrants more detailed attention. The wider philosophical implications of connectionism is underscored by the challenge it provides to some of the basic assumptions of artificial intelligence (AI) research, and in general to our understanding of the relationships between brain, mind and language. In this article the aim is to provide a description of the basic characteristics of connectionist models in order to fit them into a more general philosophical framework, and to facilitate future discussions and critiques. A connectionist system that models certain linguistic capabilities is described as an example of their practical capabilities. The central part of the article examines an important critique of connectionism from a non-structural perspective (Fodor & Pylyshyn 1988), something that was not done in the previous article. The main concern here will be the role played by rules in linguistic and mental activities. Because post-structuralism does not form part of the connectionist debate yet, a debate towards which it can certainly contribute important insights, I will briefly revisit some points of interaction between the two.

Originally published in the South African Journal of Philosophy, 1991, 10(2): 49–55. © South African Journal of Philosophy.

40 

 Paul Cilliers

1 Theoretical background Connectionism is a theory of mental activity inspired by models based on the functioning of the brain. Functionally the nervous system consists only of neurons, cells richly interconnected by means of synapses. The synapses convey stimulation generated in a previous neuron to the dendrites of the next neuron in line. If this stimulation exceeds a certain threshold, the neuron is triggered and an impulse is sent down the axon of the neuron. This impulse in turn provides the synaptic input to a number of other neurons. This input is modulated by the transfer characteristics of the synaps, as well as the structure of the dendrites. Complex patterns of neural excitation seem to be the basic feature of brain activity. Figure 1 indicates how a simple model of a neuron can be constructed. A neural unit uses the sum of its inputs to decide what output to generate. Each input is however first multiplied with a certain value or “weight”, W. This weight determines the connection strength between two specific units and models the characteristics of the synapses in the nervous system. The output of one neuron thus becomes one of the inputs to the next, after it has been modulated by the value of the weight in that pathway. Neurons are interconnected in large networks with complex connection patterns, and since the weights determine the influence of one neuron on another, the characteristics of a network is determined by the values of these weights.1 When a number of neurons are interconnected in a network, each neuron is continuously calculating its output in parallel with all the others, and patterns of activity, determined by the values of the weights, flow through the network. The topology of the network, that is, the way in which the neurons are interconnected, is also important. A network can be sparsely connected, richly connected or fully connected. A fully interconnected network is one where every neuron is connected to every other neuron.

1 An equation for the output of a specific neuron can be given:  

Ø = f  ( Σ ᷠ   On  Wn  )

  with   n   Ø   On   Wn   f

1

the amount of neurons this one is connected from the output of this neuron the output of the nth previous neuron the weight associated with the nth previous neuron the transfer function of the neuron



Rules and relations 

w₁ w₂

 41

1

ς

2

f

m

wn Fig. 1: A simple neuron

In

Layer 1

Out

Layer 2

Fig. 2: A simple neural network

In the network portrayed in Figure 2, information flows only from the input side to the output side, and the neurons are arranged in layers. Each neuron is connected from every neuron in the previous layer. There are no connections between neurons in the same layer, and there are no feedback loops. The activity of any specific neuron is influenced by many others, and it in turn has an effect on many others. Information is therefore not localised in any place in the network, but distributed over a large amount of units. The characteristics of the network are determined by the two layers of weights. The next thing to do is to show how such a network, even though it is simple, can process information. If the neurons in the input layer are activated in a certain way, a certain pattern will be generated by the output layer as the input values are multiplied through the two layers of weights. The input and the output can of course mean something. In the example we will look at later, the input represents the present tense of English verbs, and the output their past tense. If the network is presented with a verb, the pattern of weights will generate an output that represents its past tense. The crucial question to ask now is: Where do these weights come from? They could be predetermined by the designer of the network, but even if that was made possible in the case of a simple problem and a simple network, it is not necessary. By providing

42 

 Paul Cilliers

the network with enough examples, it will generate these weights by itself. A neural network is trained. It “evolves” in the direction of a solution. How is this done? By presenting the network with both the input pattern and the correct output, it can be made to adjust its weights in such a way as to match the two patterns closer. If enough examples are provided for long enough, and the network is complex enough to cope with the problem, a set of weights will be generated automatically that will produce the appropriate output for each input. After the network has been trained, it will not only recognise those examples it has been taught, but will take an educated guess at any previously unseen examples as well. Let me summarise the basic theory of connectionist models by pointing to a few crucial features. A neural network consists of large numbers of simple neurons that are richly interconnected. The weights associated with the connections between neurons determine the characteristics of the network. During a training period, the network can generate these weights automatically. Any specific weight has no significance, it is the patterns of weights in the whole system that bear information. Since these patterns are complex, and are generated by the network itself (by means of a general learning strategy applicable to the whole network), there is no algorithm available to describe the process used by the network to solve the problem. There are only complex patterns of relationships. To make the philosophical importance of all this rather dry information a little more apparent, I will try to place connectionism in a more general theoretical framework by identifying two paradigms. They will both stand under the name of a famous linguist  – the one Noam Chomsky, the other Ferdinand de Saussure. Such a dichotomy is always somewhat artificial, and I do not want to imply that one is Right and the other Wrong, but there are differences that make a difference. What is more, the link between connectionism and contemporary continental theory has to my knowledge not been made in the general debate of the subject, certainly not in the context of cognitive science. The two paradigms are summarised in Table 1. The way the Chomskian paradigm sticks together is well known: The formal grammars of production rules Chomsky employs to describe language are identical to the mathematical models of Alan Turing. These models, known as Turing machines2 also provide the mathematical foundation for the description of

2 The concept of Turing machines can be misleading. They are abstract mathematical entities that provide formalisations with which to describe computational procedures, that is, processes that follow an algorithm in discrete steps, including those implemented on digital computers



Rules and relations 

 43

digital computers. Hilary Putnam proposed that the Turing machine is an adequate model for the brain [a theory that has become known as functionalism, and which Putnam (1988) no longer supports], and Jerry Fodor extended Chomsky’s rationalist programme by linking the rule-based system of language with the functional working of the mind, irrespective of the hardware in which it is implemented. This is the paradigm that formed the basis for cognitive science and artificial intelligence (AI) research. It is only natural that under these circumstances intelligent behaviour should be described as rule-following, and hence we have the rule-based AI simulations known as expert systems. The paradigm standing under the name of Saussure is not so well structured. It can hardly be said to exist outside these descriptions, because cognitive science in general and connectionism in particular is practised in a context where Saussurian principles are unfamiliar. Part of the aim of this article is to establish the sensibility of such a paradigm. There are relationships between Saussure’s description of language (where each component of a language only has meaning in terms of the degrees to which it differs from all the other components of a language), and the way the brain works, especially in the way Saussurian theory was elaborated and criticised by post­structural thinkers like Jacques Derrida (see Cilliers 1990). To try and use neural network theory to provide a mathematical description of Saussurian linguistics, equivalent to the way in which formal grammars provide the mathematical description of Chomskian linguistics, could be part of a future project. In this article I am primarily arguing for the importance of connectionism as a development that has powerful and interesting implications for a large range of disciplines. I will also attempt to show that there are important differences between connectionist and conventional models, and how these differences hinge on the role and status of the rule.

with a van Neumann architecture. The fact that Turing describes these “machines” in terms of a “tape head” that reads and prints symbols on a (sometimes infinitely long) “tape” does not make them “real”. For a more detailed discussion of Turing machines and their relationships to computers see Haugeland (1985: 133–140) or Penrose (1989: 40–86).

44 

 Paul Cilliers

Table 1: Two linguistic paradigms Chomsky

Saussure

works with systems of production rules

works with systems of relationships

the functioning of the mind is coupled to language (Fodor)

the functioning of the brain and of language is mirrored in each other (Freud, Lacan, Derrida)

mathematical model supplied by Turing machine

mathematical model supplied by connection machine

machine produced: digital computer

machine produced: neural network

“intelligent” application: expert systems

“intelligent” application: pattern recognition

2 Rules or no rules As part of their exciting connectionist research programme, David Rumelhart and James McClelland developed a neural network that generates the past tense of the present tense English verbs it is presented with (Rumelhart & McClelland 1986, vol. 2: 216–271). They did not develop a large scale linguistic system, but chose an area of language that was confirmed enough to be manageable, yet rich enough to allow them to argue their position. The generation of past-tense forms is usually described by a fair number of rules, and also include a fair number of irregularities. The network they employed was a simple feed forward network similar to the one described above. The input and output layers each consisted of 460 neurons (239). During the learning phase the network was presented with phonological representations of the present tense of English verbs at the input, and the representation of the past tense at the output. At each presentation the weights of the network were adjusted using a process known as the “perceptron convergence procedure” (225). The output generated by the network for each input was constantly monitored throughout the learning phase. It was found that the network captured most of the features of both regular and irregular verbs in the same collection of weights, and that the network could respond appropriately to verbs it had not encountered before. Furthermore, during the training phase the network performed in ways similar to the way in which children acquire the past tense (219, 240). A child at first knows only the small number of past-tense verbs that is often used. Most of these are irregular, but are used correctly. In a second phase certain patterns are noticed, and a process of over-regularisation takes place. More verbs are used, but irregular verbs previously used correctly are now reg-



Rules and relations 

 45

ularised. In a third phase the differences are noticed, and regular and irregular forms are allowed to co-exist. These three phases were mimicked by Rumelhart and McClelland’s network. This performance of a network that only employs patterns of association led them to the following conclusion: We have, we believe, provided a distinct alternative to the view that children learn the rules of English past­tense formation in any explicit sense. We have shown that a reasonable account of the acquisition of past tense can be provided without recourse to the notion of a ‘rule’ as anything more than a description of the language (267).

That is, even though rules may be useful to describe linguistic phenomena, explicit rules need not be employed either when language is acquired or used. These unorthodox claims were sure to generate a response from the Chomskian camp. Let us take a look at how Jerry Fodor and Zenon Pylyshyn dismissed connectionism as a model for cognition and language. The context in which Fodor and Pylyshyn (1988) enters the discussion is one where connection systems are contrasted with symbol systems. In a symbol system information is represented by discrete symbols that are structured and manipulated by means of rules. The nature of a symbol system is determined by that which the symbols represent, and not by the way in which the system is implemented. If a biological system and a digital computer manipulate the same symbols in the same way, they are functionally equivalent. Fodor and Pylyshyn identifies two major traditions in “modern theorizing about the mind”: one they call “Representationalist” and the other “Eliminativist”. Representationalists hold that postulating representational (or ‘intentional’ or ‘semantic’) states is essential to a theory of cognition; according to Representationalists, there are states of the mind which function to encode states of the world. Eliminativists, by contrast, think that psychological theories can dispense with such semantic notions as representation. According to Eliminativists the appropriate vocabulary for psychological theorizing is neurological or, perhaps behavioural, or perhaps syntactic; in any event, not a vocabulary that characterizes mental states in terms of what they represent (Fodor & Pylyshyn 1988: 7).

Strangely enough, they then claim that “connectionist modelling is consistently Representationalist” (8), and therefore on the same side of the divide as “classical” cognitive theory, the position they defend. The only difference between them, on this issue, is that classicists assign semantic content to “expressions”, and that connectionists assign it to nodes in a network. They therefore claim that “representation” is not part of the dispute. What then are the issues? This is how they summarise it:

46 

 Paul Cilliers

Classical and Connectionist theories disagree about the nature of mental representation; for the former, but not for the latter, mental representations characteristically exhibit a combinatorial constituent structure and a combinatorial semantics. Classical and Connectionist theories also disagree about the nature of mental processes; for the former, but not for the latter, mental processes are characteristically sensitive to the combinatorial structure of the representations on which they operate (32).

What do these two differences amount to? A representational system works with symbols. For symbols to be meaningful, they firstly have to be structured, and it must secondly be possible to manipulate them. Unrelated collections of symbols are meaningless. Connectionist systems are representational, but because they merely “associate” representations, and have no rules to manipulate them with, they cannot model the mind. [...] since the Connectionist architecture recognizes no combinatorial structure in mental representations, gaps in cognitive competence should proliferate arbitrarily. It’s not just that you’d expect to get them from time to time; it’s that, on the ‘no-structure’ story, gaps are the unmarked case. It’s the systematic competence that the theory is required to treat as an embarrassment But, as a matter of fact, inferential competences are blatantly systematic. So there must be something deeply wrong with Connectionist architecture. What’s deeply wrong with Connectionist architecture is this: Because it acknowledges neither syntactic nor semantic structure in mental representations, it perforce treats them not as a generated set but as a list. But lists, qua lists, have no structure; any collection of items is a possible list. And, correspondingly, on Connectionist principles, any collection of (causally connected) representational states is a possible mind. So, as far as Connectionist architecture is concerned, there is nothing to prevent minds that are arbitrarily unsystematic. But that result is preposterous. Cognitive capacities come in structurally related clusters; their systematicity pervasive (49).

It is interesting to note that several of the objections stated here can be interpreted extremely positively from a post­modern or post-structural perspective (unstructured lists, “gaps are the unmarked case”, minds that are arbitrarily unsystematic), but let us remain on the cognitive science territory for the time being. Their objections certainly hold water if one is committed to a system built up out of atomic representations. They are however, gravely mistaken in thinking that connectionist systems fit that description. By insisting that information is represented in the nodes of a network (12), they miss the real significance of the distributed representations in neural networks, a point also made by Smolensky (1987: 137), Derthick (1990: 255, 257) and Bechtel (1987: 22). The example Fodor and Pylyshyn choose to demonstrate a connectionist network with, is not a neural network at all. It is a semantic network, a well-known technique from traditional AI that merely depicts a few relationships



Rules and relations 

 47

between a number of atomic representations.3 They acknowledge the existence of distributed representations, but for them it is merely local representations “sliced thinner”. The point is: in a full-blown neural network no node has any specific significance. As I have explained earlier, the significance lies in the values of the weights. Not, and this is crucial, in the value of any specific weight or even group of weights, but in the way they are related and activated each time. Information is not stored in a symbol and recalled when necessary, as in traditional cognitive models, it is reconstructed each time that that part of the network is activated (Bechtel 1987:22). Smolensky (1987) would argue that distributed representation is representation of some form, but so different to the Classical models, that they cannot be placed under the same heading. The differences between the two systems are far more radical than Fodor and Pylyshyn would admit, and the argument has to be taken up on a different level. In terms of the limited networks currently used to demonstrate connectionist principles or to solve specific problems, I think Smolensky is correct. They address specific domains, in other words, they are well framed. As models they do represent their problem areas, even if it is in some unconventional way. When we move to large systems where the frame is not apparent, I am convinced that the concept of mental representation has to be dropped altogether. Objects in the world are not represented in the brain, just as a word in a natural language does not represent a specific meaning.4 Although there are many issues that need further clarification, one can already pose the question whether connectionist systems and rule-based systems exclude each other. Fodor and Pylyshyn (1988: 64) acknowledge the biological plausibility of connectionist architectures, and the fact that neural nets can be used to implement a Turing machine. They suggest this as a possible option, and

3 Their failure to distinguish between a semantic net and a distributed representation in a neural net also results in their equating of connectionism with associationism (Fodor & Pylyshyn 1988: 31). A connectionist network does not encode a number of relationships between specific ideas, because specific nodes in the network does not correspond to specific ideas, as they do in a semantic network. Some of this confusion is a result of connectionist terminology. Connectionists tend to refer to the relationships in a network as “sub-symbolic” or as “micro-features”, as if a group of micro-features lumped together will add up to one whole feature or symbol. A better understanding of what a weight is emerges when it is compared with Derrida’s notion of a “trace” (see below). 4 The problems surrounding the concept of “mental representation” deserves a separate, more detailed study with reference to recent developments in analytic philosophy (e.g. Putnam 1988) and continental theory. In a certain sense the meaning of the word “representation” breaks down if used in the context of a “distributed representation”.

48 

 Paul Cilliers

then urge connectionists to direct their research at coming up with good implementations of “classical” architectures.5 This strategy, however, denies any significance to the important differences between local representations and fully distributed systems that have capabilities beyond that of “classical” approaches. Smolensky (1987: 137–143) suggests five possible ways of dealing with the “soft” connectionist option on the one hand, and the “hard” symbol system on the other: 1. Deny one and continue only with the other. The denial of the “soft” is also known as “rationalism”. Denial of the “hard” leads (according to Smolensky) to the “intuitive” approaches of for example Dreyfus and Dreyfus (1986). 2. Allow the two systems to “cohabitate” as two separate “processors” next to each other. In a way this is an extreme formulation of the split brain theory. 3. Soften the “hard” approach by means of fuzzy logic. This merely blurs the edges, and softness becomes degrees of hardness. 4. Make the system which is “hard” at bottom complex enough that softness will emerge. This is a sophisticated approach, but the “brittleness” of expert systems with many rules remains discouraging. 5. Make a system which is “soft” at bottom complex enough that hardness will sometimes appear when viewed at a higher level. This last suggestion is certainly intriguing. It postulates a system that does not function on a basis of rules, but where certain systematic properties can be described by means of rules if they prove to be useful. I cannot see why this suggestion should not satisfy those who like to look for structures and patterns. They can make interesting and useful classifications without having to elevate them to Final Truths. I suspect however, that this will be a problematic position for the True Scientist for whom, like Fodor and Pylyshyn (1988: 64), “truth is more important than respectability”.

3 Connectionism and post-structuralism In many areas of science, where both theory and practice are concerned, there is a growing discontent with analytical and deterministic methods and descriptions. One of the first responses to this unease was a rapid growth in statistical

5 Such research has received serious attention among connectionists long before Fodor et al. suggested it. See for example Touretzky and Hinton (1985).



Rules and relations 

 49

approaches, not only in the interpretation of experiments, but in the explanation of the results as well. However, like fuzzy logic, statistical methods do not imply a break with deterministic methods. It is a way of blurring the edges, but a statistical analysis remains a tool in the process of establishing the true mechanisms of the phenomena being investigated. To think in terms of relationships, rather than in terms of deterministic rules, is not a novelty for science, but it has always been seen as a part of qualitative descriptions and not as part of the quantitative descriptions and calculations deemed necessary ever since Kepler’s insistence that “to measure is to know”. Many phenomena, especially in the life sciences, but also in physics and mathematics, simply cannot be understood properly in terms of deterministic, rulebased or statistical processes. Quantum-mechanical descriptions of sub­atomic processes are essentially relational, and even on a more macroscopic level, relations determine the nature of matter. The carbon atoms in my body can all be interchanged with carbon atoms from the wood of my desktop, and there will be no noticeable difference (Penrose 1989: 32). The significance of matter is therefore not determined by the nature of its basic constituents, but is a result of a large number of relationships between the constituents. A striking example of the importance of relationships comes from a fascinating new development in mathematics known as fractal geometry. Normal Euclidean descriptions are often quite useless in describing natural shapes like clouds, rivers, mountains, turbulent flow, etc. Nature does not often produce straight lines or smooth curves. To cope with this complexity, a new type of mathematical description is being developed, based mainly on the strikingly beautiful and original work of the French mathematician Benoit Mandelbrot (1983). Another prominent figure, Michael Barnsley, describes it in the following way: In deterministic geometry, structures are defined, communicated, and analysed, with the aid of elementary transformations such as affine transformations, scalings, rotations, and congruences. A fractal set generally contains infinitely many points whose organisation is so complicated that it is not possible to describe the set by specifying directly where each point in it lies. Instead, the set may be defined by ‘the relations between the pieces’. It is rather like describing the solar system by quoting the law of gravitation and stating the initial conditions. Everything follows from that. It appears always to be better to describe in terms of relationships (Barnsley 1988: 5).

In the light of these examples, it is certainly strange that when it comes to descriptions of the functioning of the brain, an obviously relational structure, there is still such a strong adherence to deterministic algorithms. One of the reasons for this must surely be that cognitive science inherited its methodological framework

50 

 Paul Cilliers

from a deterministic, analytical tradition. This is in serious need of revision, and in this revision post-structural theory will have to play an important role. The influence of post-structuralism on cognitive science could operate on two levels. On a methodological level, a post-structural approach could affirm the non-algorithmic nature of cognition. It could help to suppress the desire to find complete and deterministic models by arguing that relationary models operating in a wider context (a boundless context, as a matter of fact) are less restrictive, and just as useful. It can also help to legitimise activities that do not aspire to fill in the “big picture”, but only to be of local use. On a more practical level, there seems to be striking similarities between connectionism and Derrida’s model of language. A detailed description of Derrida’s critique and elaboration of Saussarian linguistics, including analyses of the concepts of “trace” and “différance”, has been provided in a previous article (Cilliers 1990). These descriptions were expanded to include models of the brain and will not be repeated here, except for one representative quote:6 Whether in written or in spoken discourse, no element can function as a sign without relating to another element which itself is not simply present. This linkage means that each ‘element’ – phoneme or grapheme – is constituted with reference to the trace in it of the other elements of the sequence or system [...] Nothing, either in the elements or in the system, is anywhere simply present or absent. There are only, everywhere, differences and traces of traces (Derrida 1981: 26).

If one can point out some interaction between the concepts of connectionism and post-structuralism, it will not only enrich the context of connectionism, but also provide some substance to post-structural concepts. The best way to understand the significance of a “weight” in a neural network is to compare it with Derrida’s concept of the “trace”, the term Derrida uses to point to the influence that each component in the system of language has on every other component. Because of the “distributed” nature of these relationships, a specific trace has no ideational content, but only gains significance in large patterns of interaction. Similarly, Derrida’s concept of différance can be used to describe the dynamics of complex neural networks. The analogy works in the following way: If an ensemble of neurons (whether real or artificial) generates a pattern of activity, traces of these activities reverberate through the network. If there are loops in the network, these traces will be reflected back after a certain

6 Derrida discusses Saussure at length in Of grammatology (Derrida 1976) and Freud’s model of the brain in Freud and the scene of writing (Derrida 1978: 196–231). The essay “Différance” appears in Margins of philosophy (Derrida 1982: 1–27).



Rules and relations 

 51

propagation delay, and will alter (make different) the activity that produced it in the first place. As soon as a complex system contains loops and feedback,7 delayed self-altering will be one of its characteristics. This characteristic has much in common with Derrida’s concept of différance – a concept that indicates difference, deference, is suspended between the passive and active modes and that has both spatial and temporal components (Derrida 1982: 1–27). According to the post-structural logic of “trace” and “différance” no word in language (or neuron in the brain) has any significance by itself. Meaning is determined by the dynamic relationships between the components of the system. In the same way, no node in a neural network has any significance by itself – that is the central implication of a distributed representation. Significance is derived from patterns of activity involving many units, patterns that result from a dynamic interaction between large numbers of weights. If the link between “weights” and “traces” holds, it will also give more substance to post-structural models that are generally seen to be fairly abstract, despite Derrida’s claim for their “materiality” (Derrida 1981: 64). In a distributed representation a weight plays the same role as a “trace”, but there is nothing mystic, strange or abstract about it. Weights operate at such a low level of significance that they cannot be given a symbolic, or even sub-symbolic; interpretation, yet they have a specific, finite value. Bearing in mind their dynamic character (they change – at least during learning), the weights in a connectionist network can be seen as a “material” example of traces. Turning to practical neural networks being used at present to solve problems, mainly in the field of pattern recognition, a number of qualifications have to be made. Practical networks are generally designed to perform specific tasks. They have a limited number of neurons, usually a limited pattern of interconnections and limits are also placed on the values of the weights and the transfer functions of the neurons. Furthermore, a network is designed for a specific problem, and the weights are usually only altered during a learning phase. They are, generally speaking, not flexible enough to address wide-ranging problems. In this sense, neural networks are structural rather than post-structural, and can be described quite adequately, in Saussurian terms. Post-structural concepts do become important, however, if we want to stretch the capabilities of present networks, especially in the context of AI. Networks that mimic human behaviour better will have to be much more flexible. They will have to be creative (implying the generation of new meaning, not following old rules) and they will have to be able to move

7 Complex systems of feedback loops do exist in the cortex (see Edelman 1978) and are also employed in certain neural network algorithms, e.g. adaptive resonance theory (Grossberg 1988).

52 

 Paul Cilliers

beyond predetermined limits. Towards these ends, the energetic and penetrating ideas of post-structuralism may contribute enormously. If we return to the examples at the beginning of this section, it should be apparent that post-structuralism also has a lot of significance for many other scientific endeavours in a postmodern context. One of the shifts that have taken place in science as a result of the unease with analytical paradigms, is to incorporate (or attempt to incorporate) large chunks of eastern philosophy.8 Although there is nothing wrong with this, there is an inevitable flavour of mysticism associated with these strategies that will always place it at a disadvantage where the scientific mainstream is concerned. Similarly, post­structural theory has been made suspect in the way it was championed by literary theorists under the name of “deconstruction”. By emphasising the inherent subversive qualities of these theories, they created the impression that deconstruction is essentially anti-scientific. Derrida himself, though, is certainly not opposed to science as such. He is opposed to the metaphysical components that are dragged into science by elevating “logos” to a transcendental status. He also does not hesitate to call grammatology a “science” (Derrida 1976: 3, 4). Adopting a post-structural perspective on science will certainly be in conflict with much of what is accepted as canonical theory of science, but may have less radical effects on the practice of science than one expects. Unless one would want to call the opening up of new space for creative thought something radical.

References Barnsley, M. 1988. Fractals everywhere. Boston: Academica Press. Bechtel, W. 1987. Connectionism and philosophy of mind: An overview. In: The Southern Journal of Philosophy, XXVI (Supplement): 17–41. Cilliers, F.P. 1990. The brain, the mental apparatus and the text: A post-structural neuropsychology. In: South African Journal of Philosophy, 9(1): 1–8. Derrida, J. 1976. Of grammatology. Baltimore: The Johns Hopkins University Press. Derrida, J. 1978. Writing and difference. Chicago: Chicago University Press. Derrida, J. 1981. Positions. Chicago: Chicago University Press. Derrida, J. 1982. Margins of philosophy. Sussex: Harvester Press. Derthick, M. 1990. Review of Pinker, S. & Mehler, J. 1988. Connections and symbols (MIT Press, Cambridge, MA). In: Artificial Intelligence, 43: 251–265. Dreyfus, H.L. & Dreyfus, S.E. 1986. Mind over machine. The power of human intuition and expertise in the era of the computer. New York: The Free Press.

8 Examples would be the works of Fritjof Capra and Gary Zukav. They should not be confused with a serious scientist like David Bohm.



Rules and relations 

 53

Edelman, G.M. 1978. Group selection and phasic reentrant signaling: A theory of higher brain function. In: Edelman, G.M. & Mountcastle, V.B. The mindful brain. Cambridge, Massachusetts: MIT Press, 51–100. Fodor, J.A. & Pylyshyn, Z.W. 1988. Connectionism and cognitive architecture: A critical analysis. In: S. Pinker, S. & Mehler, J. (eds.). Connections and symbols. Cambridge, Massachusetts: MIT Press, 3–71. Grossberg, S. (ed.). 1987. The adaptive brain (2 volumes). Amsterdam: North Holland. Grossberg, S. (ed.). 1988. Neural networks and natural intelligence. Cambridge, Massachusetts: MIT Press. Haugeland, J. 1985. Artificial intelligence: The very idea. Cambridge, Massachusetts: MIT Press. Horgan, T. & Tienson, J. 1987. Settling into a new paradigm. In: The Southern Journal of Philosophy, XXVI (Supplement): 97–113. Mandelbrot, B.B. 1983. The fractal geometry of nature. New York: W.H. Freeman. Penrose, R. 1989. The emperor’s new mind. Oxford: Oxford University Press. Putnam, H. 1988. Representation and reality. Cambridge, Massachusetts: MIT Press. Rumelhart, D.E. & McClelland, J.L. 1986. Parallel distributed processing. Explorations in the Microstructure of cognition (2 volumes). Cambridge, Massachusetts: MIT Press. Smolensky, P. 1987. The constituent structure of connectionist mental states: A reply to Fodor and Pylyshyn. In: The Southern Journal of Philosophy, XXVI (Supplement): 137–161. Touretzky, D.S. & Hinton, G.E. 1985. Symbols among the Neurons: Details of a connectionist inference architecture. In: Proceedings of the 9th International Joint Conference on Al. Los Angeles, CA., 238–243.

Paul Cilliers

Rules and complex systems A central philosophical problem, one that has concerned scientists as much as philosophers, is the relationship between our descriptions of the world and the world itself. This problem is present in one way or another in many different theoretical discourses: in discussions of the status of models and theories in science (instrumentalism, reductionism, realism, etc.), in theories of representation, theories of meaning, and in the realm of law and ethics.1 I do not want to propose a final solution to this problem, but in order to clarify the issue, I want to analyse one of the central notions in most of the discourses mentioned, namely, that of rules. More specifically, I want to investigate the use and the status of rules when we deal with complex phenomena like the brain, language or social and cultural systems.2 In an extended analysis of complex systems (Cilliers 1998), I argue that a rulebased approach is not adequate when we want to model complex systems. The argument employs post-structural perspectives (mainly those of Derrida) in order to show that the intricate and dynamic network of relationships between the components of a complex system can be understood better in terms of connectionist (or neural network) models. The practical limitations of current neural networks, specifically the severe limitations of feedforward networks, are acknowledged, but nevertheless it is proposed that distributed systems with recurrency3 can serve as general models for complex systems. I will not repeat all those arguments here; however, I do want to address an (incorrect) impression that may arise from them: that rules are not important or useful. As a matter of fact, we cannot do without them, but we need to analyse more closely what their function and status could be. It is also necessary to investigate how descriptions based on rules and

1 ln the case of law and ethics, this discussion usually concerns the grounding of ethics. i.e., whether we can find an objective basis to justify ethical positions. Arguments include the problem of natural law, the naturalistic fallacy as well as perspectives from evolutionary ethics. See Singer (1994). 2 I use the notion of “system” in an open sense. It does not refer to systems theory or cybernetics in any direct way. For a characterisation of “complex systems”, see Cilliers (1998: 2–7). 3 A recurrent neural network is one in which the information does not flow in only one direction through the network. Recurrent nets are highly, or even fully, connected, i.e., each neuron is connected to many (or all) of the other neurons. Consequently, there are lots of feedback paths. These networks are notoriously difficult to train. Originally published in Emergence, 2000, 2(3): 40–50. © 2000 Lawrence Erlbaum Associates, Inc.

56 

 Paul Cilliers

descriptions based on relationship relate to each other. I will start with a firstlevel consideration of this last question, and return to it in more detail later.

1 Complex and complicated In order to demonstrate the difference between a rule-based and a relationship-based description, one can make use of two different approaches to language. On the one hand, there is the attempt to model language in terms of a formal system of rules. This approach is exemplified best by the early work of Chomsky. On the other hand, there is a description of language in terms of a system of differences, best exemplified by Saussure. Although any sophisticated theory of language will not be limited to one of these, the distinction remains useful. The relationship-based approach attempts to deal with patterns within the system as a whole,4 while the rule-based approach attempts to find fundamental units in the system, and identify the rules that combine and transform them. This distinction remains central to many disputes in cognitive science, as well as between different philosophical schools.5 A similar kind of distinction is made in complexity theory between things that are “complicated” and things that are “complex”. Something that is complicated can have many components, and can be quite intricate, but the relationships between the components are fixed and clearly defined. We can use the analytic method to analyse complicated things, i.e., we can take them apart and put them back together again, like a jumbo jet. Something that is complex, on the other hand, is constituted through a large number of dynamic, nonlinear interactions. Therefore the important characteristics of a complex system are destroyed when it is taken apart, i.e., when the relationships between components are broken. Living things, language, cultural and social systems are all complex. The behaviour of complicated things can be described by rules; the behaviour of complex systems is constituted through relationships. This distinction can also be used to make another point: complex things have emergent properties, complicated things do not. Emergent properties are those we cannot predict merely by analysing the components of the system. Consciousness is an emergent property of the brain that cannot be predicted by examining

4 This is difficult of course, if the boundaries of the (linguistic) system are not well defined. I take this to be an important part of Derrida’s critique of Saussure (Derrida, 1976). 5 I discuss this distinction in more detail elsewhere (Cilliers 1991, 1998:30–31).



Rules and complex systems 

 57

a neuron. The behaviour of complicated things, however, is predictable  – as it mostly should be. No one would fly in a jumbo jet with emergent properties. The notion of emergent properties is a problematic one for many. They feel that it refers either to something mystical, or to something we cannot, or do not want to, analyse. One can attempt to counter these anxieties by arguing that emergent properties have nothing to do with mysticism, that emergence does not involve metaphysical components, that we are merely talking about properties that arise because of non­linear, dynamic interactions of which there are so many that we cannot hope to contain all the information involved. According to this argument, “emergent properties” can be renamed to something like “relational properties”. Although I concur with this argument, there is a price to pay for it: it undermines the very distinction between complex and complicated. If we claim that complexity is not a metaphysical thing, that we can talk about complex systems in a materialist way, then we have to grant that the small-scale interactions in a complex system are such that basic physical principles, like causality, should hold. Although it may be impossible to describe these micro-interactions in practice, it should in principle be possible to describe them in terms of rules, and therefore, at bottom, everything is only complicated. To grant that the distinction between complex and complicated is an analytic one – useful in theory, but not always maintainable in practice – does not change our understanding of complex systems significantly. The more pressing issue here, however, is whether this implies that the distinction between relationships and rules is also dissolved. I wish to argue that it is not. In order to do that, we need to analyse the notion of a rule in more detail.

2 The status of rules The rule-based approach to complex systems is one we know well from mainstream artificial intelligence (AI) research in the last 30 years (see Pinker & Mehler 1988).It has recently been restated forcefully by John Holland in his book Emergence (1998).6 I will examine this work briefly in order to develop a more differentiated understanding of the notion of a rule.

6 He summarised his position in an article (Holland 1997). Although the publication date is earlier, it was written after completing the book.

58 

 Paul Cilliers

Holland develops ideas from mainstream AI, influenced by the work done at the Santa Fe Institute.7 He is quite clear about his intentions at the start of the book. He acknowledges that “emergent phenomena also occur in domains for which we presently [sic] have few accepted rules”, like ethics and the evolution of nations, but that he will restrict his study to “systems for which we have useful descriptions in terms of rules and laws” (1998: 3). Although he adheres for the most part to these intentions, one gets the impression that ultimately he thinks that all complex systems are underpinned by a precise set of rules. Indeed, he ends the book talking about “life” and “consciousness” (1998: 246–8).The approach to complexity he suggests is a familiar one, based on formal models. Such models consist of atomistic building blocks (1998: 24–6) whose interactions are determined by a set of formal production rules. The components of the model represent elements of reality on a one-to-one basis, i.e., each component of the model stands for a specific element of reality (1998: 29). It is usually clear that he views these rule-based models as descriptions of reality, although he does sometimes talk about rule-governed systems (1998: 6), which creates the impression that rules are somehow more fundamental to the working of the systems he describes. He claims: A well-conceived model exhibits the complexity and emergent phenomena, of the system being modeled, without the obscuring effects of incidentals (Holland 1997: 17).

His favourite examples are, predictably, board games like chess and checkers. Although Holland is correct in arguing that models have to reduce the complexity of the phenomenon being modelled, he does not acknowledge that this is exactly the reason that we cannot have exact rule-based models of complex systems. Because of the nonlinearity of the interactions constituting a complex system, it cannot be “compressed”. Any simplifying model will have to leave out something, and because of the non­linearity, we cannot predict the significance of what is suppressed. In order to capture all the complexity, we will have to “repeat” the system in its entirety. This is just as problematic. Since complex systems inter-

7 In many respects the book is somewhat anachronistic. The feeling that much of it could just as well have been written in the 1960s is confirmed by looking at his bibliography. The subtitle of the book “From Chaos to Order”, is a serious misnomer for a book that argues that formal systems of rules – something that can hardly be seen as chaotic – can have emergent properties. Chaos, in either the technical or the colloquial sense, receives virtually no attention in the book whatsoever. It is not my intention to berate the book: there is much of interest and importance in it, not the least of which is Holland’s insistence on the importance of art and metaphor when trying to describe the world.



Rules and complex systems 

 59

act with their environment in intricate ways, it is never obvious where the limits of the system are. When we deal with complexity, we cannot avoid framing our description thereof in some way or another. Models can, therefore, not function in an objective way, they have to be interpreted. I have criticised purely formal models in detail (Cilliers 1998: 13–15, 58–74), mainly because of the shortcomings of a traditional theory of representation. Something that can be understood fully in terms of a set of rules can, at best, be complicated. It is, therefore, my contention that the formal systems that Holland describes cannot show emergent behaviour. Chess may indeed have many possibilities that have not yet been realised, but all of these novelties can still be understood in terms of the basic, static, timeless rules of the game. This is not comparable to something like consciousness as an emergent property of a large number of neurons. Holland (1998: 81–114) does, however, also employ the example of neural networks. The connectionist models I suggest as an alternative to formal rule-based models are, according to Holland, also rule based at bottom. His argument is that the functioning of recurrent neural nets is based on the working of Hebb’s rule.8 This “rule” describes the local interaction between neurons that is responsible for the organisation of structure in a network. If Holland is correct, neural nets are also rule­based systems.9 This would be another way of collapsing the distinction between complex and complicated, and a vindication of his position that formal systems can have emergent properties. However, I think he is incorrect and that his mistake is at heart a semantic one: Hebb’s rule is not a rule. The logic of the notion “rule” implies a certain generality. A rule should apply “without exception to the cases subsumed by the description incorporated in the rule” (Winston, 1971: 177). In order for a rule to apply, it has to be established that the specific case at hand is the same as the general case described by the rule. A “good” rule is one that could apply to many cases. It must say something about the system it describes, or, in Holland’s terms, the model of the system must be simpler than the system it describes (Holland 1998: 24). We formulate a rule in order to describe a pattern that we have come to recognise, to formalise a regularity. Furthermore, when working with a system of rules, the different rules should

8 Hebb’s “rule” in its basic formulation, states that the connection strength between two neurons will increase if the two neurons are active simultaneously (Hebb 1949). A similar notion, what one can call the “use principle”, was employed by Freud in his early work on models of the nervous system (see Cilliers 1998:45–6). 9 It should come as no surprise that Holland (1998: 87–8), via the McCullogh-Pitts formalisms, associates neural networks with the Chomskian approach to language.

60 

 Paul Cilliers

be “articulated” properly, i.e., rules must be linked in such a way that the output of one rule can be used by a next rule as an input. This characteristic of a system of rules is described by the notion of an algorithm. Such an understanding of the notion “rule” is clearly at odds with a low level, nonalgorithmic principle like Hebb’s rule. This is a principle that applies locally between the components of a complex system. It describes a process that has no ideational content, it operates on contingent, low-level information, is not selective, and provides us with no general information about the system as a whole, or even parts of it. The same rule – perhaps it is better to start talking of something like a principle, rather than a rule  – operates everywhere in the system. More importantly, it operates in all recurrent neural systems, irrespective of what the function of that specific system is. In the brain, for example, the same principle can be used to organise the motor cortex and the visual cortex; it does not tell us anything about the differences between perception and action. In Derridean terms, we can say that Hebb’s rule functions on the level of the trace. There is no similarity between this and a rule that is supposed to capture some essential or general aspect of a system. To summarise, we can say that Hebb’s rule (or the use principle) and rules in a rule-based model of a system do not have the same status. The use principle describes a mechanism for organisation in complex systems. It does not tell us anything about the system. Rule-based models generate descriptions of what such systems do, and perhaps how they are supposed to do it. They are not “wrong” or useless, they are all we have when we want to develop an understanding of complex systems. We must merely be clear about the limitations of our models. In order to do this, we cannot just talk of rules in a blanket fashion, we need to differentiate between different kinds of rules.

3 Differentiating rules A distinction is often made between descriptive and prescriptive rules. This distinction relies on the existence of a clear differentiation between facts and values, one that is problematic to maintain,10 especially in the context of complex

10 Lyotard makes this distinction, and insists on the irreducibility of the one to the other (1986: 213). From this he concludes (mistakenly, I think) that there is a plethora of incommensurable “regimes of phrases” (1986: 218). For an example of the way in which the distinction between values and facts is problematised, see Derrida’s discussion of Austin (Derrida, 1988: 15).



Rules and complex systems 

 61

systems. Since I do not want to move into the normative domain directly,11 I will focus on another distinction, that between regulative and constitutive rules.12 Constitutive rules provide the framework within which one can understand a set of facts. The best way to describe these rules is in terms of the rules of a game. The game does not exist if the framework is not accepted. Regulative rules, on the other hand, are those rules that determine or constrain permissible moves within that framework. A regulative rule thus only makes sense in the context generated by the set of constitutive rules. To serve a double fault only makes sense when one is playing tennis. One is playing tennis only when those participating agree to submit themselves to the rules that constitute the game. There is often no compelling reason to do this other than to play the game. This distinction can, I believe, be extended to scientific models. In the Newtonian framework, for example, one would use the regulative rule that states “if a force is applied to a mass it will accelerate”. For this rule to make sense, the Newtonian framework must already be constituted. This may bear some relationship to a Kuhnian understanding of science, but my point is a little more specific: If one wants to generate a formal model of a complex system in terms of a set of rules, a model that can be simulated on a computer, the distinction between regulative and constitutive rule should be clearly understood. The nitty-gritty rules of the model, those that determine or constrain its behaviour, are regulative rules. One can spend a lot of time developing and refining them, but one should not forget that they only have meaning in terms of the framework constituting that meaning. When dealing with complicated things, the constitutive framework can normally be determined precisely, at least in principle. When modelling complex systems, however, the constitutive framework is not given, nor is it self-evident in a straightforward fashion. In order to generate some general understanding, the framework has to reduce the complexity. A framework is selected in terms of the aims of our description of the system. The quality and usefulness of the model are primarily determined by this selection. One should also bear in mind that although the constitutive framework may reduce complexity from a certain perspective, it may increase it from another perspective. This is especially the case, ironically perhaps, when we have developed

11 Since we cannot give a complete description of a complex system, we have to select the aspects we are going to consider. This selection cannot be based on calculation, it involves choice. Consequently, we cannot escape the domain of values. The relationship between complexity and ethics is explored in more detail elsewhere (Cilliers 2000). 12 The distinction between regulative and constitutive rules is well known. It is used by Kant, by speech-act theorists, and in theories of law and justice (see Rawls 1955). For an interesting discussion of the distinction, in a somewhat different context, see Reddiford (1985).

62 

 Paul Cilliers

a good framework, one that allows us to see things that were not apparent from previous models. Thus, our models can both conceal and reveal complexity. How we employ our models often has a lot to do with prevalent values in a discipline, something that, without too much exaggeration, we can call style. From the claim that the constitutive framework for a complex system is not naturally determinable, it may be inferred, incorrectly, that the framework is arbitrary. This mistake is exactly equivalent to the claim, also mistaken, that deconstruction implies relativism.13 However, at this point we are returned to the problem with which this article opened: What is the relationship between our descriptions, or models, of the world and the world itself?

4 What is described by a system of rules? A brief recapitulation of the claims made so far may be helpful. There is nothing mystical about the workings of a complex system. However, since the nature of the system is the result of countless, local, nonlinear, nonalgorithmic, dynamic interactions, it cannot be described completely and accurately in terms of a set of rules. These local interactions themselves can also not be understood in terms of the notion “rule”. Yet, we cannot avoid rules when trying to describe complex systems. We must, nevertheless, always bear in mind that these rules only make sense in terms of a framework that is not naturally given, but that is generated by the description itself. This precludes the possibility that any model of a complex system can be a perfect or exact one. We cannot avoid the reduction of complexity in the process of modelling. However, if we want to argue that models are merely linguistic entities, we are caught in an instrumentalist or relativist position, a position that cannot be coherently maintained. A possible answer begins to suggest itself when we move away from the foundationalist/relativist dichotomy. In order to do this, we have to acknowledge that the language we use to describe the world is not completely independent from it. There is two-way communication between the two, but since both are complex, there is also no perfect fit between them. There is no reason to conclude from this position that our models are arbitrary. The rules of our descriptions, I would argue, are attempts to say something about the structure of the patterns of interaction in a complex system. This structure is not rule following or rule governed in itself (in the same sense that a stone

13 Derrida dismisses these allegations in the important “Afterword” to Limited Inc. He also discusses the use and function of rules in several contexts (Derrida 1988).



Rules and complex systems 

 63

does not solve differentiate equations when tailing to the ground), nor is it timeless or fixed. A specific system of rules, especially if it is sensitive to the historical nature of complexity, may at times give a very accurate description of this structure. A specific structure may be fairly stable over time and contained within clear limits. The rules used to describe this may be so accurate that we are quite happy to say that they are “true”. Other structures may be much more open, volatile and interdependent, and therefore not amenable to a rule-based description. If we want to describe such structures, we have no choice but to impose limits on it in order to make the description possible. In the latter case, the distortion caused by the model will be more significant. Two problems compound the issue even further. The first is that patterns of all kinds are possible in a complex system, from the stable to the ephemeral and everything in between. We have no a priori method able to provide a guarantee that we are dealing with an aspect of the system that is stable, and likely to remain so. The second problem has to do with the algorithmic nature of a system of rules, i.e., that they have to articulate cleanly. The patterns of structure in a complex system are much messier. Their borders are not clearly defined. They overlap and interpenetrate each other. If a complex system is critically organised, it will also have structural components on all scales of magnitude, and will therefore be maximally sensitive to influences that can change its structure.14 These considerations lead me to conclude that rule-based models will not be able to provide general and accurate descriptions of complex systems, particularly not of things like human beings. The final question to consider, then, is whether connectionist models can do so.

5 What can we do with neural nets? Can an accurate model of a complex system be implemented in a neural net? For now, my answer is a qualified “no”. All networks used in practice have a specific function and clear limits. Although their internal logic does differ substantially from traditional formal systems, and despite the fact that they can solve certain pattern-recognition problems with greater ease than formal, rule-based models, ultimately they are well framed, receive specific inputs and produce specific

14 See Cilliers (1998: 96–8) and Bak (1996) for more detail on self-organised criticality.

64 

 Paul Cilliers

outputs. It may be somewhat tedious, but I can see no reason that they cannot be reduced to a system of rules. Call this Holland’s revenge, if you wish. What, then, becomes of the argument that neural networks can help us in our understanding of complex systems? I still maintain that large-scale, highly connected recurrent networks can serve as a general model of complex systems, but there are two important qualifications. In the first place, I am not aware of any real practical applications of such nets. They are just too cumbersome to work with. In the second place, even if it would become possible to build these models, they would cease to be “models” in the sense that they reduce complexity and thereby improve our understanding of the system.15 From the argument for the conservation of complexity  – the claim that complexity cannot be compressed (Cilliers 1998: 9–10) – it follows that a proper model of a complex system would have to be as complex as the system itself. As a result, the behaviour of the model will be as complex – and unpredictable – as that of the system itself. Is this conclusion a desperate one? It may be for those scientists and managers who still dream of a perfect grip on reality, usually in order to control it. For the rest of us, it serves as a reminder that our capabilities are limited, that there are limits to our understanding of the world.16 Nothing of what I said can be construed as an argument not to engage with those limits enthusiastically. A previous version of this paper was published in Théorie Littérature Enseignement (1999. vol. 17: 39–50) under the title “Règles et systèmes complexes”.

References Bak, P. 1996. How Nature Works. New York: Springer. Cilliers, F.P. 1991. Rules and relations: Some connectionist implications for cognitive science and language. In: South African Journal of Philosophy 10(2, May): 49–55.

15 Holland (1998: 24) provides this description of the conventional understanding of models: “Shearing away detail is the very essence of model building. Whatever else we require, a model must be simpler than the thing modeled, in certain kinds of fiction a model that is identical with the thing modeled provides an interesting device, as with Borges’ [...] map to the same scale as the land being mapped; but it never happens in reality. Even with virtual reality, which may come close to this literary [sic] identity one day, the underlying model obeys laws that have a compact description in the computer – a description that generates the details of the artificial world”. The possibility of this “compact description” of complexity is argued against in this article. 16 The relationship between the notion of the limit and deconstruction is developed in detail by Cornell (1992).



Rules and complex systems 

 65

Cilliers, P. 1998. Complexity and Postmodernism: Understanding Complex Systems. London: Routledge. Cilliers, P. 2000. What can we learn from a theory of complexity?. In: Emergence 2(1): 23–33. Cornell, D. 1992. The Philosophy of the Limit. London: Routledge. Derrida, J. 1976. Of Grammatology. Baltimore: John Hopkins University Press. Derrida, J. 1988. Limited Inc. Evanston: North-Western University Press. Hebb, D.O. 1949. The Organization of Behavior, New York: Wiley. Holland, J.H. 1997. Emergence. In: Philosophica 59(1): 1–40. Holland, J.H. 1998. Emergence: From Chaos to Order. Reading, MA: Addison-Wesley (Helix Books). Lyotard, J.F. 1986. Rules and paradoxes and svelte appendix. In: Cultural Critique 5:209–219. Pinker, S. & Mehler. J. (eds.). 1988. Connections and Symbols. Cambridge, MA: MIT Press. Rawls, J. 1955. Two concepts of rules. In: The Philosophical Review 64 (1, January): 3–32. Reddiford, G. 1985. Constitutions, institutions and games. In: Journal of the Philosophy of Sport, XII: 41–51. Singer, P. (ed.). 1994. Ethics. Oxford, UK: Oxford University Press. Winston, K.I. 1971. Justice and rules: A criticism. In: Logical Analysis 14: 177–82.

Paul Cilliers

What can we learn from a theory of complexity? The aim of this article is to investigate the implications of a general theory of complexity for social institutions and organisations, such as business corporations. Complexity theory has implications for the way we conceive of the structure of an organisation, as well as for the way in which complex organisations should be managed. However, a preliminary warning is necessary: The lessons to be learned from the study of complexity are somewhat oblique. Any hope that a study of complex systems will uncover the way of running an organisation is in vain. While we will not come up with a quick fix, the lessons are most certainly important. The first half of the article will investigate what we can learn from a theory of complexity. Most of these insights are widely accepted, but it is useful to revisit them briefly. This general understanding of complex systems also provides the background to the second half of the article, in which I investigate what we cannot learn from complexity theory. The “negative” part of the article is at least as important as the “positive” part. There I will investigate the unavoidability of an ethical dimension to all decisions made in a complex environment.

1 Complexity in a nutshell I will not provide a detailed description of complexity here, but only summarise the general characteristics of complex systems as I see them.1 1. Complex systems consist of a large number of elements that in themselves can be simple. 2. The elements interact dynamically by exchanging energy or information. These interactions are rich. Even if specific elements only interact with a few others, the effects of these interactions are propagated throughout the system. The interactions are nonlinear. 3. There are many direct and indirect feedback loops.

1 This summary is based on an extended analysis of complex systems in Cilliers (1998). Originally published in Emergence, 2000, 2(1): 23–33. © 2000 Lawrence Erlbaum Associates, Inc.

68 

 Paul Cilliers

4. Complex systems are open systems – they exchange energy or information with their environment – and operate at conditions far from equilibrium. 5. Complex systems have memory, not located at a specific place, but distributed throughout the system. Any complex system thus has a history, and the history is of cardinal importance to the behaviour of the system. 6. The behaviour of the system is determined by the nature of the interactions, not by what is contained within the components. Since the interactions are rich, dynamic, fed back, and, above all, nonlinear, the behaviour of the system as a whole cannot be predicted from an inspection of its components. The notion of “emergence” is used to describe this aspect. The presence of emergent properties does not provide an argument against causality, only against deterministic forms of prediction. 7. Complex systems are adaptive. They can (re)organise their internal structure without the intervention of an external agent. Certain systems may display some of these characteristics more prominently than others. These characteristics are not offered as a definition of complexity, but rather as a general, low-level, qualitative description. If we accept this description (which from the literature on complexity theory appears to be reasonable), we can investigate the implications it would have for social or organisational systems.

2 Complexity and organisations The notion of complexity has been applied to organisations in a number of different ways, and with varying degrees of rigor. I would like to emphasise two things. In the first place, the principles discussed here are of a very general nature. The contingent conditions at stake when investigating a specific case will be relevant, and may radically affect the importance of some of the implications. Despite this remark, I wish to stress, secondly, that this does not mean that the acknowledgement of the complexity of a situation allows us to be vague, nor does it imply a chaotic state of affairs. Complexity theory has important implications for the general framework we use to understand complex organisations, but within that (new) framework we must still be clear, as well as decisive. 1. Since the nature of a complex organization is determined by the interaction between its members, relationships are fundamental. This does not mean that everybody must be nice to each other; on the contrary. For example, for self-organisation to take place, some form of competition is a requirement



2.

3.

4.

5.

What can we learn from a theory of complexity? 

 69

(Cilliers 1998: 94–5). The point is merely that things happen during interaction, not in isolation. Complex organisations are open systems. This means that a great deal of energy and information flows through them, and that a stable state is not desirable. More importantly, it means that the boundaries of the organisation are not clearly defined. Statements of “mission” and “vision” are often attempts to define the borders, and may work to the detriment of the organisation if taken too literally. A vital organisation interacts with the environment and other organisations. This may (or may not) lead to big changes in the way the organisation understands itself. In short, no organisation can be understood independently of its context. Along with the context, the history of an organisation co-determines its nature. Two similar-looking organisations with different histories are not the same. Such histories do not consist of the recounting of a number of specific, significant events. The history of an organisation is contained in all the individual little interactions that take place all the time, distributed throughout the system. Unpredictable and novel characteristics may emerge from an organisation. These may or may not be desirable, but they are not by definition an indication of malfunctioning. For example, a totally unexpected loss of interest in a well-established product may emerge. Management may not understand what caused it, but it should not be surprising that such things are possible. Novel features can, on the other hand, be extremely beneficial. They should not be suppressed because they were not anticipated. Because of the nonlinearity of the interactions, small causes can have large effects. The reverse is, of course, also true. The point is that the magnitude of the outcome is not only determined by the size of the cause, but also by the context and by the history of the system.2 This is another way of saying that we should be prepared for the unexpected. It also implies that we have to be very careful. Something we may think to be insignificant (a casual remark, a joke, a tone of voice) may change everything. Conversely, the grand five-year plan, the result of huge effort, may retrospectively turn out to be meaning-

2 In this regard I have to stress that the butterfly metaphor borrowed from deterministic chaos is very misleading. There is no way in which the statement “a butterfly flapping its wings in Borneo could ‘cause’ a hurricane in Florida” can have any sense. The notion of causality loses all its meaning. There are many better ways of talking about a hurricane in Florida, despite the fact that we cannot be sure about exactly what caused it. Causes can be investigated, even if at best retrospectively.

70 

 Paul Cilliers

less. This is not an argument against proper planning; we have to plan. The point is just that we cannot predict the outcome of a certain cause with absolute clarity. 6. We know that organisations can self-organise, but it appears that complex systems also organise themselves toward a critical state.3 This not only means that at any given point we can expect the system to respond to external events on all possible scales of magnitude, but also that the system will organise itself to be maximally sensitive to events that are critical to the system’s survival. Think of language as a complex system. If there is a desperate need for new terms to describe important events, the system will organise itself to be critically sensitive to those terms specifically, and not necessarily to other novel terms. The “need” is determined by the context and the history of the system, not by a specific “decision” by some component of the system. Similarly, an organisation will self-organise to be critically sensitive to specific issues in the environment that may affect its wellbeing. The implications of self-organised criticality for organisational systems seems to be a subject that demands further investigation. 7. Complex organizations cannot thrive when there is too much central control. This certainly does not imply that there should be no control, but rather that control should be distributed throughout the system. One should not go overboard with the notions of self-organisation and distributed control. This can be an excuse not to accept the responsibility for decisions when firm decisions are demanded by the context. A good example here is the fact that managers are often keen to “distribute” the responsibility when there are unpopular decisions to be made – like retrenchments – but keen to centralise decisions when they are popular. 8. Complex organisations work best with shallow structures.4 This does not mean that they should have no structure. This point requires a little elaboration. Complexity and chaos  – whether in the technical or the colloquial sense – have little to do with each other. A complex system is not chaotic, it has a rich structure. One would certainly not describe the brain or language, prime examples of complex systems, as “chaotic.”5 I certainly would not put

3 For an introduction to self-organised criticality, see Bak (1997). For a discussion of some implications, see Cilliers (1998: 96–8). 4 The notion of “structure” here refers to the relationships among the various components of the system. Some of these relationships can be fairly fixed and static, others fairly fluid. 5 I am not implying that there are no lessons to be learned from chaos theory, but that they are more limited than is often believed. The notion of the “edge of chaos” is often useful, but even here I think we are better served by using the idea of critical organisation.



What can we learn from a theory of complexity? 

 71

my trust in a chaotic organisation. A complex system does have structure, but not a strictly hierarchical structure; perhaps not even a shallow structure. Structure can be shallow, but still extremely hierarchical. Perhaps the best way to think of this would be to say that there should be structure on all scales, and much interaction between different structural components. This is another aspect of complex organisations that could be fleshed out with insights from self-organised criticality. These few implications of complexity theory for organisations are important, and can dramatically affect our understanding of complex organisations. They can be spelled out in much more detail, but as I insisted above, this will have to be done in the context of specific organisations and their contingent conditions. In order to do that, we should also be clear about what we cannot learn from a theory of complexity.

3 W  hat we cannot learn from a theory of complexity I hope to show that the implications of this negative part of the article are at least as important as those following from the positive part. Acknowledgement of the limitations of our knowledge lies at the root of the whole western tradition of Socratic philosophical reflection, but I am sure that the mere acknowledgement of limitations is not enough. On the one hand, it suppresses the challenge to shift the boundaries of our knowledge. On the other hand, it stops short of investigating the ramifications of this limitation. I want to argue that one important consequence is that we are forced to take up an ethical position. What are the limits of a theory of complexity? Looking at the positive aspects we discussed above, you will notice that none is specific. They are all heuristic, in the sense that they provide a general set of guidelines or constraints. Perhaps the best way of putting it is to say that a theory of complexity cannot help us to take in specific positions, to make accurate predictions. This conclusion follows inevitably from the basic characteristics discussed above. In order to predict the behaviour of a system accurately, we need a detailed understanding of that system, i.e., a model. Since the nature of a complex system is the result of the relationships distributed all over the system, such a model will have to reflect all these relationships. Since they are nonlinear, no set of interactions can be represented by a set smaller than the set itself – superposition does not hold. This is one way of saying that complexity is not compressible. More-

72 

 Paul Cilliers

over, we cannot accurately determine the boundaries of the system, because it is open. In order to model a system precisely, we therefore have to model each and every interaction in the system, each and every interaction with the environment – which is of course also complex – as well as each and every interaction in the history of the system. In short, we will have to model life, the universe and everything. There is no practical way of doing this. Before I continue, two qualifications are required in order to prevent misunderstanding. The first is to re-emphasise that this is not the same as saying that complex systems are chaotic. Emergence is not a random or statistical phenomenon. Complex systems have structure, and, moreover, this structure is robust. Secondly, this does not imply that there is no point in developing formal models of complex systems. We can develop models on the basis of certain assumptions and limitations, just as with any scientific model. Let me put the matter in slightly different terms. The prediction of complex behaviour is only possible as a form of generalisation. However, when we deal with a complex system, we can never escape the necessity of facing the particular nature of the system at any given moment. Since we do not know the boundaries of the system, we never know if we have taken enough into consideration. We have to make a selection of all the possible factors involved, but under nonlinear conditions we will never know if something that was left out because it appeared to be insignificant was indeed so. What does this amount to in practice? It means that we have to make decisions without having a model or a method that can predict the exact outcome of those decisions. A theory of complexity cannot provide us with a method to predict the effects of our decisions, nor with a way to predict the future behaviour of the system under consideration. Does this mean we should avoid decisions, hoping that they will make themselves? Most definitely not. We cannot avoid them. Without activity in the system, without the energy provided by engaging with the system, it would probably wither away into a state of equilibrium, another word for death. Not to make a decision is of course also a decision. What, then, are the nature of our decisions? Because we cannot base them on calculation only – calculation would eliminate the need for choice – we have to acknowledge that our decisions have an ethical nature.

4 Ethics and complexity I want to make clear how the notion of ethics is used here. I do not take it to mean being nice or being altruistic. It has nothing to do with middle-class values, nor



What can we learn from a theory of complexity? 

 73

can it be reduced to some interpretation of current social norms. I use the word in a rather lean sense: it refers to the inevitability of choices that cannot be backed up scientifically or objectively. Why call it ethics? First, because the nature of the system or organisation in question is determined by the collection of choices made in it. There are, of course, choices to be made on all scales: major ones, as well as all the seemingly insignificant small ones made all the time – and remember that the scale of the effect is not related to the scale of the cause. In a way, the history of the organisation is nothing else but the collection of all these decisions. Secondly, since there is no final objective or calculable ground for our decisions, we cannot shift the responsibility for the decision on to something else – “Don’t blame me, the genetic algorithm said we should sell!” We know that all of our choices to some extent, even if only in a small way, incorporate a step in the dark. Therefore we cannot but be responsible for them. This may have a pessimistic ring to it, but that need not be the case. An awareness of the contingency and provisionality of things is far better than a false sense of security. Such an awareness is also an integral part of the notion “adaptive”. Of course, this does ultimately translate into a value system, but this system is not a given, something that is governed by a priori notions of good and bad. The system of values is itself a matter of choice. Our decisions are guided by some notion of what we think the organisation should be – and it is in this “should” that the ethical dimension is contained. If an organisation decides “The bottom line is our first priority”, then that is the kind of organisation it would be: nothing comes in the way of money. The central issue here is that a system of values is exactly that. Values are not natural things that we can read off the face of nature; we choose them. It is not written in the stars that the bottom line is vital to the survival of a company, it comes with accepting a certain understanding of what a company should be under, say, capitalist conditions. Of course, it is not only the nature of the organisation that is determined by choices, but also our nature as individuals. We are also the result of our choices. Thieves are not thieves when they are caught out, or found guilty under some legal system. Thieves are thieves when they steal. A further implication of this “ethical” position needs to be spelled out. “Ethics” is part of all the different levels of activities in an organisation. These ethical components, related to the values and preferences of the members of the organisation, are often referred to as merely “politics”, something separate to the organisation’s real operation and goals. The argument here is that the political aspects of the interactions in an organisation are not something extraneous to the workings of that organisation. It is not something that has to be dealt with in order to guarantee the proper working of the organisation, it is integral to its

74 

 Paul Cilliers

proper working. The individual and collective values of members of the system cannot be separated from their functional roles. This point is probably instinctively accepted by most good managers. The fact of the matter is that this is the case, whether it is accepted by management as such or not. To summarise the argument: The ethical position is not something imposed on an organisation, something that is expected of it. It is an inevitable result of the inability of a theory of complexity to provide a complete description of all aspects of the system.6

5 Modelling and calculation It may appear at this stage as if I am arguing against any kind of calculation, that I am dismissing the importance of modelling complex systems. Nothing is further from the truth. The important point I want to make is that calculation will never be sufficient. The last thing this could mean is that calculation is unnecessary. On the contrary, we have to do all the calculation we possibly can. That is the first part of our responsibility as scientists and managers. Calculation and modelling will provide us with a great deal of vital information. It will just not provide us with all the information. Perhaps I am wrong here: it may become possible for some sophisticated model to provide all the information about a specific system. The problem would remain, however, that this information has to be interpreted. All the models we construct – whether they are formal, mathematical models, or qualitative, descriptive models – have to be limited. We cannot model life, the universe, and everything. There may not be any explicit ethical component contained within the model itself, but ethics (in the sense in which I use the term) has already played its part when the limits of the model were determined, when the selection was made of what would be included in the frame of the investigation. The results produced by the model can never be interpreted independently of that frame. This is no revelation, it is something every scientist knows, or at least should know. Unfortunately, less scrupulous people, often the popularisers of some scientific idea or technique, extend the field of applicability of that idea way beyond the framework that gives it sense and meaning.

6 This argument can also be made from a strictly philosophical position, particularly from the perspective of deconstruction. Despite the resistance to Derrida’s post-structural insistence on undecidability, it is a strongly argued position that does not imply indecision or relativism. For a good philosophical introduction to this perspective, see Caputo (1997).



What can we learn from a theory of complexity? 

 75

My position could be interpreted as an argument that contains some mystical or metaphysical component, slipped in under the name “ethics”. In order to forestall such an interpretation, I will digress briefly. It is often useful to distinguish between the notions “complex” and “complicated”. A jumbo jet is complicated, a mayonnaise is complex (at least for the French). A complicated system is something we can model accurately (at least in principle). Following this line of thought, one may argue that the notion “complex” is merely a term we use for something we cannot yet model. I have much sympathy for this argument. If one maintains that there is nothing metaphysical about a complex system, and that the notion of causality has to be retained, then perhaps a complex system is ultimately nothing more than extremely complicated. It should therefore be possible to model complex systems in principle, even though it may not be practical. Would the advent of adequate models of complex systems relieve us from our ethical responsibility? My contention is that it would not. Here is why: We cannot make simple models of complex systems. Their non-linear nature, or, in other words, their incompressibility, demands that the model of a system be as complex as the system itself. If it is in the nature of the system to behave, at least sometimes, in novel and unpredictable ways, the model must also do so. In any case, how would we be able to determine if the model were indeed an adequate model of the system if we were already in trouble when trying to decide what constitutes the system itself? It would be as difficult to interpret the model as to interpret the system itself.7 Good models of complex systems can be extremely useful; I just do not believe that they will allow us to escape the moment of interpretation and decision.

6 Complexity and the humanities Whatever we take the notion of ethics to mean, our analysis of what we can and cannot learn from a theory of complexity has shown that a proper reflection on complex organisations will have to involve the humanities. Perhaps we can describe the humanities as those disciplines that realise that their subject matter cannot be studied only by formal means. There are, of course, a number of disciplines that immediately come to mind: political science, sociology, psychology, and, of course, philosophy. Allow me the

7 The problem of interpretation is one of the central issues in the history of philosophy, so much so that it has its own name: hermeneutics.

76 

 Paul Cilliers

opinion that philosophy, the mother of all the sciences – but in an instrumentaland outcomes-based world often seen as redundant – may yet prove to be one of our greatest resources. The need to reflect critically on the nature and the limits of our knowledge and understanding is indispensable to a study of complexity. I do not, however, want to end with that cheer for the home team. I also want to stress the importance of the arts. Artists through the ages have attempted to find new ways of portraying and understanding the complexities of our world. Under certain conditions, a good novel may teach us more about human nature than mathematical models of the brain, or the theories of cognitive psychology. An engagement with the arts should not be a luxury in which we indulge after “work”, it should be intertwined with our work. Faced with the complexities of life, we all have to be artists in some sense of the word. It is to be hoped that this will not only help us to a better understanding of our organisations, it will also make us better human beings. This article is based on a paper delivered at Managing the Complex, the Third Annual Symposium of the New England Complex Systems Institute, held in Boston, March 1999.

References Bak, P. 1997. How Nature Works: The Science of Self-organized Criticality. Oxford: Oxford University Press. Caputo, J.D. 1997. Deconstruction in a Nutshell. New York: Fordham University Press. Cilliers, P. 1998. Complexity and Postmodernism: Understanding Complex Systems. London: Routledge.

Paul Cilliers

Knowledge, complexity and understanding The strange thing about television is that it doesn’t tell you everything. It shows you everything about life on earth, but the mysteries remain. Perhaps it is in the nature of television (Thomas Jerone Newton in The Man Who Fell to Earth).

During most events concerned with knowledge management, someone starts a presentation by saying that they will not revisit the problem of the distinction between knowledge and data. Usually a sigh goes through the audience, seemingly signifying relief. But why relief? Is it because they will not be bored with an issue that has been resolved already, or because they are glad that they will not be confronted with these thorns again? I suspect that they want to believe that the first reason is the case, but that in fact it is the second. In what follows I therefore want to problematise the notion of “knowledge”. I will argue that when talking about the management of “knowledge”, whether by humans or computers, there is a danger of getting caught in the objectivist/ subjectivist (or fundamentalist/relativist) dichotomy. The nature of the problem changes if one acknowledges the complex, interactive nature of knowledge. These arguments, presented from a philosophical perspective, should have less influence on the practical techniques employed in implementing knowledge management systems than on the claims made about what is actually achieved by these systems.

1 The traditional trap The issues around knowledge – what we can know about the world, how we know it, what the status of our experiences is  – have been central to philosophical reflection for ages. Answers to these questions, admittedly oversimplified here, have traditionally taken one of two forms. On the one hand there is the belief that the world can be made rationally transparent, that with enough hard work knowledge about the world can be made objective. Thinkers like Descartes and Habermas are often framed as being responsible for this kind of attitude. It goes under numerous names including positivism, modernism, objectivism, rationalism and epistemological fundamentalism. On the other hand, there is the belief that knowledge is only possible from a personal or cultural-specific perspective, Originally published in Emergence, 2000, 2(4): 7–13. © 2000 Lawrence Erlbaum Associates, Inc.

78 

 Paul Cilliers

and that it can therefore never be objective or universal. This position is ascribed, correctly or not, to numerous thinkers in the more recent past like Kuhn, Rorty and Derrida, and its many names include relativism, idealism, postmodernism, perspectivism and flapdoodle. Relativism is not a position that can he maintained consistently1 and of course the thinkers mentioned above have far more sophisticated positions than portrayed in this bipolar caricature. There are also recent thinkers who attempt to move beyond the fundamentalist/relativist dichotomy,2 but it seems to me that when it comes to the technological applications of theories of knowledge, there is an implicit reversion to one of these traditional positions. For those who want to computerise knowledge, knowledge has to be objective. It must be possible to gather, store and manipulate knowledge without the intervention of a subject. The critics of formalised knowledge, on the other hand, usually fall back on arguments based on subjective or culture-specific perspectives to show that it is not possible, that we cannot talk about knowledge independently of the knowing subject. I am of the opinion that a shouting match between these two positions will not get us much further: The first thing we have to do is to acknowledge the complexity of the problem with which we are dealing. This will unfortunately not lead us out of the woods, but it should enable a discussion that is more fruitful than the objectivist/subjectivist debate.

2 Complexity and understanding An understanding of knowledge as constituted within a complex system of interactions3 would, on the one hand, deny that knowledge can be seen as atomised “facts” that have objective meaning. Knowledge comes to be in a dynamic network of interactions, a network that does not have distinctive borders. On the

1 If relativism is maintained consistently, it becomes an absolute position. From this one can see that a relativist is nothing but a disappointed fundamentalist. However, this should not lead one to conclude that everything that is called postmodern leads to this weak position. Lyotard’s seminal work, The Postmodem Condition (1984), is subtitled A Report on Knowledge. He is primarily concerned with the structure and form of different kinds of knowledge, not with relativism. An informed reading of Derrida will also show that deconstruction does not imply relativism at all. For a penetrating philosophical study of the problem, see (Norris, 1991). 2 The critical realism of Bhaskar (1986) is a good example. 3 Complex systems are discussed in detail in Cilliers (1998).



Knowledge, complexity and understanding 

 79

other hand, this perspective would also deny that knowledge is something purely subjective, mainly because one cannot conceive of the subject as something prior to the “network of knowledge”, but rather as something constituted within that network. The argument from complexity thus wants to move beyond the objectivist/subjectivist dichotomy. The dialectical relationship between knowledge and the system within which it is constituted has to be acknowledged. The two do not exist independently, thus making it impossible to first sort out the system (or context), and then identify the knowledge within the system. This codetermination also means that knowledge and the system within which it is constituted are in continual transformation. What appears to be uncontroversial at one point may not remain so for long. The points above are just a restatement of the claim that complex systems have a history, and that they cannot be conceived of without taking their context into account. The burning question at this stage is whether it is possible to do that formally or computationally. Can we incorporate the context and the history of a system into its description, thereby making it possible to extract knowledge from it? This is certainly possible (and very useful) in the case of relatively simple systems, but with complex systems there are a number of problems. These problems are, at least to my mind, not of a metaphysical but of a practical nature. The first problem has to do with the nonlinear nature of the interactions in a complex system. From this it can be argued (see Cilliers 1998: 9–10 and Richardson et al. 2000) that complexity is incompressible. There is no accurate (or, rather, perfect) representation of the system that is simpler than the system itself. In building representations of open systems, we are forced to leave things out, and since the effects of these omissions are nonlinear; we cannot predict their magnitude. This is not an argument claiming that reasonable representations should not be constructed, but rather one that the unavoidable limitations of the representations should be acknowledged. This problem  – which can be called the problem of boundaries4  – is compounded by the dynamic nature of the interactions in a complex system. The system is constituted by rich interaction, but since there is an abundance of direct and indirect feedback paths, the interactions are constantly changing. Any activity in the system reverberates throughout the system, and can have effects that are very difficult to predict; once again as a result of the large number of nonlinear interactions. I do not claim that these dynamics cannot be modelled. It could be possible for richly connected network models to be constructed. However, as soon as these networks become sizable, they become extremely difficult to train.

4 The problem of boundaries is discussed in more detail in Cilliers (2001).

80 

 Paul Cilliers

It also becomes rather hard to figure out what is actually happening in them. This is no surprise if one grants the argument that a model of a complex system will have to be as complex as the system itself. Reduction of complexity always leads to distortion. What are the implications of the arguments from complexity for our understanding of the distinction between data and knowledge? In the first place, it problematises any notion that data can be transformed into knowledge through a pure, mechanical and objective process. However; it also problematises any notion that would see the two as totally different things. There are facts that exist independently of the observer of those facts, but the facts do not have their meaning written on their faces. Meaning only comes to be in the process of interaction. Knowledge is interpreted data. This leads us to the next big question: What is involved in interpretation, and who (or what) can do it?

3 Knowledge and the subject The function of knowledge management seems to be either to supplement the efforts of a human subject who has to deal with more data than is possible, or to free the subject up for other activities (perhaps to do some thinking for a change). Both these functions presuppose that the human subject can manipulate knowledge. This realisation leads to questions in two directions. One could debate the efficiency of human strategies to deal with knowledge and then attempt to develop them in new directions. This important issue will not be pursued further here. There is another, perhaps philosophically more basic, question, and that has to do with how the human subject deals with knowledge at all. Given the complexities of the issue, how does the subject come to forms of understanding, and what is the status of knowledge as understood by a specific subject? This has been pursued by many philosophers, especially in the discipline known as hermeneutics. However, I am not aware that this has occurred in any depth in the context of complexity theory.5 How does one perceive of the subject as something that is not atomistically self-contained, but is constituted through dynamic interaction? Moreover, what is the relationship between such a subject and its

5 An important contribution was made by reinterpreting action theory from the perspective of complexity (Juarrero 1999). Some preliminary remarks, more specifically on complexity and the subject, are made in Cilliers & De Villiers (2000).



Knowledge, complexity and understanding 

 81

understanding of the world? A deeper understanding of what knowledge is, and how to “manage” it, will depend heavily on a better understanding of the subject. This is a field of study with many opportunities. Apart from calling for renewed effort in this area, I only want to make one important remark. It seems that the development of the subject from something totally incapable of dealing with the world on its own into something that can begin to interpret  – and change  – its environment is a rather lengthy process. Childhood and adolescence are necessary phases (sometimes the only phases) in human development. In dealing with the complexities of the world there seems to be no substitute for experience (and education). This would lead one to conclude that when we attempt to automate understanding, a learning process will also be inevitable. This argument encourages one to support computing techniques that incorporate learning (like neural networks) rather than techniques that by to abstract the essence of certain facts and manipulate them in terms of purely logical principles. Attempts to develop a better understanding of the subject will not only be helpful in building machines that can manage knowledge, they will also help humans better understand what they do themselves. We should not allow the importance of machines (read computers) in our world to lead to a machine-like understanding of what it is to be human.

4 Implications In Nicholas Roeg’s remarkably visionary film The Man Who Fell to Earth (1976), an alien using the name Thomas Jerome Newton (superbly played by David Bowie) tries to understand human culture by watching television, usually a whole bunch of screens at the same time. Despite the immense amount of data available to him, he is not able to understand what is going on directly. It is only through the actual experience of political complexities, as they unfold in time, that he begins to understand. By then he is doomed to remain earthbound. I am convinced that something similar is at stake for all of us. Having access to untold amounts of information does not increase our understanding of what it means. Understanding, and therefore knowledge, follows only after interpretation. Since we hardly understand how humans manage knowledge, we should not oversimplify the problems involved in doing knowledge management computationally. This does not imply that we should not attempt what we can – and certain spectacular advances have been made already – but that we should be careful in the claims we make about our (often still to be finalised) achievements.

82 

 Paul Cilliers

The perspective from complexity urges that, among others, the following factors should be kept in mind: –– Although systems that filter data enable us to deal with large amounts of it more effectively, we should remember that filtering is a form of compression. We should never trust a filter too much. –– Consequently, when we talk of mechanised knowledge management systems, we can (at present?) only use the word “knowledge” in a very lean sense. There may be wonderful things to come, but at present I do not know of any existing computational systems that can in any way be seen as producing “knowledge”. Real breakthroughs are still required before we will have systems that can be distinguished in a fundamental way from database management. Good data management is tremendously valuable, but cannot be a substitute for the interpretation of data. –– Since human capabilities in dealing with complex issues are also far from perfect, interpretation is never a merely mechanical process, but one that involves decisions and values. This implies a normative dimension to the “management” of knowledge. Computational systems that assist in knowledge management will not let us escape from this normativity. Interpretation implies a reduction in complexity. The responsibility for the effects of this reduction cannot be shifted away on to a machine. –– The importance of context and history means that there is no substitute for experience. Although different generations will probably place the emphasis differently, the tension between innovation and experience will remain important. These considerations should assist in developing an understanding of knowledge management that could be called “organic”, but perhaps also “ethical”. Acknowledgement: The financial assistance of the National Research Foundation: Social Sciences and Humanities (of South Africa) toward this research is hereby acknowledged. Opinions expressed and conclusions arrived at are those of the author; and are not necessarily to be attributed to the National Research Foundation.

References Bhaskar, R. 1986. Scientific Realism and Human Emancipation. London: Verso. Cilliers, P. 1998. Complexity and Postmodernism: Understanding Complex Systems. London: Routledge. Cilliers, P. 2001. Boundaries, Hierarchies and Networks in Complex Systems.



Knowledge, complexity and understanding 

 83

Cilliers, P. & De Villiers, T. 2000. The Complex ‘I’. In: Wheeler, W. (ed.). The Political Subject. London: Lawrence & Wishart. Juarrero, A. 1999. Dynamics in Action: Intentional Behavior as a Complex System. Cambridge, MA: MIT Press. Lyotard, J.F. 1984. The Postmodern Condition: A Report on Knowledge. Manchester; UK: Manchester University Press. Norris, C. 1997. Against Relativism: Philosophy of Science, Deconstruction and Critical Theory. Oxford, UK: Blackwell. Richardson, K., Cilliers, P. & Lissack, M. 2000. Complexity Science: A ‘Grey’ Science for the ‘Stuff in Between’. In: Proceedings of the First International Conference on Systems Thinking in Management. Geelong, Australia, 532–537.

Paul Cilliers

Boundaries, hierarchies and networks in complex systems 1 Complexity Complexity theory has been a bright new star in the academic firmament for a while now. It is being pursued eagerly in a number of disciplines (see Thrift 1999), generally with a fair amount of hype. Why the enthusiasm, and more particularly, why is there so much of it in the organisational sciences? My suspicion is that the reason has a lot to do with the hope that we are finally onto a method that will improve our understanding of, and therefore our control over complex systems like organisations. The argument may go like this: if we pay enough attention to flat hierarchies, networks of interaction, non-linearity and emergence, we may finally be able to develop a general theory of complex organisations. This will, of course, be a much sought-after management tool, and it should come as no surprise that so many are looking for it. It should also not be a mystery that the Santa Fe style of approaching the problem – lots of chaos theory and mathematics – should be the most popular. We want to predict the behaviour of complex systems, and for that, we need good models. Of course, complexity theory did not appear on the scene without antecedents. In many ways, it is a continuation of what was done in cybernetics, general systems theory and chaos theory. These disciplines also generated lots of hype – and lots of results, of course – but could never quite deliver the theories and tools required for a general theory of complexity. There are a number of reasons for this, but two related reasons, I think, are central: they did not pay enough attention to the historical nature of complex systems, and consequently, did not pay enough attention to the radically contingent nature of a complex system. Complexity was taken to be symmetrical in time, a point of view no longer tenable after the work of Prigogine (see also Dasgupta 1997: 138; and Emmeche 1997: 48, 58). The burning question is whether we can take this and other characteristics of complexity (Cilliers 1998: 2–7) into account, and then succeed where previous efforts failed. Is it possible to have a general theory of complex systems? In this paper, I suggest that although we can say a lot of important things about complexity in general, it is not possible to develop a general model for complex systems. This has to do with the meaning of the notions “model” and “complexOriginally published in the International Journal of Innovation Management, June 2001, 5(2): 135–147. © Imperial College Press.

86 

 Paul Cilliers

ity”. In what follows, I will look at the limitations of models of complex systems by examining the status of boundaries and hierarchies. I will conclude by reflecting on the status of neural network models, and on some of the implications the whole argument has on our understanding of organisations. There is one question that can be pre-empted now: is it possible to have a science of complexity? I would argue that it is, but that it implies a revision of our notion of what constitutes science. In an editor’s note to a short review article by Corning (1998), the following statement is made: “Until the ‘complexity science’ researchers can develop a formal notation in symbols and syntax, while at the same time respecting its subjective nature [sic], it will not really be a ‘science’” (Corning 1998: 197). If this strict, formal and quantificatory attitude remains the way in which science is defined, then there will be no “science” of complexity. However, our knowledge of complex systems is, to my mind at least, undermining such a strict understanding of science. It forces us to consider strategies from both the human and the natural sciences, to incorporate both narratives and mathematics — not in order to see which one is best, but in order to help us to explore the advantages and limitations of all of them. Complexity studies should thus be seen not as aiming at a new ‘synthetic theory’ of complexity of any kind, but as a cross-disciplinary field of research and meeting place for dialogue between specialised groups of people such as biologists, physicists, philosophers, mathematicians, computer scientists, and, last but not least, science writers (Emmeche 1997: 43).

Before turning to how boundaries and hierarchies make the limitations of our models of complexity explicit, we should first explore the notion of a model in a little more detail.

2 Models The notion of a model is central to scientific understanding. The notion will be used here in a wide sense (i.e. theories and systems of rules can also be seen as models). In the context of complexity, the role of models is described in the following way by Csányi (in Khalil & Boulding 1996: 148): Any kind of scientific statement, concept, law and any description of a phenomenon is a model construction which tries to reflect phenomena of the external world. Reality is extremely complex; it consists of strongly or more weakly related events. Science makes an attempt to separate and isolate different effects and phenomena. It seeks the simplest relationships by which examined phenomena can at least be described or demonstrated. It



Boundaries, hierarchies and networks in complex systems 

 87

creates simplified models which only partly reflect reality, but which allow contemplation, and what is most important, pragmatic, even if sometimes modest, predictions.

We cannot deal with reality in all its complexity. Our models have to reduce this complexity in order to generate some understanding. In the process, something is obviously lost. If we have a good model, we would hope that that which is left out is unimportant. It should be clear already that purely quantitative models of complex systems, which abstract from a set of real properties to numerical values, will be problematic (Emmeche 1997: 54). The underlying problem with models of complexity is, however, even more serious. No matter how we construct the model, it will be flawed, and what is more, we do not know in which way it is flawed. In order to understand this claim, we have to remember the non-linear nature of the interactions in complex systems. This non-linearity has two important consequences. In the first place, when there are a lot of simultaneous, non-linear interactions, it soon becomes impossible to keep track of causal relationships between components. Secondly, from the non-linear nature of complex systems, we can deduce that they are incompressible (Cilliers 1998: 10). If we add to this the historical nature of complex systems, the problem should become clear: models have to reduce the complexity of the phenomena being described, they have to leave something out. However, we have no way of predicting the importance of that which is not considered. In a non-linear world where we cannot track a clear causal chain, something that may appear to be unimportant now may turn out to be vitally important later. Or vice versa, of course. Our models have to “frame” the problem in a certain way, and this framing will inevitably introduce distortions.1 This is not an argument against the construction of models. We have no choice but to make models if we want to understand the world. It is just an argument that models of complex systems will always be flawed in principle, and that we have to acknowledge these limitations. What then of the argument that it may be possible to incorporate absolutely all the information concerning a complex system into some fancy (neural network) model? I do not wish to argue that it is impossible to repeat the complexity of a system in another medium, but one should remember that we now have a “model” that is as complex as the system being modelled. It will be as difficult to understand as the system itself, and its behaviour will be as unpredictable. If

1 In this paper, the ethical issues arising from the acknowledgement of complexity will not be examined, but it should be clear that the selection of a certain frame always involves normative issues (see Cilliers 2000b).

88 

 Paul Cilliers

the history of the model and the history of the system are not kept identical (and I cannot see how this can be done in anything but the most trivial of cases), the two will soon become uncorrelated. My conclusion is that it is impossible to have a perfect model of a complex system. This is not because of some inadequacy in our modelling techniques, but a result of the meaning of the notions “model” and “complex”. There will always be a gap between the two. This gap should serve as a creative impulse that continually challenges us to transform our models, not as a reason to give up.

3 Structure The claim that our models of complex systems cannot be perfect introduces a next layer of problems: what is it then that is described by our models? Are they merely constructions or instruments, or do they reflect reality in some way? Both claims have had strong support. One way of naming these two traditions is to say that the attempt to reflect nature (accurately) is a modern approach, and that giving up that attempt is post-modern. Emmeche (1997: 46) argues that we can only deal with complexity if we adopt elements from both kinds of ethos. One can make a slightly stronger and more difficult demand: both approaches should be followed simultaneously. We are always busy with the world itself, and simultaneously, we cannot grasp it fully. Let us explore this a little further. A distinction is often made between “descriptive” and “ontological” complexity (e.g. by Emmeche 1996: 43). The first has to do with the complexity of our descriptions, the second with the “actual” complexity of things in the world. If one maintains this distinction, it would be easy to fall into the kind of dichotomy mentioned above. We would have descriptions of the world, and separate from it, the world itself. This is the trap stepped into by the classical approach to artificial intelligence: trying to make formal models that should represent the world accurately (see Cilliers, 1998: 58–88). The relationship between our descriptions of the world and the world itself is, however, more complex. There is a constant to and fro between them in which our models and, especially in the case of the human sciences, the world itself is transformed. Since our models cannot “fit” the world exactly, there are many degrees of freedom in which they can move. They are, however, simultaneously constrained by the world in many ways. There is feedback from the world that tells us something about the appropriateness of our models. The situation is the following: there is on the one hand freedom in modelling, and on the other hand, constraints from reality, but the two are not independent of each other.



Boundaries, hierarchies and networks in complex systems 

 89

We will return to the notion of constraints below when we look at boundaries, but for now it is important to realise that the notion of a constraint is not a negative one. It is not something which merely limits possibilities, constraints are also enabling. By eliminating certain possibilities, others are introduced. Constraints provide a framework that enables descriptions to be built up around it.2 When dealing with complexity, though, these frameworks cannot be fixed. They are constantly being transformed, and therefore our models will always be provisional. What then is it that is described by our models? I would argue that models attempt to grasp the structure of complex systems. Complex systems are neither homogeneous nor chaotic. They have structure, embodied in the patterns of interactions between the components.3 Some of these structures can be stable and long-lived (and are therefore easier to catch in or model), whilst others can be volatile and ephemeral. These structures are also intertwined in a complex way. We find structure on all scales.4 In order to see how difficult it is to grasp these structures, it is necessary to look at the boundaries of complex systems, and to the role of hierarchies within them.

4 Boundaries In order to be recognisable as such, a system must be bounded in some way. However, as soon as one tries to be specific about the boundaries of a system, a number of difficulties become apparent. For example, it seems uncontroversial to claim that one has to be able to recognise what belongs to a specific system, and what does not. But complex systems are open systems where the relationships amongst the components of the system are usually more important than the components themselves. Since there are also relationships with the environment, specifying clearly where a boundary could be is not obvious.

2 For a more detailed discussion of constraints, see Juarrero (1999: 131–150). 3 The notion of “structure” is used in many different and confusing ways. In this analysis, it refers to the patterns of interaction in the system, and underplays a distinction between the structure on the one hand, and activities within that structure on the other. Structure is the result of action in the system, not something that has to exist in an a priori fashion. The advantages of a network model of complexity is that we can depict rather stable structures, as well as more volatile ones using the same means (see Cilliers 1998: 99–100). 4 Structure is not chaotic, but often has a fractal nature (Csányi in Khalil & Boulding 1996: 158), especially if the system is critically organised (see Cilliers 1998: 96–98).

90 

 Paul Cilliers

One way of dealing with the problem of boundaries is to introduce the notion of “operational closure”.5 For a system to maintain its identity, it must reproduce itself (internally). These arguments often follow from the work by Maturana and Varela on autopoiesis. Zeleny (in Khalil & Boulding, 1996: 123) defines an autopoietic system as [...] a system that is generated through a closed organisation of production processes such that the same organisation of processes is regenerated through the interaction of its own products (components), and a boundary emerges as a result of the same constitutive processes.

When dealing with complex systems in an “operational” way, there is nothing wrong with this approach. One should be careful, however, not to overemphasise the closure of the boundary. The boundary of a complex system is not clearly defined once it has “emerged”. Boundaries are simultaneously a function of the activity of the system itself, and a product of the strategy of description involved. In other words, we frame the system by describing it in a certain way (for a certain reason), but we are constrained in where the frame can be drawn. The boundary of the system is therefore neither purely a function of our description, nor is it a purely natural thing. We can never be sure that we have “found” or “defined” it clearly, and therefore the closure of the system is not something that can be described objectively. An overemphasis on closure will also lead to an understanding of the system that may underplay the role of the environment. However, we can certainly not do away with the notion of a boundary. Our understanding of boundaries can be given a little more content by considering the following two issues. The first concerns the “nature” of boundaries. We often fall into the trap of thinking of a boundary as something that separates one thing from another. We should rather think of a boundary as something that constitutes that which is bounded. This shift will help us to see the boundary as something enabling, rather than as confining. To quote Zeleny (133) again: All social systems, and thus all living systems, create, maintain, and degrade their own boundaries. These boundaries do not separate but intimately connect the system with its environment. They do not have to be just physical or topological, but are primarily functional, behavioral, and communicational. They are not “perimeters” but functional constitutive components of a given system.

5 The work of Niklas Luhmann provides a good example of this approach. (For a monograph in English, see Luhmann 1989.)



Boundaries, hierarchies and networks in complex systems 

 91

As an example of this logic, think of the eardrum. It forms the boundary between the inner and the outer ear, but at the same time, it exists in order to let the sound waves through. As a matter of fact, if it was not there, the sound waves would not be able to get through at all! If the boundary is seen as an interface participating in constituting the system, we will be more concerned with the margins of the system, and perhaps less with what appears to be central.6 A second boundary issue concerns the “place” of the boundary. The propensity we have towards visual metaphors inclines us to think in spatial terms. A system is therefore often visualised as something contiguous in space. This tendency is reinforced by the prevalence of biological examples of complex systems. We think of systems in an “organistic” way. Social systems are obviously not limited in the same way. Parts of the system may exist in totally different spatial locations. The connections between different components could be seen as virtual, and therefore the system itself may exist in a virtual space. This much should be self-evident to most inhabitants of the global village, but there are two important implications to be drawn from this. The first is that non-contiguous subsystems could be part of many different systems simultaneously. This would mean that different systems interpenetrate each other, that they share internal organs. How does one talk of the boundary of the system under these conditions? A second implication of letting go of a spatial understanding of boundaries would be that in a critically organised system we are never far away from the boundary. If the components of the system are richly interconnected, there will always be a short route from any component to the “outside” of the system. There is thus no safe “inside” of the system, the boundary is folded in, or perhaps, the system consists of boundaries only. Everything is always interacting and interfacing with others and with the environment; the notions of “inside” and “outside” are never simple or uncontested. In accepting the complexity of the boundaries of complex systems, we are committed to be critical about how we use the notion since it affects our understanding of such systems, and influences the way in which we deal with them. The notion of “boundary critique” is not a new one (see Midgley et al. 1998), but in this critique we have to keep the enabling nature boundaries as well as their “displacement” in mind.

6 Although it will not be elaborated on in this text, a number of the ideas presented have a close affinity to arguments from deconstruction. For more detail, see Cilliers 1998, especially Chapter 3.

92 

 Paul Cilliers

5 Hierarchies An analysis of the importance of hierarchies has been part of the study of complex systems for a long time (Pattee 1973; Simon 1962). In his seminal paper, Simon gives at least three reasons why hierarchies are important. In the first place, a modular structure would make it easier for new complex systems to be generated. He uses the example of two watchmakers, one building each watch from scratch, the other first constructing basic sub-assemblies, and then connecting these together. The second, he argues, will be more efficient. This “hierarchical” structure would also allow the system to take better advantage of evolutionary opportunities. In the second place, hierarchies establish unambiguous routes of communication. If the system is hierarchical, an algorithm can be developed that would ensure that information would get from A to B. In the third place, Simon argues that hierarchical systems have a lot of redundancy, and that it is therefore possible to construct models of such systems that are simpler than the system itself (a claim which is obviously somewhat at odds with the position argued for here). A somewhat contrary position is taken in some contemporary discussions of complex systems. A lot of emphasis is placed on self-organisation and the “distributed” nature of the structure in a system. According to these arguments, complex systems do not have central control systems. They have to be dynamic and adaptable, not rigid or invariable. Consequently, the notion of hierarchy is resisted. In terms of the structure of organisations, it is often argued that to the extent that there should be hierarchies at all, they should be shallow and loose. There must be enough space for innovation. Both these positions oversimplify the role of hierarchies in complex systems. Hierarchies are certainly necessary, but the way in which they work differs in important respects from the classical understanding. Let us examine these differences. In the first place, it must be underscored that systems cannot do without hierarchies. Complex systems are not homogeneous things. They have structure, and moreover, this structure is asymmetrical (see Cilliers 1998: 120, 124, 147–148). There are subsections with functions, and for them to exist at all, there has to be some form of hierarchy. Problems arise, however, when these hierarchies are seen as either too clearly defined or too permanent. The classical understanding of hierarchies tends to view them as being nested.7 In reality however, hierarchies

7 This is perhaps again a legacy of biological models – subsystems are seen as “organs”. Biological systems are subjected to constraints that may not apply to all complex systems, especially



Boundaries, hierarchies and networks in complex systems 

 93

are not that well-structured. They interpenetrate each other, i.e. there are relationships which cut across different hierarchies. These interpenetrations may be fairly limited or so extensive that it becomes difficult to typify the hierarchy accurately in terms of prime and subordinate parts. Simon (1962), of course, knows this. Nevertheless, in the hope of coming up with enough hierarchical structure to enable modelling of the system, he emphasises that which falls within the hierarchies, and not the interpenetrations. He argues that many complex systems are “near decomposable”, meaning that hierarchical models will provide a fair approximation. This view would see the interpenetrations as part of the messiness of complexity, whereas I would rather see them as indispensable. Similar to the notion of boundaries discussed above, the structure of a complex system cannot be described merely in terms of clearly defined hierarchies. This is because the structure of complexity is usually fractal, there is structure on all scales. The cross-communications between hierarchies are not accidental, but part of the adaptability of the system. Alternative routes of communication are vital in order to subvert hierarchies that may have become too dominant or obsolete. Cross-connections may appear to be dormant for long, but in the right context may suddenly play a vital role. This leads directly to the next point: part of the vitality of a system lies in its ability to transform hierarchies. Although hierarchies are necessary in order to generate frameworks of meaning in the system, they cannot remain unchanged. As the context changes, so must the hierarchies. Some hierarchies may be more long-lived than others, but it is important to perceive of hierarchies as transformable entities. This may seem to be self-evident, but I do not think that managers regularly think in these terms. They may realise that they can be replaced, but they do not often perceive their positions to be, in principle, provisional. They also tend to think of the interpenetrations as obstacles to efficient management, and not as vital routes of communication. To summarise then, hierarchies are necessary, but they are not neatly nested. The hierarchies in a system have a complex structure themselves. Whatever their structure, hierarchies are furthermore not permanent, they have to be transformed. Transformation does not imply that hierarchies are to be destroyed, but that they should be shifted.8 The argument is thus not setting up an opposition between, for example, hierarchies and teams in an organisation, and then insisting that one should find a balance between the two (see, e.g. Romme 1996). Teams have hierarchies too, and this should be acknowledged. It is better to

not social systems. 8 This claim can also be substantiated by arguments from deconstruction. See footnote 5.

94 

 Paul Cilliers

make as much as possible of the structure in an organisation explicit, and then to deconstruct it, rather than to claim that there is no (or little) hierarchical structure, denying (and thereby actually affirming) the implicit structures that have to exist. The first option also makes the lines of responsibility within the organisation more explicit, and, therefore, it should not only lead to a more efficient organisation, but also to a more ethical one (see Cilliers 2000a).

6 Network models A final issue to consider briefly is the role of network models in understanding complex systems. Do they have any advantages? I have argued previously (Cilliers 1998: 18–21) that neural networks provide a better framework for modelling complex systems than rule-based models. This claim needs to be qualified. The argument that network models mimic the kind of structure found in complex systems is still, I feel, a sound one. Network models can self-organise, information is represented in a distributed fashion, and most importantly, structures which are very loose, very rigid, and everything in between can be implemented in the same medium. The qualification, however, lies in the difference between the notions “mimic” and “model”. Despite a huge amount of practical problems (for example, in the training of recurrent networks), it is in principle possible that a neural network can simulate a complex system, but I do not think that the problem of modelling complex systems discussed at the beginning of this paper can really be circumvented. Why not? Although the training of neural nets involves a process where the network is allowed to develop structure without using a pre-existent theory of how it is done, a theoretical framework is nonetheless introduced. This has only a little to do with the selection of the type of network and the training algorithm, but a lot to do with the selection of the data presented to the network. We cannot present the network in training with life, the universe and everything; we have to select. That means that a framework defining the boundary of what is in and what is out, what is important and what is marginal, has to be decided upon before training commences. This does not mean that we cannot generate some very useful network models. It just means that these models will have some a priori constraints which will have to form part of the interpretation of the results. In that respect, neural networks should not be treated in a different way to any other model. There is a rather serious problem, however. Given its distributed nature, the capabilities and limitations of the model are not available in an explicit fashion. We present the network with new data, and then we have to trust



Boundaries, hierarchies and networks in complex systems 

 95

the result – unless the network was “engineered” in such a way that we know what it does. In such a case, however, a model of the system would have had to exist beforehand in order to make the engineering possible. It is sometimes better to work with a simple model where the limitations are explicit than to work with a complex model that may turn out to be a false friend.

7 Conclusions Let me conclude with a few summary remarks highlighting the implications of the arguments presented for a theory of organisations. 1. Complexity theory increases our understanding of complex systems like organisations, but it does not present us with tools which can predict or control the behaviour of a specific organisation accurately. We may be able to learn a lot about the kind of dynamics involved in the functioning of such systems, but we will not be able to use these general principles to make accurate predictions in individual cases. Complexity theory underscores the importance of contingent factors, of considering the specific conditions in a specific context at a specific time. No general model can capture these singularities. Although we cannot escape the use of models, we can also not escape the responsibility involved in using them – a responsibility that can never be shifted onto the models themselves. 2. Organisations do have boundaries, and these boundaries play an important role in determining the identity of the organisation. However, boundaries are not clearly defined in their nature or their place. The vitality of an organisation will be improved if we do not try too hard to define or fix its boundaries, but allow for their constant renegotiation. 3. Since organisations do have structure, they inevitably also have hierarchies. We will not understand the organisation if we do not allow for the role of these hierarchies, but we have to remember that they are often not clearly determined and that they interpenetrate. We have to allow for the important role that could be played by apparently marginal elements, that is, we have to remember that the hierarchies themselves have a complex structure. In a vital organisation, it will be possible to transform existing hierarchies into different ones, but not to eliminate them. A final word can be added concerning the identity of an organisation. It may appear from these arguments that such a notion will be difficult to maintain. I think not. The identity of an organisation cannot be static, but neither should

96 

 Paul Cilliers

it be too fluid. It emerges exactly from the way in which the boundaries and the hierarchies of that organisation are simultaneously maintained and transformed. Acknowledgements: The financial assistance of the National Research Foundation: Social Sciences and Humanities of South Africa towards this research, as well as the support of the Institute for the Study of Coherence and Emergence (ISCE) in Boston, USA, is hereby acknowledged. Opinions expressed and conclusions arrived at are those of the author, and are not necessarily to be attributed to the National Research Foundation or to ISCE.

References Cilliers, P. 1998. Complexity and Postmodernism. Understanding Complex Systems. London: Routledge. Cilliers, P. 2000a. What can we learn from a theory of complexity?. In: Emergence 2(1): 23–33. Cilliers, P. 2000b. Rules and complex systems. In: Emergence 2(4): 40–50. Corning, P.A. 1998. Complexity is just a word! In: Technological Forecasting and Social Change 59: 197–200. Dasgupta, S. 1997. Technology and complexity. In: Philosophica 59(1): 113–139. Emmeche, C. 1997. Aspects of complexity in life and science. In: Philosophica 59(1): 47–68. Juarrero, A. 1999. Dynamics in Action. Intentional Behaviour as a Complex System. Cambridge, MA: MIT Press. Khalil, E.L. & Boulding, K.E. (eds.). 1996. Evolution, Order and Complexity. London: Routledge. Luhmann, N. 1989. Ecological Communication. Chicago: University of Chicago Press. Midgley, G., Munlo, I. & Brown, M. 1998. The theory and practice of boundary critique: developing housing services for older people. In: Journal of the Operational Research Society 49: 467–478. Pattee, H.H. 1973. Hierarchy Theory. The Challenge of Complex Systems. New York: George Braziller. Romme, A.G.L. 1996. A note on the hierarchy-team debate. In: Strategic Management Journal 17: 411–417. Simon, H.A. 1962. The architecture of complexity: hierarchic systems. In: Proceedings of the American Philosophical Society 106(6): 467–82. Thrift, N. 1999. The place of complexity. In: Theory Culture and Society 16(3): 31–69.

Paul Cilliers

Why we cannot know complex things completely Despite wonderful advances in the mathematics and science of complexity, despite clever modelling techniques, despite fantastic computing machines, and, above all, despite its being somewhat fashionable, I wish to argue that complexity theory will not lead to a grand science that will solve many of those difficult outstanding problems of science and philosophy. Rather, I wish to argue that the study of the characteristics of complex dynamic systems is showing us exactly why limited knowledge is unavoidable – or, to be more precise, why knowledge has to be limited. The study of complexity, in other words, is not going to introduce us to a brave new world in which we will be able to control our destiny; it confronts us with the limits of human understanding. (This position is, to my mind, supported by the large-scale problems experienced in the so-called new economy, especially in the context of the disappointing performance of so many over promoted software companies.) Before this position is elaborated, it should be made clear what is not being claimed. In the first place, the argument has nothing to do with the dispute about the so-called end of science; see among others Durlauf (1997), Horgan (1996), and Lindley (1993). No claim is made that we have already discovered most of the fundamental scientific theories, and that new science will only be derivative. This is a hubristic position that glorifies the present. Nor is a claim made that fundamental advances will not be made in those sciences (like the human sciences) normally perceived as too complex, or where empirical results have so far been disappointing. This is a defeatist position, often triggered by an over-evaluation of so-called hard scientific results or methods. There is no reason not to believe that there is much to be learned. The argument is just that, as far as complex systems are concerned, our knowledge will always be contextually and historically framed. It is also not claimed that there is something wrong with modelling complex systems. Computational and mathematical models of different kinds are doing wonderful things, and new avenues should be pursued all the time. However, we should be careful about the claims made about the “knowledge” we gain from many of these models. The models are often as complex as that being modelled, and thus do not always lead to deeper understanding of the systems at stake. In order to gain “knowledge” from complex models they have to be interpreted, and these interpretations will always involve a reduction in complexity. Thus the Originally published in Emergence, 2002, 4(1/2): 77–84. © 2002 Lawrence Erlbaum Associates, Inc.

98 

 Paul Cilliers

main argument is not that there is something metaphysically unknowable about complex systems, but rather that we cannot “know” a system in all its complexity despite the fact that we may be able to model its behaviour on a computer. (This allowance that good models of complex systems may exist is a generous one. Most models of complex systems are used to display general complex behaviour, not to model specific, empirical complex systems. This state of affairs may remain so, again not for metaphysical reasons, but because the behaviour of complex models will be as unpredictable as that of the systems they model.) We are returned to the old philosophical problem concerning the relationship between our descriptions of the world and the world itself.

1 Ontology vs. epistemology The traditional way of dealing with this problem is to distinguish between epistemological and ontological issues. Epistemology has to do with the way in which we understand and describe the world, and ontology with the way the world is. One can therefore talk of epistemological complexity (how complex are our descriptions?) and ontological complexity (how complex are things really in themselves?). Using this distinction, one could deal with the problem of our knowledge of the world in the following way: The world itself is not complex, it just is. There is nothing mystical about complex systems. It is merely that we cannot keep track of all the millions of nonlinear interactions when we have to describe it. Complexity is therefore only an epistemological matter. This is how McIntyre (1998: 28) describes this position: [Complex systems, like human systems] are not complex “as such” but only complex as described and defined by a given level of inquiry. What is the nature of our interest in human behaviour? What sort of questions do we ask about it? That is what will determine the level of complexity that we are dealing with when we seek to understand certain features of human interaction. For the subject matter of social science is not a “natural kind” just sitting out there waiting for us to discover it. A subject matter is created only when we begin to ask questions about features of reality that are puzzling us. Thus, on this interpretation, complexity is derivative rather than inherent.

This argument has the advantage that it demystifies complexity somewhat. For example, we do not have to let go of causality in order to acknowledge complexity. The world is not dependent on our descriptions. McIntyre (1998: 28), however, uses the ontological/epistemological distinction to make another point, namely that this would mean “that there is no ‘fundamental’ limit to our understanding of ‘complex’ systems.”



Why we cannot know complex things completely 

 99

Once one accepts that complex systems are only complex ‘as described,’ there is always the possibility that some alternative description – some ‘redescription’ – of the system will yield regularities that will be simpler and can be handled by science [...] The job of science, then, is to search for those descriptions of the phenomena that will unlock the regularities that are behind the surface noise of complexity (McIntyre 1998: 29).

This argument is in general a useful one, but on certain points somewhat problematic. At heart it is an instrumentalist position, made explicit by the claim that “in attempting to understand reality, we have many descriptive tools at our disposal [...] There may be one world, but there are an infinite number of alternative ways of describing it” (McIntyre 1998: 29). Despite his attempts to deny it (“nature rules out infinitely many descriptions that are inconsistent with it”), this position will have serious difficulties in defending itself against the accusation of relativism. These difficulties are the result, I would argue, of a too simplified, or perhaps even contradictory, understanding of the relationship between our description of the world and the world itself. On the one hand McIntyre (1998: 28) separates the two quite clearly, accusing others (e.g. Hayek) of failing to distinguish “sharply” between ontology and epistemology, but at the same time it wants to affirm that science is about reality. This is to have your cake and eat it. In the end such a sharp distinction between epistemological and ontological issues cannot be maintained. Even if we acknowledge that our descriptions of the world are not perfect, we would like to maintain that they are not merely instruments, but that they enhance our knowledge of the world as it is. There is a complex dialectical relationship between the world and our descriptions. When we try to understand the world we are always dealing with ontological and epistemological issues simultaneously. To maintain a clear distinction between the two is the essence of metaphysics.

2 What is knowledge? If it is argued that epistemology and ontology cannot be kept apart systematically, what becomes of the notion of “knowledge”? This is one of the words that have become commodified in our times. We talk of a “knowledge industry” and of “knowledge management.” These terms create the impression that knowledge is something in which we can trade, independently of the subject that has the knowledge. In this way knowledge is reified, turned into something that “exists”, that can be put on a disk or a website. Of course, there are many things we can put on a disk, but perhaps one should reserve the terms “data” or even “information”

100 

 Paul Cilliers

for this. The term “knowledge,” I suggest, should be reserved for information that is situated historically and contextually by a knowing subject. Knowledge is that which has meaning, it is the result of a process of interpretation (see also Cilliers 2000). There is nothing new about linking knowledge and the knowing subject. It may also appear as if it reinstates an independent epistemological level. However, from the perspective of complexity theory, these issues look a little different. In the first place, the subject is not an independent whole, not a free-floating ego that makes “subjective” observations or decisions. It is a complex thing in itself, constituted through the web of relationships with others and the world. The subject itself can therefore only be understood as something contextualised through and through (see Cilliers & De Villiers 2000). Secondly, complexity theory also helps us to understand the process by which things and concepts acquire meaning differently. I argue in detail elsewhere (Cilliers 1998: 58–88) that we cannot maintain a representational theory of meaning. Meaning is not something complete and abstract, linked to the sign that represents it, but is the result of a dynamic interaction between all the meaningful components in the system (Cilliers, 1998: 37–47), itself a complex process. If meaning is relational, not representational, there are potentially an infinite amount of relationships at stake each time the meaning of something is generated. Complex systems are open systems; interactions take place across their boundaries. However, if an infinite number of interactions have to be considered, the production of meaning will be indefinitely postponed. This, we know, is not the case. Meaning is generated in real time. How is this possible? Because meaning is constituted in a specific context where some components are included and others not. It would not be possible to have any real meaning if the number of relationships were not limited. In other words, for meaning or knowledge to exist at all, there have to be limits. We cannot comprehend the world in all its complexity. We have to reduce that complexity in order to generate understanding. This is not some terrible fate that befell human subjects, it is merely the result of having to deal with the world in real time with finite means. To summarise: We are simultaneously in the world and reflecting on the world. These processes are intertwined and involve the interaction of an infinite number of factors. The knowing subject is, however, contextualised. The context limits the number of factors, and thereby makes meaning possible. The context can change, of course, and thereby involve other factors. However, the new context will involve new limits. We cannot have knowledge without limits. An interesting question that I will not pursue here is whether we can have knowledge of emergent properties. Perhaps the answer is no!



Why we cannot know complex things completely 

 101

3 Limits and boundaries Talking about the limits or boundaries of complex systems is not an easy task. On the one hand we acknowledge that complex systems are open, that they exchange information (or matter and energy) with their environment. This would tend to underplay the role of the boundary. On the other hand, the very notion of “system” presupposes the existence of a boundary to the system. For the system to be identified as such it has to be distinguished from what is not part of the system, that is, the environment or other systems. Both positions can be problematic. One can, and often should, emphasise the interrelatedness of systems. Often the boundaries of systems are constructions that we impose in order to reduce the complexity. This can lead to oversimplifications, to reductive descriptions of the system. However, if boundaries become too vague, we end up with a kind of holism that does not allow much to be said. We cannot consider life, the universe and everything in its totality all the time. We need limits in order to say something. One can, nevertheless, also overemphasise the role played by the boundaries of a system. To my mind, this is the case with Luhmann’s position in his elaboration of Maturana and Varela’s arguments concerning autopoiesis (for an excellent discussion of these positions, see Rasch 2000). The claim that a system can only make representations in terms of its own resources results in what Luhmann calls “operational closure”. Thus the legal system, for example, can only operate in legal terms. It organises legal procedures, so there is change in the system, but this change is always in terms of evolutionary processes taking place within the system. This position makes it difficult to see how any intervention in the dynamics of the system can take place. The claim for operational closure leads to a self-sufficient conceptualisation of the system. Since the “knowledge” contained within the system has to be constructed in terms of the internal resources of the system, it is again difficult to see how this position can escape the charge of relativism. Perhaps one can evade some of these complexities by making a distinction between boundaries and limits. Since this distinction attempts to reduce complexity, it will, like most distinctions, come under pressure in certain contexts. However, it also allows us to say new things about complex systems. The suggestion is that a boundary is something with two sides, like the boundary of a country. A limit, on the other hand, we can only know from one side, that is, we cannot know what is beyond it. Let us examine the two concepts briefly. The notion of a boundary seems fairly clear cut. It refers to that which contains and constrains a system. The skin is the boundary of the body; a dam ends

102 

 Paul Cilliers

where the water ends. However, more often than not it is extremely difficult to determine where exactly the boundary is. Think, for example, of the boundary as those elements of a system that interact directly with the environment of the system. If one conceives of a complex system as something constituted through a rich interaction of all its components, there is only a short route between any element and the environment. In a sense, the whole system is close to the boundary, the boundary is “folded in”, and one is never quite sure whether one is dealing with the inside or the outside of the system. The boundary is there, but one cannot pin it down. At the same time, one should also not think of the boundary as something confining the system, but rather as something that constitutes the system. By differentiating the system from the environment, and simultaneously allowing for the transcending of the boundary, the system can be and become what it is. A good example to illustrate this principle is that of the eardrum. It separates the inner and outer ear, but exists in order to let sound come through. Moreover, it would not have been possible for the sound to come through if the boundary were not there. (See Cilliers 2001, for a further discussion of boundaries. There the distinction between limits and boundaries is not made explicit.) The notion of the limit is a difficult one (and needs a more detailed discussion than will be attempted here). For example, if we concede that there are limits to our knowledge, how do we know when we have reached that limit? It is exactly this claim – that we have reached the limit and that we know it – that leads to the “end of science” argument. Furthermore, how do we talk about limits if we do not know what lies beyond? Do we maintain a Wittgensteinian silence, or do we make assumptions about what lies beyond – a move that will return us to the traditional world of metaphysics? Perhaps complexity theory can help us deal with this problem in somewhat different terms. Without falling back into a crude dichotomy between epistemology and ontology, we could argue that the world itself does not have limits, only boundaries. Limits exist in our understanding and descriptions of the world (keeping in mind that these descriptions are not arbitrary constructions, but that they are constrained by reality, that they are “about” the world). The limits are not transcendentally given, but a result of having to deal with complexity with finite means. If this is the case, then there is no reason that the limits cannot be shifted. There will always be limits, thus there will always be something that eludes our understanding of a complex system, but from different perspectives, following different strategies, these limits will be different. To keep on confronting these limits is what science – and life – is all about. Nevertheless, they will remain limits in the sense that we cannot say what it is that eludes us. We cannot calculate what it is that escapes our grasp.



Why we cannot know complex things completely 

 103

What we need, therefore, are ways of dealing with that which we can not calculate, of coping with our ignorance. There is a name for this. It is called “ethics”, and no amount of complexity theory will allow us to escape it.

References Cilliers, P. 1998. Complexity and Postmodernism: Understanding Complex Systems. London: Routledge. Cilliers, P. 2000. Knowledge, complexity, and understanding. In: Emergence 2(4): 7–13. Cilliers, P. 2001. Boundaries, hierarchies and networks in complex systems. In: International Journal of Innovation Management 5(2): 135–47. Cilliers, P. & De Villiers, T. 2000. The complex ‘I’. In: Wheeler, W. (ed.). The Political Subject. London: Lawrence & Wishart. Durlauf, S.N. 1997. Limits to science or limits to epistemology. In: Complexity 2(3): 31–7. Horgan, J. 1996. The End of Science. Menlo Park, CA: Addison-Wesley. Lindley, D. 1993. The End of Physics: The Myth of a Unified Theory. New York: Basic Books. McIntyre, L. 1998. Complexity: A philosopher’s reflections. In: Complexity 3(6): 26–32. Rasch, W. 2000. Immanent systems, transcendental temptations, and the limits of ethics. In: Rasch, W. & Wolfe, E. (eds). Observing Complexity: Systems Theory and Postmodernity. Minneapolis, MN: University of Minnesota Press. Rasch, W. & Wolfe, E. (eds). 2000. Observing Complexity: Systems Theory and Postmodernity. Minneapolis, MN: University of Minnesota Press.

Paul Cilliers

Knowledge, limits and boundaries There’s danger on the edge of town (The Doors).

1 Introduction1 I understand this special edition to be primarily concerned with the problem of knowledge. How do we understand the world, what is “scientific” knowledge, and to what extent is this knowledge limited by the fact that the world in which we live is complex? The problems associated with the status of our knowledge of the world have been central to philosophy all along. Here I will focus on the way in which the acknowledgement of complexity transforms some of the traditional conceptions of (especially scientific) knowledge. I will also examine the notions of boundaries and limits, arguing that these notions are not problems we have to get out of the way, but that they are inevitable as soon as we start talking of “knowledge”.

2 The problem As science confronted more and more complex problems, various manifestations of the problem of limits appeared: relativity theory introduced the speed of light as absolute limit, quantum theory made us aware of inescapable uncertainty, and Gödel and Turing brought us face to face with limits of deductive logic. Influential as these ideas were, they are all still largely part of an attempt to describe the world in purely objective terms. The speed of light is a constant of nature; undecidability is an inevitable characteristic of formal systems. Limits, therefore, are natural things. The perspective introduced by complexity is rather different. Here the argument is the following: a complex system is constituted through a large amount of nonlinear interactions and cannot be separated from its environment. It is thus not possible (in practice or in principle, the argument goes), to give a complete,

1 This paper is based on material used in Cilliers (2000; 2001). Originally published in Futures 37, 2005: 605–613. © 2004 Elsevier Ltd. All rights reserved.

106 

 Paul Cilliers

analytical and formal description of a complex system. We have to frame the system in a certain way when we want to describe it. There is, however, no “pure” position outside the system we can assume in order to determine the parameters of this frame (unless we are dealing with well-defined and closed systems which are normally at most complicated, and not complex). The result is that we cannot determine the limits of our description objectively. Limits are determined by strategic considerations. Even though this does not necessarily mean that limits are arbitrary, it does mean that considerations of power and expediency affect the way in which we understand the world. These ideas are disconcerting for those believing in science as something that has to maintain some form of objectivity. It seems to open the door to a relativism that would destroy the notion of “scientific knowledge”. In what follows I will try to take the argument from complexity seriously, but in such a way that we do not fall prey to relativism. It will, however, necessitate a re-examination of what we understand as “knowledge”.

3 The problem of knowledge 3.1 What qualifies as knowledge? The immense usefulness of mathematics has led to an understanding of scientific knowledge that is linked to formal models: one has “knowledge” of a subject to the extent that it can be described in terms of a set of (objective) rules. The knowledge is contained in an algorithm, and the complexity of the knowledge is equivalent to the length of the algorithm. This has resulted – according to, e.g. Robert Rosen (1996) – in a shift towards methodology, and away from the content of scientific knowledge. I think it is important to ask if one can talk of “formal knowledge” at all. Perhaps one should reserve the notion “knowledge” for something that has meaning, that is meaningful to a human subject (I will return to this below). Knowledge would then refer to something with empirical content, not an abstract algorithm. The notion of empirical content is also important if we want to make some ontological claims beyond a pure and formal, Platonic kind of epistemology. The issue is interesting particularly because we can do things technologically that we do not understand. For example, I can build a neural network that can perform some kind of pattern recognition, like recognising faces. I know how the network works, but I do not know how it solves the problem. Can I now claim that I “know” how to recognise faces? I do not think so. This would be to confuse data



Knowledge, limits and boundaries 

 107

with knowledge. Perhaps “knowledge” is a concept used with too much ease, as in “knowledge management”. We cannot “know” a complex thing in all its complexity, we reduce the complexity in order to be able to say something about it within the finite means of our comprehension. Knowledge and data-reduction are intertwined. We can have knowledge because we draw boundaries. Let us examine these concepts in a little more detail. The issues around knowledge – what can we know about the world, how do we know it, what is the status of our experiences – have been central to philosophical reflection for ages. Answers to these questions, admittedly oversimplified here, have traditionally taken one of two forms. On the one hand there is the belief that the world can be made rationally transparent, that with enough hard work knowledge about the world can be made objective. Thinkers like Descartes and Habermas are often framed as being responsible for this kind of attitude, and it goes under numerous names including positivism, modernism, objectivism, rationalism and epistemological fundamentalism. On the other hand, there is the belief that knowledge is only possible from a personal or cultural- specific perspective, and that it can therefore never be objective or universal. This position is ascribed, correctly or not, to numerous thinkers in the more recent past like Kuhn, Rorty and Derrida, and its many names include relativism, idealism, post-modernism, perspectivism and flapdoodle. Relativism is not a position that can be maintained consistently,2 and of course the thinkers mentioned above have far more sophisticated positions than portrayed in this bipolar caricature. There are also recent thinkers who attempt to move beyond the fundamentalist/relativist dichotomy, but it seems to me that when it comes to the technological applications of theories of knowledge, there is an implicit reversion to one of these traditional positions. For those who want to computerise knowledge, knowledge has to be objective. It must be possible to gather, store and manipulate knowledge without the intervention of a subject. The critics of formalised knowledge, on the other hand, usually fall back on arguments based on subjective or culture-specific perspectives to show that it is not possible, that we cannot talk about knowledge independently of the knowing subject.

2 If relativism is maintained consistently, it becomes an absolute position. From this one can see that a relativist is nothing else but a disappointed fundamentalist. However, this should not lead one to conclude that everything that is called postmodern leads to this weak position. Lyotard’s seminal work, The Postmodern Condition (1984), is subtitled A Report on Knowledge. He is primarily concerned with the structure and form of different kinds of knowledge, not with relativism. An informed reading of Derrida will also show that deconstruction does not imply relativism at all. For a penetrating philosophical study of the problem, see Against Relativism (Norris 1997).

108 

 Paul Cilliers

4 Complexity and understanding An understanding of knowledge as constituted within a complex system of interactions3 would, on the one hand, deny that knowledge can be seen as atomised “facts” that have objective meaning. Knowledge comes to be in a dynamic network of interactions, a network that does not have distinctive borders. On the other hand, this perspective would also deny that knowledge is something purely subjective, mainly because one cannot conceive of the subject as something prior to the “network of knowledge”, but rather as something constituted within that network. The argument from complexity thus wants to move beyond the objective/subjective dichotomy. The dialectical relationship between knowledge and the system within which it is constituted has to be acknowledged. The two do not exist independently, thus making it impossible to first sort out the system (or context), and then to identify the knowledge within the system. This co-determination also means that knowledge and the system within which it is constituted is in constant transformation. What appears to be uncontroversial at one point may not remain so for long. The points made above are just a restatement of the claim that complex systems have a history, and that they cannot be conceived of without taking their context into account. The burning question at this stage is whether it is possible to do that formally or computationally. Can we incorporate the context and the history of a system into its description, thereby making it possible to extract knowledge from it? This is certainly possible (and very useful) in the case of relatively simple systems, but with complex systems there are a number of problems. These problems are, at least to my mind, not of a metaphysical, but of a practical nature. The first problem has to do with the non-linear nature of the interactions in a complex system. From this it can be argued (Cilliers 1998: 9–10) that complexity is incompressible. There is no accurate (or rather, perfect) representation of the system, which is simpler than the system itself. In building representations of open systems, we are forced to leave things out, and since the effects of these omissions are nonlinear, we cannot predict their magnitude. This is not an argument claiming that reasonable representations should not be constructed, but rather an argument that the unavoidable limitations of the representations should be acknowledged. This problem  – which can be called the problem of boundaries  – is compounded by the dynamic nature of the interactions in a complex system. The

3 Complex systems are discussed in detail in Cilliers (1998).



Knowledge, limits and boundaries 

 109

system is constituted by rich interaction, but since there are an abundance of direct and indirect feedback paths, the interactions are constantly changing. Any activity in the system reverberates throughout the system, and can have effects that are very difficult to predict – once again as a result of the large amount of nonlinear interactions. I do not claim that these dynamics cannot be modelled. It could be possible that richly connected network models can be constructed. However, as soon as these networks become sizeable, they become extremely difficult to train. It also becomes rather hard to figure out what is actually happening in them. This is no surprise if one grants the argument that a model of a complex system will have to be as complex as the system itself. Reduction of complexity always leads to distortion. What are the implications of the arguments from complexity for our understanding of the distinction between data and knowledge? In the first place it problematises any notion that data can be transformed into knowledge through a pure, mechanical and objective process. It, however, also problematises any notion that would see the two as totally different things. There are facts that exist independently of the observer of those facts, but the facts do not have their meaning written on their faces. Meaning only comes to be in the process of interaction. Knowledge is interpreted data. This leads us to the next big question: what is involved in interpretation, and who (or what) can do it?

5 Knowledge and the subject Before talking about specific forms of knowledge (scientific, algorithmic, knowledge which can be managed) we have to deal with the question of how the human subject deals with knowledge in the first place. Given the complexities of that which we wish or have to know, how does the subject come to forms of understanding, and what is the status of knowledge as understood by a specific subject? This issue has been pursued by many philosophers, especially in the discipline known as hermeneutics. However, I am not aware that this has been done in any depth in the context of complexity theory.4 How does one perceive of the subject as something that is not atomistically self-contained, but is constituted through dynamic interaction? Moreover, what is the relationship between such a subject

4 An important contribution was made by reinterpreting action theory from the perspective of complexity (Juarrero 1999). Some preliminary remarks, more specifically on complexity and the subject, are made in Cilliers and De Villiers (2000).

110 

 Paul Cilliers

and its understanding of the world? A deeper understanding of what knowledge is, and how to “manage” it, will depend heavily on a better understanding of the subject. This is a field of study with lots of opportunities. Apart from calling for renewed effort in this field, I only want to make one important remark. It seems that the development of the subject from something totally incapable of dealing with the world on its own into something that can begin to interpret  – and change  – its environment is a rather lengthy process. Childhood and adolescence are necessary phases (sometimes the only phases) in human development. In dealing with the complexities of the world there seems to be no substitute for experience (and education). This would lead one to conclude that when we attempt to automate understanding, a learning process will also be inevitable. This argument leads one to support computing techniques, which incorporate learning (like neural networks) rather than techniques which attempt to abstract the essence of certain facts and manipulate them in terms of purely logical principles. Attempts to develop a better understanding of the subject will not only be helpful in building machines that can manage knowledge, it will also help humans to better understand what they do themselves. We should not allow that the importance of machines (read computers) in our world leads to a machine-like understanding of what it is to be human. Knowledge as something that has meaning for a subject will always be contextualised. It will form part of our experience of the world, and will therefore be influenced by relationships of power. Knowledge cannot be symmetrical, pure, complete or ahistorical. It is always bounded. The status and function of boundaries, when dealing with complex systems, therefore need closer analysis.

6 The nature of boundaries In order to be recognisable as such, a system must be bounded in some way. However, as soon as one tries to be specific about the boundaries of a system, a number of difficulties become apparent. For example, it seems uncontroversial to claim that one has to be able to recognise what belongs to a specific system, and what does not. But complex systems are open systems where the relationships amongst the components of the system are usually more important than the components themselves. Since there are also relationships with the environment, specifying clearly where a boundary could be, is not obvious. One way of dealing with the problem of boundaries is to introduce the notion of “operational clo-



Knowledge, limits and boundaries 

 111

sure”.5 For a system to maintain its identity, it must reproduce itself (internally). These arguments often follow from the work by Maturana and Varela on autopoiesis. Zeleny (in Khalil & Boulding, 1996: 123) defines an autopoietic system as [...] a system that is generated through a closed organisation of production processes such that the same organisation of processes is regenerated through the interaction of its own products (components), and a boundary emerges as a result of the same constitutive processes.

When dealing with complex systems in an “operational” way, there is nothing wrong with this approach. One should be careful, however, not to overemphasise the closure of the boundary. The boundary of a complex system is not clearly defined once it has “emerged”. Boundaries are simultaneously a function of the activity of the system itself, and a product of the strategy of description involved. In other words, we frame the system by describing it in a certain way (for a certain reason), but we are constrained in where the frame can be drawn. The boundary of the system is therefore neither purely a function of our description, nor is it a purely natural thing. We can never be sure that we have “found” or “defined” it clearly, and therefore the closure of the system is not something that can be described objectively. An overemphasis on closure will also lead to an understanding of the system that may underplay the role of the environment. However, we can certainly not do away with the notion of a boundary. Our understanding of boundaries can be given a little more content by considering the following two issues. The first concerns the “nature” of boundaries. We often fall into the trap of thinking of a boundary as something that separates one thing from another. We should rather think of a boundary as something that constitutes that which is bounded. This shift will help us to see the boundary as something enabling, rather than as confining. To quote Zeleny (Khalil & Boulding, 1996) again: All social systems, and thus all living systems, create, maintain, and degrade their own boundaries. These boundaries do not separate but intimately connect the system with its environment. They do not have to be just physical or topological, but are primarily functional, behavioral, and communicational. They are not ‘perimeters’ but functional constitutive components of a given system.

5 The work of Niklas Luhmann provides a good example of this approach. (For a monograph in English, see Luhmann (1989)).

112 

 Paul Cilliers

As an example of this logic, think of the eardrum. It forms the boundary between the inner and the outer ear, but at the same time it exists in order to let the sound waves through. As a matter of fact, if it was not there, the sound waves would not be able to get through at all! If the boundary is seen as an interface participating in constituting the system, we will be more concerned with the margins of the system, and perhaps less with what appears to be central.6 A second boundary issue concerns the “place” of the boundary. The propensity we have towards visual metaphors inclines us to think in spatial terms. A system is, therefore, often visualised as something contiguous in space. This tendency is reinforced by the prevalence of biological examples of complex systems. We think of systems in an “organistic” way. Social systems are obviously not limited in the same way. Parts of the system may exist in totally different spatial locations. The connections between different components could be seen as virtual, and therefore the system itself may exist in a virtual space. This much should be self-evident to most inhabitants of the global village, but there are two important implications to drawn from this. The first is that non-contiguous sub-systems could be part of many different systems simultaneously. This would mean that different systems interpenetrate each other, that they share internal organs. How does one talk of the boundary of the system under these conditions? A second implication of letting go of a spatial understanding of boundaries would be that in a critically organised system we are never far away from the boundary. If the components of the system are richly interconnected, there will always be a short route from any component to the “outside” of the system. There is thus no safe “inside” of the system, the boundary is folded in, or perhaps, the system consists of boundaries only. Everything is always interacting and interfacing with others and with the environment; the notions of “inside” and “outside” are never simple or uncontested. In accepting the complexity of the boundaries of complex systems, we are committed to be critical about how we use the notion since it affects our understanding of such systems, and influences the way in which we deal with them. The notion of “boundary critique” is not a new one (see Midgley et al. 1998), but in this critique we have to keep the enabling nature of boundaries in mind, whilst simultaneously trying to displace (deconstruct) them. The argument for an understanding of boundaries and constraints as being enabling, and the observation that in a complex system one is never far away

6 Although it will not be elaborated on in this text, a number of the ideas presented have a close affinity to arguments from deconstruction. For more detail, see Cilliers (1998), especially chapter three.



Knowledge, limits and boundaries 

 113

from a boundary,7 has certain implications. One can, for example, deal with a system as if it is a pre-given and objectively defined entity. Then the boundaries will be clear to the extent that one could say that they are “natural”. There are systems like this, but they are usually neither complex nor interesting – systems like machines. If one acknowledges the complexity of a system, it becomes more difficult to talk about “natural” boundaries. Boundaries are still required if we want to talk about complex systems in a meaningful way – they are in fact necessary, as argued above  – but there are strategic considerations at stake when drawing them. These considerations may include subjective, or intersubjective components, but this does not mean that they are arbitrary. A complex system has structure and patterns that would render some descriptions more meaningful than others, but the point is that we do not have an a priori decision procedure for determining when we are dealing with something “more meaningful”. The contingent and historic nature of complex systems entails that our understanding of the system will have to be continually revised; the frames of our models will have to change. The boundaries of complex systems cannot be identified objectively, finally and completely. This supports the argument that our knowledge of complex systems cannot be reduced to formal algorithms, but has to incorporate considerations of what the knowledge is for. The criteria used to evaluate the knowledge are not independent things; they co-determine the nature of the knowledge (see Rosen 1996). Knowledge cannot be abstract and complete – we cannot “know” something like that. For us to have knowledge about something, it has to be limited. I want to stress that this does not imply a subjective relativism. It merely acknowledges the inevitability of choice when trying to understand a complex system, and it is exactly at this point that we encounter the ethical domain.

7 The challenge of the limit In Nicholas Roeg’s remarkably visionary film The Man Who Fell to Earth (1976), an alien using the name Thomas Jerome Newton (superbly played by David Bowie), tries to understand human culture by watching television  – usually a whole bunch of screens at the same time. Despite the immense amount of data available to him, he is not able to understand what is going on directly. It is only through

7 For the sake of clarity, I propose that we talk about the limits of knowledge and of the boundaries of systems.

114 

 Paul Cilliers

the actual experience of political complexities, as they unfold in time, that he begins to understand. By then he is doomed to remain earthbound. It is only from a situated position that we can have knowledge, never from an abstract or divine one – and the computer will not be able to replace God in this argument either. It should be stressed, however, that when we conclude that limits are necessary for the generation of meaning, then there is no need to feel despondent when we encounter them. They form an integral part of the development and transformation of knowledge. The fact that it becomes a little more difficult to talk about “objective” knowledge should also not lead to despair, but to humility. The issues in need of urgent attention seem, to me at least, to be the following: –– Developing the notion of “scientific” knowledge in order to go beyond an abstract objectivity without falling prey to relativism. –– Elaborating the ethical considerations inherent to all forms of knowledge. Complexity theory may turn out to be central to these explorations. It may also be just what we need in order to start building bridges between the “two cultures”.

References Cilliers, P. 1998. Complexity and Postmodernism. Understanding Complex Systems. London: Routledge. Cilliers, P. 2000. Knowledge, complexity and understanding. In: Emergence 2(4): 7–13. Cilliers, P. 2001. Boundaries, hierarchies and networks in complex systems. In: International Journal of Innovation Management 5(2): 135–147. Cilliers, P. & De Villiers, T. 2000. The complex ‘I’. In: Wheeler, W. (ed.). The Political Subject. London: Lawrence and Wishart. 226–245. Juarrero, A. 1999. Dynamics in Action. Intentional Behaviour as a Complex System. Cambridge, MA: MIT Press. Khalil, E.L. & Boulding, K.E. (eds.). 1996. Evolution, Order and Complexity. London: Routledge. 133. Luhmann, N. 1989. Ecological Communication. Chicago: University of Chicago Press. Lyotard, J.F. 1984. The Postmodern Condition: A Report on Knowledge. Manchester: Manchester University Press. Midgley, G., Munlo, I. & Brown, M. 1998. The theory and practice of boundary critique: developing housing services for older people. In: Journal of the Operational Research Society 49: 467–478. Norris, C. 1997. Against Relativism. Philosophy of Science, Deconstruction and Critical Theory. Oxford: Blackwell Publishers. Rosen, R. 1996. On the limitations of scientific knowledge. In: Casti, J.L. & Karlqvist, A. (eds.). Boundaries and Barriers. On the Limitations of Scientific Knowledge. Reading, MA: Addison Wesley. 199–214.

Part 1: Single-authored Papers Theme 2: Complexity and Philosophy

Paul Cilliers

Postmodern knowledge and complexity (or why anything does not go) 1 Introduction The term “postmodern” is used in a number of ways, many of them quite confusing. I will not attempt to clarify this confusion here since I feel that this self-indulgent search for a clear identity of what postmodernism is, is still part of the heritage of modernism. In a sense the term has outlasted its usefulness but as yet there seems to be no good alternative. I will therefore accept the term without argument and I will also accept that we do indeed live in the postmodern era, at least in the sense that we have given up the project of modernity. What I will explore in more detail is what this means in terms of the way in which we understand our world and the way in which we perform judgements on aspects of our world. The implications of a postmodern theory of knowledge and understanding are disconcerting for many. They feel that it is destabilising to the extent that we are not capable of making any judgements, whether they be aesthetic, moral or scientific. I wish to provide an approach to postmodernism that does not imply that “anything goes”, but rather provides us with a strategy for coping with the complexities we have to deal with when we wish to talk about our world. Our traditional epistemological systems  – predicate logic, the “scientific method”, rule-based language with transparent semantics  – have coped adequately with those descriptions of a world in which the phenomena encountered can be reduced to clearly demarcated categories with finite contents. The limitations of these systems have become more and more apparent with the failure of traditional approaches to provide adequate descriptions of complex phenomena like language, art, society and life (in a wide sense that includes biological aspects). Postmodernism, in the sense that I will use the concept, has an inherent sensitivity for complexity. It is sensitive to the fact that complex phenomena cannot be given simple descriptions. The price paid for this sensitivity is, in conventional terms, quite high: an abandoning of the quest for universal criteria of truth, knowledge and judgement. This may generate a feeling of loss, but I wish to argue that the nostalgia for such criteria has prevented us from engaging with our world in a responsible way.

Originally published in the South African Journal of Philosophy, 1995, 14(3): 124–132. © South African Journal of Philosophy.

118 

 Paul Cilliers

I cannot deal with the various aspects of complexity and complex systems in detail, but I will introduce some of the important concepts briefly.1 I will then point out the affinities between my description of complex systems and postmodern theory. The discussion of the conditions for knowledge in the postmodern era (as suggested by Lyotard) will show that the lack of universal criteria – or of a coherent metanarrative – does not imply that all forms of judgement are relativised. I will close with a few remarks on the status of the judgements we make when we evaluate, for example, works of art.

2 Characterising complexity When dealing with complex systems we cannot employ traditional analytical methods. The nature of complex systems, like language, the brain or social systems, is determined by large numbers of elements that interact richly in a nonlinear fashion. The complexity referred to here is therefore not equivalent to either chaos or to entities we cannot grasp merely because of their magnitude (like trying to understand a beach in terms of the grains of sand). Complex systems are structured in the sense of containing meaningful patterns of interaction between their constituent components. These interactions are not fixed in time – complex systems are dynamic. The kind of complexity exhibited by a snowflake is therefore also excluded since it shows no real dynamic behaviour, except perhaps decay. Although not quite accurately, I will refer to the “snowflake” kind of complexity as “passive complexity”. The inaccuracy is caused by the fact that the dynamical, interactive kind of complexity that we find in living systems cannot really be described as “active” since their behaviour has to be described as is suspended somewhere between active and passive.2 With the growth of sophisticated computer technology, a new interest in the nature, and especially in the modelling of complex systems has developed. Various techniques including neural networks, cellular automata, random Boolean nets and several methods from the study of artificial intelligence (like expert systems) are employed to do this. Before discussing some of them, I wish to list a number of general characteristics of complex systems. This is done in order to develop a feeling for the nature of complexity, to show that although it is

1 They are discussed in detail in Cilliers (1994). 2 This suspension between active and passive is a dynamic quality equivalent to the process described by Derrida’s notion of différance.



Postmodern knowledge and complexity (or why anything does not go)  

 119

very difficult, usually impossible, to give clear descriptions of the behaviour of a complex system, it does not mean that anything metaphysical or transcendental is involved. I will clarify these characteristics with an example. Later I will show how these characteristics are echoed in postmodern theory. i. Complex systems consist of a large number of elements. If the number is small, the behaviour of the elements can often be given a formal description in conventional terms. When the number becomes large, conventional means (e.g. a system of differential equations) not only become impractical, they also cease to assist in our understanding of the system. ii. A large number of elements is not sufficient. The grains of sand on a beach do not constitute an interesting complex system. The elements have to interact, and this interaction must be dynamic. A complex system changes with time. The interactions do not have to be given a physical interpretation. It merely means that information is transferred from one element to another. iii. The interaction is fairly rich, that is, any element in the system influences, and is influenced by, quite a few other ones. The exact level of .interaction, however, is not important. If there are enough elements in the system of which some are redundant, a number of sparsely connected elements can perform the same function as one richly connected element. iv. The interactions themselves have a number of important characteristics. Firstly, the interactions are nonlinear. Nonlinearity means that the law of superposition does not hold. In a linear system one can analyse the components separately and then add the results together. A large system of linear elements can be collapsed into an equivalent system that is very much smaller. Nonlinearity denies this possibility. Nonlinearity also guarantees that small causes can have large results and vice versa. This is a precondition for complexity. v. The interactions usually have a fairly short range, that is, elements receive information primarily from their immediate neighbours. Long-range interaction is not impossible, but practical constraints usually force this consideration. This does not preclude wide-ranging influence. Since the interaction is rich, the route from one element to any other is usually short. What it does mean, is that the influence gets modulated along the way. It can be enhanced, suppressed or altered in a number of ways. vi. There are loops in the interactions. The effect of any activity can feed back onto itself, sometimes directly, sometimes after a number of intervening stages. This feedback can be positive (enhancing, stimulating) or negative (detracting, inhibiting). Both kinds are necessary. The technical term for this aspect of a complex system is recurrence.

120 

 Paul Cilliers

vii. Complex systems are usually open systems, i.e. they interact with their environment. As a matter of fact, it is often difficult to define the border of a complex system. Instead of being a characteristic of the system itself, the scope of the system is usually determined by the purpose of the description of the system, and is thus largely a function of the observer. This process is called “framing”. Closed systems can be quite complex too, but their complexity is usually of the “passive” kind. The final state of a closed system is equilibrium. viii. Complex systems operate under conditions far from equilibrium. There has to be a constant flow of energy to drive the evolution of the system and to ensure its performance. Equilibrium is another word for death. ix. Complex systems have a history. Not only do they evolve through time, but their past is co-responsible for their present behaviour. Any analysis of a complex system that ignores the dimension of time is incomplete, or at most a snapshot of a changing process. x. Each element in the system is ignorant of the behaviour of the system as a whole, it responds only to information that is available to it locally. This point is vitally important. If each element “knows” what was happening to the system as a whole, all the complexity would have to be represented in that element. This would entail either a physical impossibility in the sense that a single element is not capable of such complexity, or constitute a metaphysical move in the sense that “consciousness” of the whole is locked up in the particular. Complexity is the result of a rich interaction between simple elements that only respond to the limited information they are presented with. Complex behaviour emerges from a large ensemble of interactions where complexity is neither determined from the outside nor from the characteristics locked up in specific elements. When we look at the behaviour of a complex system as a whole, our focus shifts from the individual element in the system to the complex structure of the system. The complexity lies in the patterns of interaction, not in the elements. The importance of these ten characteristics will become clear when we examine some examples. Consider a snowflake. From a distance it appears to be a pretty simple object, but when we examine it closer, much closer, as through a microscope, it reveals a remarkably complex structure. The snowflake is arranged hexagonally with each of the six “branches” showing an elaborate and beautifully patterned structure. Although all snowflakes share this form, they all differ in detail. A snowflake consists of a large amount of elements (water molecules) interacting through its crystalline structure. Each molecule is influenced only by local information (there is no external decision as to where the molecule must sit



Postmodern knowledge and complexity (or why anything does not go)  

 121

in the snowflake), but these relationships are fairly fixed. There are however no real feedback loops or evolution (except perhaps decay). As far as its structure is concerned, it is not really an open system. It is in temporary equilibrium, cannot adapt to its environment, and therefore quickly loses its structure. A snowflake, although wondrously complex in appearance, is at most an example of “passive” complexity. Since it is true complexity that we are really interested in, let us look at such systems. The human brain is considered by many to be the most complex object known. Although our understanding of the brain is growing rapidly, many of its most important features remain largely hidden. Similarly, the language with which we communicate daily does not yield to analytical descriptions. These two complex systems – the brain and natural language – need detailed attention and this makes them unsuitable for use as didactic examples. I will therefore elucidate the ten characteristics of a complex system using another example, even if it is somewhat contrived and inaccurate: the economic system. The first step to take is to decide our “distance” from the system. What level of detail are we going to consider? If we stand far away, we could only consider the activity of large financial institutions – banks, large corporations, even countries. Obviously a lot of detail will get lost in the process. If we examine the system in microscopic detail, we may want to keep track of the status of every individual penny. In this case we run the risk of all meaningful patterns being obscured by the buzzing activity on the lowest level. For argument’s sake, let us frame the system in such a way that we consider individual human beings in their capacity as economic agents as the elements in our complex system, and draw the border at the economic activity of a single country. The ten characteristics of complex systems will then manifest themselves in the following way: i. The economically active people in a country certainly comprise a large number of elements, usually several million. ii. The various individuals interact by exchanging money for goods, lending, borrowing and investing. These relationships continually change. iii. Economic agents interact with a large number of the other elements: shops, banks, other agents. Some are more active than others, but their influence on the system is not determined merely by the amount of money they transfer. iv. The interaction is nonlinear. Money can receive compounded interest, small amounts can receive large returns (e.g. buying the right shares at the right time, or vice versa). v. Economic agents primarily interact with others that are in their near vicinity (not necessarily in a spatial sense): their colleagues, partners, friends, etc. They can, however, easily interact with more distant parties via intermediaries like banks or brokers.

122 

 Paul Cilliers

vi. The activity of an agent may eventually reflect back on himself. A good investment can produce good returns (positive feedback) and overspending can result in a shortage in the money supply (negative feedback). Without feedback there would be no economic system –who would invest if there were no returns? Activities can also reflect back after a large number of intermediary steps. The complexities of inflation serve as a good example. vii. The economic system is certainly open. It is virtually impossible to draw its borders. There is continuous interaction with the political system, agriculture (and therefore the climatic conditions), science and technology, international relationships, the stability of the society, etc. There is a constant flow of commodities, products, money and information through the system. viii. Since the economic system is driven by the dynamics of supply and demand it is never in a state of equilibrium. It may be growing or shrinking, swing up or down, but it never stands still, not even in a recession. Even when we talk about a stable economy, the “stability” has to be understood in dynamic terms. ix. Economic systems are greatly influenced by their history. Today’s prices largely depend on yesterday’s. Many important economic trends change fairly slowly over long periods of time. x. Each economic agent has to act on the information that is available to it. It does not know what all the other agents are doing. When an agent, for example, wants to purchase a commodity, a decision is based on a number of “local” factors: how much do I want it, can I afford it, in place of what else will it be purchased, etc. The effects of this act on the inflation rate, the balance of payment, investor’s confidence, interest rates and the like are not normally taken into account even though this act does affect (minutely, but no less than other similar acts) all these factors. Our discussion of the economic system may seem a little thin (which in many respects it is) but there are good reasons for this. We have been describing the elements of the system and their interactions on the level at which they operate. If we want to shift the discussion to more complex economic phenomena (gross national product, stock-market indexes, the gold price, etc.), there is nothing extra we have to add to the system – these phenomena emerge only as a result of the interactions between the various elements of the system. These interactions often take the form of clusters of elements that cooperate with each other and compete with other clusters. A bank, for example, is in a certain sense nothing more than a number of individuals grouped together to perform specific functions. The components of the complex economic system do not consist of different types of things (banks, the state, corporations and individuals), it consists of



Postmodern knowledge and complexity (or why anything does not go)  

 123

individual agents that cluster together to form the larger scale phenomena. The higher order complexities that we hope to get an understanding of do not reside in any of the individual agents, but in the rich pattern of interactions between them. The example of the economic system allows us to make a last significant point. An element in the system may belong to more than one clustering. A person may bank with more than one bank, work for a big corporation and play the stock market on his own. Clustering should not be interpreted in a spatial way or seen as resulting in fixed, hermetically sealed entities. Clusters can grow or shrink, be subdivided or absorbed, flourish or decay. They are dynamic and always interact, not only through the interactions between their various constituents, but also through the members they share with each other. If we wish to model complex systems we must employ methods that mimic their characteristics. The traditional approach was to find the fundamental laws that describe system and then use them to “calculate” the system’s behaviour. This is, however, not adequate since the dynamic nature of these systems denies us the possibility of constructing such laws. The model must be able to incorporate these dynamics and perform the same functions as the system being modelled. The implication of this is unfortunate but clear: the model must be at least as complex as the system being modelled. This is mainly the result of the nonlinear nature of complex systems. If the system can be reduced to something simpler, it was probably not a complex system to start off with. In modelling complexity we are bound to the principle of the “conservation of complexity”. Fortunately rapid progress is being made in a modelling technique known as neural networks or connectionism. This technique is based on mathematical models of neurons (brain cells) connected in vast networks. The workings of these networks are simulated on powerful computers and they have certain capabilities that make them extremely powerful. The most important of these are that they can learn and that they can self-organise their internal structure. I will not discuss them further, but will continue to refer to connectionist models of complex systems.3 Let us now turn to postmodernism and see to what extent our understanding of complexity is compatible with it.

3 For more detail see Cilliers (1991) and (1994). It is also possible to show that there is an equivalence between models of the brain and Saussure’s model of language. See Cilliers (1990).

124 

 Paul Cilliers

3 Complexity and postmodernism The postmodern position is often a controversial one and the literature on postmodernism has proliferated to the extent that it has become difficult to know exactly what position has to be defended (or criticised). For that reason I will focus on a classical text in this field, The postmodern condition (Lyotard 1984). A further reason for this focus is that I do not so much want to provide a survey of postmodernism as to give it a specific interpretation in terms of our conceptual models of complexity. My argument will be that the proliferation of discourses and meaning resulting from postmodern analyses is not merely the result of wilful and disruptive acts by theorists, but an inescapable effect of the complexities of our linguistic and social spaces. The proliferation of information as well as the way in which the media collapse the international public space into a local private space prevents us from coming up with unifying, coherent descriptions of our world. It is along these lines that Lyotard develops his description of the postmodern condition. His aim is to study the conditions of knowledge in developed societies (xxiii). Scientific knowledge, he claims, has traditionally legitimated itself by appealing to a coherent meta discourse which has to perform a general unifying function. All forms of knowledge should be incorporated into one grand narrative. The dream of modernism was that this is possible. I will use the term modem to designate any science that legitimates itself with reference to a meta discourse of this kind making explicit appeal to some grand narrative, such as the dialectics of the Spirit, the hermeneutics of meaning, the emancipation of the rational or working subject, or the creation of wealth (Lyotard 1984: xxiii).

Postmodernism is then defined as an “incredulity towards meta­narratives” (xxiv).4 Instead of looking for a simple discourse that can unify all forms of knowledge,

4 Although a characterisation of postmodernism as an opposition to or a radicalisation of modernism constitutes a useful theoretical or didactic approach to the subject, it is also problematic. It triggers an endless discussion on what is actually the difference between the two as well as an interminable classification game, e.g. is James Joyce modern or postmodern, or Marcel Duchamp, or Dadaism? I feel that a postmodern approach need not legitimate itself by contrasting itself to modernism, since it then remains parasitic on the modem discourse. Postmodernism should move beyond its present narcissistic phase of self-definition and start doing things a little less self-consciously. To incorporate self-consciousness as one of the defining characteristics of postmodernism complicates the matter. If it is used as a technique for criticising present conditions it can produce useful results. Self-indulgence, however, confirms a certain decadent aspect of postmodernism. This decadence could form an important aspect of postmodern art, but when we examine science and the conditions of knowledge, it is more important to come to grips with



Postmodern knowledge and complexity (or why anything does not go)  

 125

we have to cope with a multiplicity of discourses, many different language games, which are determined locally, not legitimated externally. Different institutions and different contents produce different narratives that are not reducible to each other. The narrative function is losing its functions, its great hero, its great voyages, its great goal. It is being dispersed in clouds of narrative language elements -narrative, but also denotative, prescriptive, descriptive, and so on. Conveyed within each cloud are pragmatic valencies specific to its kind. Each of us lives at the intersection of many of these. However, we do not necessarily establish stable language combinations, and the properties of the ones we do establish are not necessarily communicable (Lyotard 1984: xxiv).

This description of knowledge as the result of a multiplicity of local narratives, it must be stressed, is not an argument against scientific knowledge as such, but against a certain understanding of such knowledge. Lyotard denies an understanding that claims that science represents the totality of all true knowledge and argues for a narrative understanding of knowledge, smaller stories that function well within the contexts where they apply (7). Instead of claiming the impossibility of knowledge, “it refines our sensitivity to differences and reinforces our ability to tolerate the incommensurable. Its principle is not the experts homology, but the inventor’s paralogy” (xxv).5 Let me summarise Lyotard’s position. Different groups (institutions, disciplines, communities) each tell their own story about what they know and what they do. Their knowledge is not structured logically or systematically in a general way, but in the form of a narrative that allows them to make sense of what they are doing and to achieve their goals. Since these narratives are all local, they cannot be fused to form a grand narrative that unifies all knowledge. The postmodern condition is characterised by the co-existence of a multiplicity of heterogeneous discourses that can be experienced in different ways. Those who have a nostalgia for a unifying metanarrative – a dream that has been central to Western metaphysics – experience the postmodern condition as fragmented, full

the complexity of contingency. For that we have to put the mirror down. 5 The choice of the word “paralogy” is significant. It is usually employed to designate false reasoning, to mark something as illogical. The word literally means “beside” logic and in this sense Lyotard employs it to show that logical descriptions are not adequate when dealing with the richness and contradictions of contingent complexity. Many stories, even contradictory ones, can be told about single events or phenomena. Lyotard (1984: 60) distinguishes paralogy from “innovation”, which, he claims, is still under the command of the system, used to improve its efficiency. Paralogy is “a move played in the pragmatics of knowledge”, the consequences of which cannot be determined a priori.

126 

 Paul Cilliers

of anarchy and therefore ultimately meaningless. It leaves them with a feeling of vertigo. Those who embrace postmodernism find it challenging, exciting and full of uncharted spaces. It fills them with a sense of adventure. Which of these two evaluations applies is often determined by whether one feels comfortable without fixed points of reference. The choice between the two, therefore, is just as much a result of psychological as of theoretical considerations. Here, however, I will confine myself to theoretical aspects. There is one important theoretical argument often used against postmodernism (e.g. Parushnikova 1992). It claims that if all narratives have only local legitimation, the resulting fragmentation of the social fabric will relativise all knowledge. Since there is no external “check” on any discourse, no local narratives can be seriously criticised. Each discourse becomes independent of all others. Discursive communities will become closed and isolated. Ultimately it would imply that each individual is referred to herself only, with no way of grounding any knowledge objectively. This results in a situation where “anything goes”, a situation that is clearly unacceptable, especially within the context of philosophy of science. I wish to show here that this conclusion is incorrect, at least as far as Lyotard is concerned and that “anything goes” is not a necessary result of postmodernism. The faulty conclusion is a result of a specific understanding of the role of the individual, one that is explicitly rejected by Lyotard (1984: 15). This breaking up of the grand Narratives [...] leads to what some authors analyze in terms of the dissolution of the social bond and the disintegration of social aggregates into a mass of individual atoms thrown into the absurdity of Brownian motion. Nothing of the kind is happening: this point of view, it seems to me, is haunted by the paradisaic representation of a lost ‘organic’ society (15).6

The erroneous understanding of the individual is to see her as an isolated, autonomous agent ... A careful reading of Lyotard shows that his understanding of the individual is formulated in such a way that it counters the idea of fragmentation and isolation that could result from a dismissal of the grand narrative. The following quotation contains a number of important points that will be analysed closely. It also makes the relationships between Lyotard’s model of postmodern society and the models supplied by connectionism clearly visible.

6 Lyotard here explicitly refers to Jean Baudrillard (e.g. Baudrillard 1988) whose analyses of postmodern society often have a very negative flavour.



Postmodern knowledge and complexity (or why anything does not go)  

 127

A self does not amount to much, but no self is an island; each exists in a fabric of relations that is now more complex and mobile than ever before. Young or old, man or woman, rich or poor, a person is always located at ‘nodal points’ of specific communication circuits, however tiny these may be. Or better: one is always located at a post through which various kinds of messages pass. No one, not even the least privileged among us, is ever entirely powerless over the messages that traverse and position him at the post of sender, addressee, or referent. One’s mobility in relation to these language game effects (language games, of course, are what this is all about) is tolerable, at least within certain limits (and the limits are vague); it is even solicited by regulatory mechanisms, and in particular by the self-adjustments the system undertakes in order to improve its performance. It may even be said that the system can and must encourage such movement to the extent that it combats its own entropy; the novelty of an unexpected ‘move’, with its correlative displacement of a partner or group of partners, can supply the system with that increased performativity it forever demands and consumes (Lyotard 1984: 15).

The relevance of the connectionist model is clearly indicated by this passage. The self is understood in terms of a “fabric of relations”, a node in a network, and not in terms of atomistic units standing by and for themselves. Since it is the relationships that are important, and not the node as such, “a self does not amount to much”. Lyotard’s description of the postmodern condition is in fact a description of the network of our society and how it produces and reproduces knowledge. His point is that this network has become too complex for general or overarching descriptions. It has all the characteristics of a complex system. The argument for a multiplicity of discourses is not a wilful move, it is an acknowledgement of complexity. It allows for the explosion of information and the inevitable contradictions that form part of a truly complex network. The critique that the postmodern condition results in many isolated discourses in which anything goes is countered in two ways. In the first place, society forms a network. Although different discourses form “clusters” in this network, they cannot isolate themselves from the network. There are always connections to other discourses. The different local narratives interact, some more than others, but no discourse is fixed or stabilised by itself. Different discourses  – clusters in the network  – may grow, shrink, break up, coalesce, absorb others or be absorbed. Local narratives only make sense in terms of the contrast or difference with their surroundings. This is a self-organising process where meaning is generated through a dynamic process, not the passive reflection of an autonomous agent that can make “anything go”. Instead of being self-sufficient and isolated, discourses are in constant interaction with each other in a battle for territory, so

128 

 Paul Cilliers

to speak, where the provisional boundaries between them are the very stakes in the game. This is what Lyotard calls “the agonistic aspects of society” (16–17).7 The second aspect of the network of society that counters the idea of isolation is the distributed nature of the network. In a connectionist network information is not “represented” by specific nodes, but encoded in patterns distributed over many nodes. Conversely, any specific node forms part of many different patterns. In the social network, discourses are similarly spread over many “selves”. No discourse “represents” some aspect of a metanarrative, there is merely the “pattern of activity” over a large group of individuals exchanging local information. One should not make the mistake of assuming that a person is described by a single node in the network. A human being is far more complex than that. Furthermore, each person is also part of many larger patterns. One can be a mother, a scientist, a consumer, a political activist, an artist and a lover, all at the same time. The social network (similar to connectionist networks) is highly distributed. The argument that claims that postmodernism results in isolation misses the target completely. We only find  – and define  – ourselves within the rich and shifting patterns of social interaction, not as a coherent, self-contained individuals. Another aspect of the social network that Lyotard refers to in the passage cited above is that of self-organisation: the “self­adjustments the system undertakes to improve its own performance”. I have already pointed to the fact that the discarding of a determining metanarrative on the one side and the autonomous agent on the other, suspends the self-organising process somewhere between active and passive. The social fabric is not “designed” through some transcendental principle, but develops as a result of the way in which it responds to contingent information in a dynamic fashion. This process is a complex one that involves many individuals with complex, nonlinear relationships between them, including feedback relations. Individuals cooperate to form clusters, but also compete for the resources in the network. The system is therefore not, and can never be, symmetrical.8 The history of the system is vitally important for the way in which meaning is generated in any part of it. The evolution and continuous alteration of structures in the social fabric is an integral part of its dynamics.

7 It is possible to interpret Lyotard, especially in later writings, as placing more weight on the “closed”aspect of local discourses. If he should then conclude that these discourses are hermetically separated, I will disagree. For me, at least, “local” in no way implies “closed”. 8 Symmetry would imply that the relationship of A to B is the same as the relationship of B to A. If for example, A is a child and B is a parent, one would say that their relationship is not symmetrical.



Postmodern knowledge and complexity (or why anything does not go)  

 129

Lyotard is quite clear on the point that the complexity of the social system does not automatically result in randomness or noise. From the passage above it is clear that the system “combats entropy”, that it generates meaning, not noise. To optimise this process, the system has to be as diverse as possible, not as structured as possible. This, for Lyotard, is the function of paralogy, and is equivalent to the process of self-organised criticality (Bak & Chen 1991).9 Self-organised criticality is the mechanism by which networks diversify their internal structure maximally. The more diverse the structure, the richer is the information that can be stored and manipulated. In order to be rich in its diversity, the network has to distribute its resources as widely as possible. This must be done in a flexible way while at the same time maintaining the integrity of the information stored in the relationships. The network has to walk a tightrope between a structure that is reliable but rigid on the one hand and a structure that is flexible but chaotic on the other. In our network model, this process is the result of the competition between units or groups of units. For Lyotard (1984: 61), the driving force in a social system is that of paralogy and dissension:10 [...] it is now dissension that must be emphasized. Consensus is a horizon that is never reached. Research that takes place under the aegis of a paradigm tends to stabilize; it is like the exploitation of a technological, economic, or artistic ‘idea’. It cannot be discounted. But what is striking is that someone always comes along to disturb the order of ‘reason’. It is necessary to posit the existence of a power that destabilizes the capacity for explanation, manifested in the promulgation of new norms for understanding or, if one prefers, in a proposal to establish new rules circumscribing a new field of research for the language of science.

Lyotard’s insistence on dissension and destabilising forces has serious implications for philosophy in general and specifically for philosophy of science. The role of science has often been understood in a positivistic sense, as one that has to fix knowledge in a permanent grid. Experimental evidence is used to verify theories. Sufficient verification would ensure a permanent place in the grid. It soon became clear, however, that the conditions for objective verification are problematic, that experimental evidence can support a theory, but cannot prove it. The experimental process cannot include all the factors that could possibly be involved, nor can it predict how new knowledge would change the interpretation of experimental results. Since one can disprove theories, the process of verifica-

9 The principles of self-organisation and self-organised criticality are discussed in detail in Cilliers (1994, chapter 5). 10 The importance of dissension, as opposed to consensus, form the core of Lyotard’s critique of Habermas, a subject that falls outside the scope of this study.

130 

 Paul Cilliers

tion was replaced by one of falsification. If one cannot add to the grid, you could at least disqualify unwanted members. This strategy of “throwing away” has the result that the body of knowledge that can be called scientific becomes leaner and leaner. Everything that is too complex or that contains uncertainties or unpredictability is, for the time being at least, out. Subsequently, large parts of the totality of human knowledge are disregarded as unscientific – most of the arts, most of psychology (for many scientists Freud remains the paradigm example of a scientific charlatan), most of the human sciences in general. With a strict understanding of what science is, even the life sciences (biology) and the empirical sciences (engineering) become suspect. Pushed to its limits, the theory of falsification implies that only abstract, a priori truths are really scientific. If one denies the existence of this metaphysical category, the body of knowledge that can truly be called “Scientific” becomes the empty set. Lyotard’s suggestion is that we discard the idea of consensus since it is impoverishing. To proliferate knowledge, we have to proliferate discourses without trying to fix them into a permanent grid. This position has some affinity with the position of Paul Feyerabend (1975). Feyerabend insists on a scientific “anarchy” in which all the marginalised voices should participate. There should be no immutable “method” that determines what forms part of the canon and what does not. Instead of throwing away everything that does not fit into the scheme, one should try to find meaningful relationships between the different discourses. The connectionist model provides us with an extremely important insight into these ideas. If it is granted that all knowledge is embedded in the larger social network (a point that becomes a watershed), then this proliferation of meaning and discourses is an inevitable characteristic of a complex, self-organising network. Lyotard and Feyerabend are not wilfully disruptive, anti­scientific anarchists, they are carefully considering the conditions of knowledge in a complex society. To allow previously marginalised voices equal opportunity, once again does not imply that “anything goes”. Dissenting voices receive no special privilege, they have to enter into the “agonistics of the network” where their relevance is dynamically determined through competition and cooperation in terms of the history as well as the changing needs and goals of the system. To conclude this section, a cautionary note: Since all these networks we have talked about are contingent entities, they are finite. Even the most complex ones have a finite capacity for handling information. A network can therefore suffer from an overload, especially when confronted with too much novelty. An overloaded network will show “pathological” behaviour, either in terms of chaotic behaviour or in terms of catatonic shutdown. This may actually be the state of affairs that many critics of postmodernism fear, that we are being overloaded with information and, in the process, annihilated (e.g. Baudrillard 1988). The



Postmodern knowledge and complexity (or why anything does not go)  

 131

point is, however, that there is little escape. Reverting to rigid, central control or the reintroduction of grand narratives will not make the information go away. We have to learn to cope with it by being more discriminating, by filtering out some of the excesses. Once again, the connectionist model is the most effective one for performing this “filtering”. In a rule-based system, preferences have to be programmed in and can be adjusted only with difficulty. Such systems remain paradigmatic of the modernist approach working with abstract forms of meaning (representation) and central control. Connectionist models can dynamically adjust themselves in order to select that which is to be inhibited and that which is to be enhanced. Robustness and flexibility are two sides of the same coin. In terms of our social condition, this would mean that we would experience less postmodern stress if we became less rigid in our interaction with each other and our environment. This does not mean that one should give up, or go with the flow. It means that we all have to enter into the agonistics of the network.

4 Postmodern society and complex systems In this section I wish to analyse postmodern society explicitly in terms of the ten characteristics of complex systems described in Section 2. Some of them have interesting implications for social theory which will be highlighted. Bear in mind that the connectionist model remains a model of our complex society and that it should not be hypostatised. The usefulness of the model will only become clear to the extent we allow it to play out its possibilities. Perhaps one should also add that although our present (postmodern) society is a particularly complex one, this analysis holds for the concept of society in general. Human affairs  – and nature – have always been complex.

i. Complex systems consist of a large number of elements. If we analyse the social system at such a level of detail that we consider human individuals to be the elements, the number of elements in the social system is certainly huge.

132 

 Paul Cilliers

ii. The elements in a complex system interact dynamically. Individuals are engaged in a constant exchange of information. Remember that a specific node in a neural network has limited significance, that it is the patterns of interconnections that encode information and generate meaning. Similarly, no human individual is meaningful in isolation, “the self does not amount to much” (Lyotard 1984: 15). The individual is constituted by its relationships with others.

iii. The level of interaction is fairly rich. Human individuals interact with many others in a vast array of different capacities. In the postmodern society the level of interaction is growing continuously.

iv. These interactions are nonlinear. Nonlinearity is a precondition for complexity, especially where self-organisation, dynamic adaptation and evolution are at stake. Closely related to the principle of nonlinearity is the principle of asymmetry. Linear, symmetrical relationships result in simple systems with transparent structures. In complex systems mechanisms have to be found to break symmetry and to exploit the magnifying power of nonlinearity. This is ensured by the rich level of interaction and by the competition for resources. The social system is nonlinear and asymmetrical as well. The same piece of information has different effects on different individuals and small causes can have large effects. The competitive nature of social systems is often regulated by relations of power, ensuring an asymmetrical system of relationships. This, it must be emphasised strongly, is not an argument in favour of relations of domination or exploitation. The argument is merely one for the acknowledgement of complexity. Nonlinearity, asymmetry, power and competition are inevitable components of complex systems. It is what keeps them going, their engine. If there was a symmetrical relationship between infants and adults, the infants would never survive. If there was a symmetrical relationship between teacher and student, the student will never be able to learn something new. If the state had no power, it would have no reason to exist. If women and men were the same, our world would be infinitely less interesting. These considerations have important implications for social theory. The fact that society is held together by asymmetrical relations of power does not mean that these relationships are never exploited. To the contrary, they are continu-



Postmodern knowledge and complexity (or why anything does not go)  

 133

ously exploited by parents, by lecturers, by the state and by men, but also by children, by students, by citizens and by women. The point is that the solution to these forms of exploitation does not lie in some symmetrical space where power is distributed evenly. Such spaces cannot exist in complex systems that are driven by nonlinearity. The hope that such spaces could be created in any enduring fashion is false. To combat exploitation, as a matter of fact, to fulfil any ethical obligation, you cannot rely on some form of universal, a priori discourse. There is only one option: you have to enter into the agonistics of the network. You have to make your hands dirty.

v. The interactions have a fairly short range. The elements in a complex network usually interact primarily with those around them. In large networks this results in groups or assemblies of elements clustering together to perform more specific functions. Lyotard (1984: xxiv, 61, 66) describes this phenomenon as “local determination”. Elements operate on information that is available to them locally. They have no other option, since that is the only place where information is available. In the complex (postmodern) system there is no meta-level of control, no central distribution centre that determines the patterns of distribution of censored information. The behaviour of the system is therefore characterised best in terms of a multiplicity of local “discourses”. At the risk of repetition, it must be emphasised that these locally determined groups are not isolated from each other. Despite the short range of the interactions, nothing precludes wide-ranging influence. The different clusters are interconnected and since these connections are nonlinear, they can have large effects, even if the interconnections are sparse. Important events can reverberate through the system quite rapidly, but they are never propagated in a pure form since they are constantly modulated by the clusters they pass through.

vi. There are loops in the interconnections. Feedback is an essential aspect of complex systems. Not feedback as understood simply in terms of control theory, but as intricately interlinked loops in a large network. This means that the activity of an element can directly or indirectly influence itself. In postmodern theory this manifests itself as the problem of reflexivity (see Lawson 1985). If one accepts that information is proliferated throughout the system and that it is continually transformed, by other bits of information and by itself, then it becomes impossible to stipulate a “true” interpretation for any

134 

 Paul Cilliers

piece of information. Information can only be interpreted locally and then only through the dynamics of différance – as reflecting upon and transforming itself. These dynamics preclude the definition of truth or origin on a meta-level and is therefore referred to as the postmodern predicament: “a crisis of our truths, our values, our most cherished beliefs” (9). There may indeed be a crisis of knowledge, but, and this must be underscored, the crisis is not the result of the disruptive activity of wilful theoreticians like Nietzsche, Heidegger and Derrida. It is a direct result of the complexity of our postmodern society. This is the point that Lyotard is also trying to make when he insists that the conditions for knowledge are locally determined. Reflexivity does lead to paradox, but this is a problem only if all paradox has to be resolved on a meta-level. If one remains within the agonistics of the network you have to cope with the paralogy of the postmodern condition. This does not imply that it is impossible to interpret information, it merely means that all interpretations are contingent and provisional, pertaining to a certain context and a certain history within a certain timeframe.

vii. Complex systems are open systems. We have already made the point that local discourses are not closed off, but interact with each other. The social system also interacts with nature. Under the postmodern condition this relationship has also come under new scrutiny, giving rise, for example, to strong political groupings with the environment as their prime concern. Nature is no longer the passive object of Man’s exploitation, but is being given new meaning in the context of a process of interaction.

viii. Complex systems operate under conditions far from equilibrium. Complex systems need a constant flow of energy to change, to evolve and to survive as complex entities. Equilibrium and symmetry (stasis) means death. Just as the flow of energy is necessary to fight entropy and maintain the complex structure of the system, society can only survive as a process. It is not defined by its origins or its goals, but by an ever present flux. In postmodern society this flux, this lack of equilibrium has been fully exposed, particularly through the role of the mass media. This has an unsettling effect on many and one has to develop certain skills to cope with these conditions. For example, one has to filter out



Postmodern knowledge and complexity (or why anything does not go)  

 135

some of the excess information. However, to yearn for a state of complete equilibrium is to yearn for a sarcophagus.11

ix. Complex systems have a history. The importance of history has been emphasised. One point bears repetition. The history of a complex system is not objectively available, it is a collection of traces distributed over the system that can never be given a single interpretation. In this sense of the word, postmodern society, is certainly not a­‑historical. It does deny an interpretation where history becomes a master key for unlocking the true meaning of present conditions. It cannot, however, deny being caught up in and being influenced by the unstoppable flow of time.

x. Individual elements are ignorant of the behaviour of the complete system in which they are embedded. This is a more complex point that needs careful consideration. We have already pointed to the fact that elements in a system can only respond to local information, given that this information can be quite rich. We have also shown that single elements of a complex system are not significant by themselves, but through their patterns of interaction. The point made here is slightly different. Single elements cannot contain the complexity of the whole system and can therefore neither control nor comprehend it fully. Because of the overwhelming amount of information available in postmodern society we often live under the illusion that we get the complete picture. Because of the complexity of our society this is not possible. Since we are in part creating the society through our actions, there is no complete picture possible. No other element is better off either. Similarly, single elements cannot exert complete control over a decentralised system. For example, I may want to combat inflation, but I have no way of measuring the effect of my own spending pattern. These effects only become apparent when my behaviour combines with those of a large number of other economic agents.

11 This yearning is not an unfamiliar one in the history of philosophy. It can be found in Aristotle’s wish for the quiet life of contemplation, in the abstract peace yearned for in many religions, including Buddhism and many forms of Christianity, as well as in Freud’s death principle (Thanatos as opposed to Eros).

136 

 Paul Cilliers

The claim that the structure of society is an emergent property of the social system may create a feeling of irrelevance in some people. This need not be the case. In the first place the relevance of our activities is determined by the effectiveness with which we enter into the agonistics of the network, not by understanding it through a view from God’s eye. Secondly, it must be kept in mind that small causes can have large effects – the interactions are nonlinear. It also means, however, that the effects of our actions are somewhat unpredictable. To conclude this section I wish to tum briefly to Lyotard’s claims that his analysis of the postmodern condition provides us with “the outline of a politics that would respect both the desire for justice and the desire for the unknown” (67). Here Lyotard rejects the claim that the absence of all meta-descriptions (or prescriptions) makes the postmodern condition fundamentally unethical. This claim rests on the assumption that any form of critique that is not backed up by objective criteria, or at least by some form of consensus, can be dismissed readily. Lyotard singles out Habermas as one who would adopt this kind of approach but finds it “neither possible, nor even prudent” (45). Lyotard claims that the Habermasian approach, using a dialogue of argumentation, rests on two assumptions. In the first place, it assumes that “it is possible for all speakers to come to agreement on which rule or metaprescriptions are universally valid for all language games” and in the second place it assumes “that the goal of dialogue is consensus” (65). Lyotard finds neither of these assumptions acceptable, primarily because they deny the complexity of postmodern society, the nature of which he describes in the following way: It is a monster formed by the interweaving of various networks of heteromorphous classes of utterances (denotative, prescriptive, performative, technical, evaluative, etc.). There is no reason to think that it could be possible to determine metaprescriptives common to all of these language games or that a revisable consensus like the one in force at a given moment in the scientific community could embrace the totality of metaprescriptions regulating the totality of statements circulating in the social collectivity. As a matter of fact, the contemporary decline of narratives of legitimation – be they traditional or ‘modern’ (the emancipation of humanity, the realization of the Idea) – is tied to the abandonment of this belief (65).

The first assumption of the Habermasian approach is directly opposed to Lyotard’s emphasis on the proliferation of heterogeneous discourses  – the role of paralogy  – while the second is opposed to his insistence on the importance of dissent. Not that consensus is always impossible, it can be achieved, but only as a local phenomenon limited in both time and space. Consensus as a goal would attempt to freeze the social system in a particular state. Since this will not happen, one has to be sensitive to the process of social transformation. This



Postmodern knowledge and complexity (or why anything does not go)  

 137

may indicate that “consensus has become an outmoded and suspect value” but, claims Lyotard, “justice as a value is neither outmoded nor suspect” (66). Lyotard does not really develop this concept of justice, perhaps because of his acknowledgement of its complexity, but he suggests two important, if predictable moves: to recognise the heteromorphous nature of language games and to recognise that all agreements on the rules of any discourse as well as the “moves” allowed within that discourse must be local, in other words, agreed on by its present players and subject to eventual cancellation (66). How this proposal sketches an idea and practice of justice can perhaps be best understood in the following way: it becomes the responsibility of every player in any discursive practice to know the (local) rules of the language game involved and to assume responsibility for both the rules themselves and the effects of that specific practice. This responsibility cannot be shifted to any universally guiding principles or institutions – whether they be the State, the Church or the Club. A postmodern principle of justice is summed up once again in what has become a bit of a Leitmotiv: you cannot escape the agonistics of the network.

5 Conclusion I have given a characterisation of complex systems and have argued that human society is such a system. Although this means that a coherent meta-description of our society is not possible, it does not mean that society is either random or chaotic. It is therefore incorrect to conclude that the absence of a metanarrative implies that anything goes. Our society is richly patterned, containing many complex structures that cannot be understood in an abstract sense, but only by becoming part of the process by which they are constituted and transformed. In a subsequent article I will explore the consequences of complexity theory for our understanding of language and for a number of contemporary arguments concerned with the status of scientific knowledge. Here we will also reach the conclusion that we cannot escape the acceptance of a responsibility within the agonistics of the network. This is fundamentally an ethical position and needs deeper analysis than Lyotard’s elliptical remarks. The way in which post-structural thinkers are engaging with the radical ethics of Emmanuel Levinas may serve as good starting point for the project of giving substance to the claim that the absence of a guiding metanarrative leaves us primarily in a position of responsibility.

138 

 Paul Cilliers

References Bak, P. & Chen, K. 1991. Self-organized criticality. In: Scientific American January: 26–33. Baudrillard, J. 1988. The ecstasy of communication. New York: Semiotext(e). Cilliers, F.P. 1990. The brain, the mental apparatus and the text: a post-structural neuropsychology. In: South African Journal of Philosophy 9(1): 1–8. Cilliers, F.P. 1991. Rules and relations. Some connectionist implications for cognitive science and language. In: South African Journal of Philosophy 10(2): 49–55. Cilliers, F.P. 1994. Modelling complexity. D Phil thesis (unpublished): University of Stellenbosch. Feyerabend, P. 1975. Against method. Outline of an anarchistic theory of knowledge. London: Verso. Lawson, H. 1985. Reflexivity. The post-modern predicament. London: Hutchinson. Lyotard, J.F. 1984. The postmodern condition: a report on knowledge. Manchester: Manchester University Press. Parushnikova, A. 1992. Is a postmodern philosophy of science possible? In: Studies in the History and Philosophy of Science 23(1): 21–37.

Paul Cilliers

Complexity, deconstruction and relativism 1 Introduction It is easily acknowledged that different intellectual traditions have different understandings of what the nature and status of meaningful knowledge is. This would not have been a problem if these discourses operated in isolation. However, different epistemological positions interact and compete with one another. This competition is necessary, of course, but it is rarely an amicable one, probably because our basic understanding of the world, and of our role in it, is at stake. Thus there is no agreement even on the criteria for what would count as meaningful knowledge. The need for clarity and certainty has often, and increasingly, been bolstered by an appeal to science, or at least to a certain understanding of what it means to be scientific. This has indeed led to a deeper understanding of the world, but it has also resulted in reductionist strategies of thinking that underestimate the complexity of much of what we try to understand. Fortunately we no longer have to fight against a crude positivism, but at the same time there seems to be a growing resistance against theoretical positions which emphasise the interpretative nature of knowledge. More specifically, there seems to be a need to dismiss positions that can be called postmodern, post-structural or deconstructive. This need is best exemplified by (but certainly not restricted to) the so-called Sokal’s hoax and the subsequent dismissal of a number of important postmodern thinkers (Sokal & Bricmont 1998).1 There is at least one important lesson to be learnt from this affair: one has to be very careful when using and criticising work from a foreign discipline (irrespective of whether one is a social or natural scientist).2 If this remained the central contribution of this storm in a thimble, much good could come of it. However, Sokal’s hoax is still being used, especially by those promoting science in a new-positivistic way, to dismiss important contributions from thinkers perceived not to be adequately informed about what is

1 Other examples include Ellis (1989) and O’Neill (1995). For a more balanced engagement, see Wheeler (2000) and Luntley (1995). 2 For some responses to Sokal (none of them from a “French” perspective), see Beller (1998), Guillory (2002) and Haworth (1999). Originally published in Theory, Culture and Society, 2005, 22(5): 255–267. DOI: 10.1177/0263276405058052, © 2005 SAGE Publications.

140 

 Paul Cilliers

really happening in science. When important writers, such as Richard Dawkins (2002: 47–53), use Sokal’s writings to continue the disparagement of those critical of the role that science plays in establishing certain cultural and political conditions, the matter develops an ethical dimension which should not be concealed under what is offered to the general public as ‘scientific facts’. The very appeal to scientific objectivism becomes a political move in establishing a certain mode of understanding as the privileged one. It may well be that a certain kind of new-positivism is necessary to serve as a correction for some of the excesses of postmodernism. Many postmodern positions are so open and vague that they really do not contribute to our knowledge of the world. If that was all there was to it, we could merely let the debate be. However, if we acknowledge that the world in which we have to live is complex; we also have to acknowledge the limitations of our understanding of this world.3 What is more, some of the theoretical positions that are being dismissed so assertively, such as deconstruction, help us to cope with these limitations and should not be relegated to the junkyard of history. They should be developed in conjunction with our growing scientific knowledge. The opposition sketched above can be generalised into an opposition between what could be called self-confident or assertive positions and modest positions. The term “modest” will be used to describe reflective positions that are careful about the reach of the claims being made and of the constraints that make these claims possible. The aim of this article is to argue for the importance of modest positions when trying to deal with complex problems. Deconstruction serves as an example and I will argue that the view from complexity serves as another, or rather, as a complementary one.4 The dismissal of everything postmodern will therefore include the dismissal of a number of important insights from our understanding of complexity.

3 This is argued for in the next section. 4 For other discussion of the relationships between complexity and post-structuralism, see Taylor (2001) and Dillon (2000). Some of the critics of deconstruction, as well as some of its more radical (but uninformed) supporters, see it as a new form of nihilism, as something which contemplates emptiness, or the “void”. This is, to my mind, an incorrect interpretation of Derrida’s position. He is at pains to show that there is always a plenitude of meaning, not a lack of it. The play of différance creates meaning, it does not destroy it. It is exactly in this respect that there is a close link between deconstruction and complexity. Both emphasise that we deal with a world of growing plenitude, that our understanding of that world involves a reduction of the plenitude, and that there is no meta-method for doing such a reduction. For some of Derrida’s clearest articulations on this issue, see Derrida (1988), especially the Afterword.



Complexity, deconstruction and relativism 

 141

Attempts to reject modest positions are based on a number of important arguments which have to be taken seriously. Before this is done, it has to made clear that what is at stake here is not an apology for modesty, but an argument for the importance of modesty. The failure to acknowledge the complexity of a certain situation is not merely a technical error, it is also an ethical one. A modest position should not be a weak position, but a responsible one. Such a position will be developed by examining three arguments: the argument that modest positions lead to relativism, the argument that modest positions are subject to the performative contradiction and the argument that modest positions are vague. Before tackling each of these, the view from complexity, at least to my understanding, should be presented briefly.

2 The view from complexity there are different understandings of complexity theory and its implications. On the one hand, there is a more strictly mathematical and computational view. This view is often developed via insights from chaos theory. In the cases where such a “hard” understanding is uncritically appropriated by the human sciences, it can lead to exactly the kind of positivism that is being argued against in this article. On the other hand, there is a more critical understanding of complexity. This view argues that complexity theory does not provide us with exact tools to solve our complex problems, but shows us (in a rigorous way) exactly why these problems are so difficult.5 This second view may have a more sceptical perspective on what can be done with complexity theory, but it is developed from an understanding that is not really at odds with a generally accepted scientific characterisation of complexity. These characteristics can be summarised in the following way:6 1. Complex systems are open systems. 2. They operate under conditions not at equilibrium. 3. Complex systems consist of many components. The components themselves are often simple (or can be treated as such). 4. The output of components is a function of their inputs. At least some of these functions must be nonlinear.

5 See Richardson and Cilliers (2001) for a discussion of some of these issues. 6 These characteristics were formulated in collaboration with Fred Boogerd and Frank Bruggemans at the Department of Molecular Cell Physiology at the Free University, Amsterdam. Similar lists by Holland (1998: 225–31), Emmeche (1997) Kauffman (1971) and Cilliers (1998) were consulted in the process.

142 

 Paul Cilliers

5. The state of the system is determined by the values of the inputs and outputs. 6. Interactions are defined by actual input-output relations and they are dynamic (the strength of the interactions change over time). 7. Components on average interact with many others. There are often multiple routes possible between components, mediated in different ways. 8. Some sequences of interaction will provide feedback routes, whether long or short. 9. Complex systems display behaviour that results from the interaction between components and not from characteristics inherent to the components themselves. This is sometimes called emergence. 10. Asymmetrical structure (temporal, spatial and functional organisation) is developed, maintained and adapted in complex systems through internal dynamic processes. Structure is maintained even though the components themselves are exchanged or renewed. 11. Complex systems display behaviour over a divergent range of timescales. This is necessary in order for the system to cope with its environment. It must adapt to changes in the environment quickly, but it can only sustain itself if at least part of the system changes at a slower rate than changes in the environment. This part can be seen as the “memory” of the system. 12. More than one description of a complex system is possible. Different descriptions will decompose the system in different ways. Different descriptions may also have different degrees of complexity. If one considers the implications of these characteristics carefully a number of insights and problems arise: 1. The structure of a complex system enables it to behave in complex ways. If there is too little structure, i.e. many degrees of freedom, the system can behave more randomly, but not more functionally. The mere “capacity” of the system (i.e. the total amount of degrees of freedom available if the system was not structured in any way) does not serve as a meaningful indicator of the complexity of the system. Complex behaviour is possible when the behaviour of the system is constrained. On the other hand, a fully constrained system has no capacity for complex behaviour either. (This claim is not quite the same as saying that complexity exists somewhere on the edge between order and chaos. A wide range of structured systems display complex behaviour.) 2. Since different descriptions of a complex system decompose the system in different ways, the knowledge gained by any description is always relative to the perspective from which the description was made. This does not imply that any description is as good as any other. It is merely the result of the fact that only a limited number of characteristics of the system can be taken into



Complexity, deconstruction and relativism 

 143

account by any specific description. Although there is no a priori procedure for deciding which description is correct, some descriptions will deliver more interesting results than others.77 3. In describing the macro behaviour (or emergent behaviour) of the system, not all the micro features can be taken into account. The description is a reduction of complexity. Nevertheless, macro behaviour is not the result of anything else but the micro activities of the system. Yet, to describe the macro behaviour purely in terms of the micro features is a difficult task. When we do science, we usually work with descriptions which operate mainly on a macro level, but these descriptions will, more often than not, be approximations of some kind. These insights have important implications for the knowledge claims we make when dealing with complex systems. To fully understand a complex system, we need to understand it in all its complexity. Furthermore, because complex systems are open systems, we need to understand the system’s complete environment before we can understand the system, and, of course, the environment is complex in itself. There is no human way of doing this. The knowledge we have of complex systems is based on the models we make of these systems, but in order to function as models – and not merely as a repetition of the system – they have to reduce the complexity of the system. This means that some aspects of the system are always left out of consideration. The problem is compounded by the fact that that which is left out, interacts with the rest of the system in a nonlinear way and we can therefore not predict what the effects of our reduction of the complexity will be, especially not as the system and its environment develops and transforms in time.8 We cannot have complete knowledge of complex systems; we can only have knowledge in terms of a certain framework. There is no stepping outside of complexity (we are finite beings), thus there is no framework for frameworks. We choose our frameworks. This choice need not be arbitrary in any way, but it does mean that the status of the framework (and the framework itself) will have to be continually revised. Our knowledge of complex systems is always provisional.9 We have to be modest about the claims we make about such knowledge.

7 This issue will be returned to when we deal with relativism. 8 These ideas are elaborated upon in Cilliers (2000) and (2001). 9 For a similar view, see Najmanovich (2002).

144 

 Paul Cilliers

The links with postmodern positions, specifically with deconstruction, should now be obvious.10 Deconstruction argues for the irreducibility of meaning. Meaning and knowledge cannot be fixed in a representational way, but is always contingent and contextual. The context itself is not transparent, but has to be interpreted. Derrida (1988: 118–119) explicitly links the problem of meaning and context to the fact that these things are complex. The critical understanding of complexity theory presented here, and deconstruction, therefore, make a very similar claim: knowledge is provisional. We cannot make purely objective and final claims about our complex world. We have to make choices and thus we cannot escape the normative or ethical domain. This is, of course, a contested position. The same arguments used to dismiss deconstruction can now also be used to dismiss the view from complexity. Nevertheless, the question remains whether these arguments are effective, whether they actually show that this position is a weak one that should no longer be taken seriously. That is why they have to be examined more carefully.

3 Against relativism Perhaps it is not necessary to spend too much time in defending deconstruction and the view from complexity from the accusation that they lead to relativism. This accusation usually comes as a kind of knee-jerk reaction in a bid to dismiss or demolish deconstruction and usually after it has superficially – and erroneously  – been associated with postmodernism. A good example of this position can be found in Sweetman (1999). After claiming that the work of Derrida is “an ideal representative of postmodern philosophy in general” (1999: 5–6), he proceeds to criticise it on the following five points: 1. it confuses aesthetics with metaphysics 2. it mistakes assertion for argument in philosophy 3. it is guilty of relativism (both epistemological and moral) 4. it is self-contradictory

10 Dillon (2000: 4) describes the relationship between post-structural positions and complexity theory as a commitment to the “anteriority of radical relationality”. He proceeds to argue for certain differences between the two. His categorisation of “complexity” is, however, a little general. I argue for an understanding of complexity which is not primarily concerned with intelligence, survival or fitness (2000: 22), but with the limits of our knowledge, and thus with the inevitability of normative components. This interpretation of complexity is to my mind compatible with post-structuralism, at least in its Derridean form.



Complexity, deconstruction and relativism 

 145

5. it is guilty of intellectual arrogance because its proponents insist that its critique of traditional philosophy can still succeed even though its positive claims have not been established (1999: 6, 14). These kind of criticisms have been addressed in some detail by Derrida, for example in the Afterword to Limited Inc. (Derrida 1988) and by others (for example Norris 1997), and will, therefore, not be repeated here. The argument is mostly pursued by those still working with a strict (and hierarchical) distinction between analytical and continental philosophy,11 or between natural science and the humanities. This kind of distinction is also active in the complexity community between those arguing for a strict scientific or mathematical foundation for complexity theory, and those seeing complexity as something more metaphorical.12 Generally speaking, these dichotomies serve mostly as stumbling blocks, or as ways to dismiss intellectual opponents and not as a framework for fruitful discussion. Virtually nobody claims to be a relativist; it is a self-refuting position. Why then is the accusation that a certain position implies relativism used so often? It has to be because of a deeply held fear that perhaps “true” knowledge will continue to elude us. We have to keep on convincing ourselves that relativism is bad. But there is more to it. A true relativist, i.e. somebody that argues that there are no grounds for any form of knowledge is, in a way, nothing but a disappointed foundationalist. If he cannot find objective and universal points of reference to guarantee knowledge, then he may as well give up. The argument between foundationalists and relativists is a dead end – a family fight. What then is the status of the claim that we cannot know complex things completely if it does not imply relativism? In the first place, one should realise that the claim that we cannot have complete knowledge does not imply that anything goes. “Limited” knowledge is not equivalent to “any” knowledge. If this were so, any modest claim, i.e. any claim with some provisionality or qualification attached to it, would be relativistic. The only alternative then would be an arrogant self-assurance. Such a self-assured position is deeply problematic since its complacency forecloses further investigation. Modest claims are not relativistic and, therefore, weak. They become an invitation to continue the process of generating understanding.

11 See Nuyen (1989) for a discussion of why a systems approach is difficult to maintain from the analytical perspective. 12 See note 5.

146 

 Paul Cilliers

4 Against the performative contradiction A serious philosophical argument often brought against deconstruction, for example by Habermas (1987: 185–210), is that it is subject to the performative contradiction. Simply put, this mistake is made when there is a contradiction between what you say, and the way in which you say it. Thus Habermas claims that when Derrida argues against reason, he has to make use of rational means. “Anyone who argues against reason is necessarily caught up in a contradiction: she asserts at the locutionary level that reason does not exist, while demonstrating by way of her performance in argumentative processes that such reason does in fact exist” (Fleming, 1996: 169). The claim made above  – that we can never have complete knowledge of complex systems – falls into the same trap. It looks like an absolute statement about complex things but denies that such a statement can be made. Whether Habermas is correct in his assessment that Derrida argues against reason13 is of less importance now than it is to look at the “logic” of the performative contradiction. The first thing one should notice is that most careful or modest claims will come under pressure from this test. The claim “no sentence has an exact meaning” obviously fails the test, but the claim “perhaps some sentences are not perfectly clear” is also in trouble. If it is correct, then the sentence itself is perfectly clear. If it is not correct, then perhaps all sentences are clear. This point can be made more explicit by examining what kind of statements would pass the test. The claim, “When I am rational I will always be right” passes the test with flying colours! It may not be true, but there is no contradiction between what I say and how I am saying it. I am always right, and I am also right that I am always right, and I can make this claim in an assertive tone of voice. Surely a test that will pass most self-assertive, macho claims and that will fail most modest claims, cannot be all that useful when dealing with complex things. Some reasons for this can be supplied. The performative contradiction is predicated on the assumption that one can adequately distinguish between the performative and the locutionary levels, and, in the terms Habermas uses to criticise Derrida, between logic and rhetoric. However, in order to make this distinction clearly, one would need to take in a position that can characterise what is being said from an external vantage point. In the language of complexity, that would mean that one has access to a framework that is not the result of a strategic choice, i.e. some objective meta-framework. This is exactly what the view

13 See Fleming (1996) for a further discussion of this issue. She argues that deconstruction works from within the tradition of rational argument.



Complexity, deconstruction and relativism 

 147

from complexity is sceptical about. The argument is that our frameworks are all compromised to some extent; dealing with complexity is a little messy. As Derrida (1988: 119) says: if things were simple, word would have gotten around.14 In a way, the view from complexity acknowledges that some form of performative tension is inevitable. We are playing in what Wood (1990: 150) calls the “theatre of difficulty”, and this requires a certain “performative reflexivity” (1990: 132). We need to demonstrate the difficulties we are in; also in the way we talk about them. Our discourse should reflect the complexities. To talk about the complex world as if it can be understood clearly is a contradiction of another kind15 and this is a contradiction with ethical implications. Those who claim to have access to the truth are denying us our critical perspective and, therefore, keep us in a kind of false consciousness by not restoring the world to its original difficulty. It is only by acknowledging that we are in trouble that we can start grappling with the complexities around us. To be subject to the performative contradiction would seem, at least from the perspective of a certain kind of logical argumentation, to be a weak position. Such a position is seen as not being sufficiently rational and thus unscientific and irresponsible. The view from complexity argues to the contrary, that the conditions imposed by the test for performative contradiction feeds off a kind of intellectual arrogance that is in itself irresponsible. We only have limited access to a complex world and when we are dealing with the limits of our understanding, we are dealing with ethics. In Derrida’s (2000: 467) words: “There is ethics precisely where I am in performative powerlessness”. The modest position is not weak; it is responsible.

14 In the Afterword to Limited Inc., Derrida (1988) defends his position against several accusations, including that it is a relativist position, and that it is obscure: These things are difficult, I admit; their formulation can be disconcerting. But would there be so many problems and misunderstandings without this complexity and without these paradoxes? One shouldn’t complicate things for the pleasure of complicating, but one should also never simplify or pretend to be sure of such simplicity where there is none. If things were simple, word would have gotten round, as you say in English. There you have one of my mottos, one quite appropriate for what I take to be spirit of the type of ‘enlightenment’ granted our time. Those who wish to simplify at all costs and who raise a hue and cry about obscurity because they do not recognize the unclarirty of their good old Aufklärung are in my eyes dangerous dogmatists and tedious obscurantists. No less dangerous (for instance, in politics) are those who wish to purify at all costs (1988: 119). 15 Derrida makes the same point in his defence against Habermas’ claim that deconstruction is subject to the performative contradiction. See Derrida (1988: 134, n. 9). See also Derrida (2000).

148 

 Paul Cilliers

5 Against vagueness A third objection to the view from complexity is that it results in a position which is vague.16 The argument could be made that because this position is loathe to make strong claims for the truth of its statements, it can only produce vague generalisations or platitudes which offer little resistance to being interpreted loosely. This objection is perhaps related to, but not quite the same as, the one accusing the view from complexity of relativism. In trying to avoid relativism, the argument may go, specific claims can be made; but in a way that is so watered down, or obscure, that one cannot come to grips with them. The objection is most certainly valid in many cases. In some (postmodern) circles a vague kind of chatter, employing a shared vocabulary in an uncritical way, has become acceptable – one could even say a new orthodoxy. Sokal’s hoax certainly contributed to the exposure of this. There is no excuse for academic groupies or sloppy reviewing practices (the prime reason why Sokal’s fake article was allowed to create the stir it did). The problem is exacerbated by the fact that much of the terminology used by the groupies is borrowed from decent academic disciplines. Examples of such jargon include difference, deconstruction, democratic, power, gender, rhizomatic, signifier, dialectic, quantum, chaos and complexity. It becomes difficult to establish when these terms are used with insight, and when they are only mentioned in order to make acceptable noises. (Many of these terms have, of course, been used in this article.) There is no defence for this vague groupspeak. However, it must be emphasised that there is no reason in principle why a modest position should be a vague one. For a statement to be intelligible at all, it must be possible to distinguish it from other claims. Intelligibility does not result from some external guarantee, some truth-giving process, but it is the result of a process of differentiation; a process that has nothing to do with fuzziness. Not grasping this point has led to a number of misguided dismissals of deconstruction. The deconstructive claim that meaning is not saturated, or that language has an element of “play”, does not imply that there is no meaning, or that any meaning of a term is as good as

16 The problem of vagueness has received a lot of attention in Logic where the issue at stake is the relationship between the sometimes vague sentences in natural language and the precise statements of logic and mathematics. How does logic deal with borderline cases, and how does it solve the Sorites paradox (one grain of sand is not a heap, two grains of sand is not a heap, ...)? One suggestion is to modify classical logic into fuzzy logic. Although it is related to what will be discussed, the problem of vagueness in logic will not be investigated here. See Greenough (2003), as well as the other articles in Mind Vol. 112, for more detail.



Complexity, deconstruction and relativism 

 149

any other. The deconstructive claims have to do with the limits of our claims, not with their intelligibility. In reply to questions arising from Searle’s critique of deconstruction, Derrida (1988: 114–31) discusses this issue in some detail. Every concept that lays claim to any rigor whatsoever implies the alternative of ‘all or nothing’. Even if in ‘reality’ or in ‘experience’ everyone believes he knows that there is never ‘all or nothing’, a concept determines itself only according to ‘all or nothing’. Even the concept of ‘difference of degree’, the concept of relativity is, qua concept, determined according to the logic of all or nothing, of yes or no: differences of degree or nondifference of degree. It is impossible or illegitimate to form a philosophical concept outside this logic of all or nothing (1988: 116–17).

Derrida’s point is that for communication to take place at all, concepts (or signs) have to be recognisable (iterable), and therefore they have to be differentiated from other concepts. This differentiation cannot be vague or done by statistical approximation since that would not delineate the concept in question. He continues his argument: [one] neither can nor should avoid saying: it’s serious or nonserious, ironical or nonironical, present or nonpresent, metaphorical or nonmetaphorical, [...] etc. To this oppositional logic, which is necessarily legitimately a logic of ‘all or nothing’ and without which the distinction and the limits of a concept would have no chance, I oppose nothing, least of all a logic of approximation [á peu prés], a simple empiricism of difference in degree; rather I add a supplementary complication that calls for other concepts, for other thoughts beyond the concept and another form of ‘general theory’, or rather another discourse, another ‘logic’ that accounts for the impossibility of concluding such a ‘general theory’ (1988: 117).

However, the fact that the concept has to be communicated clearly, not by approximation, does not imply that the concept now has an indisputable identity. In a different context a different set of differentiations may come into play which would give the (still clearly recognisable) concepts different meanings. For the concept to have meaning at all, it has to be limited, but these limits are not a priori or external to the situation. They are contingent and historical. The “art” of deconstruction, like the art of modelling complex systems, is in many ways nothing more than the examination of these limits. In a way similar to deconstruction, the view from complexity claims that we cannot know complex things completely (Cilliers 2002). This does not imply that we can know nothing about complex systems, or that the knowledge claims we make about them have to be vague, insipid or weak. We can make strong claims, but since these claims are limited, we have to be modest about them.

150 

 Paul Cilliers

6 Conclusion: against arrogance When dealing with complexity, modest positions are inescapable. This does not imply that they should be relative, vague or self-contradictory, nor does it imply a reason to cringe in false modesty. We can make clear, testable assertions about complex systems. We can increase the knowledge we have of a certain system, but this knowledge is limited and we have to acknowledge these limits. The fact that our knowledge is limited is not a disaster, it is a condition for knowledge. Limits enable knowledge. Without limits we would have to incorporate life, the universe and everything into every knowledge claim we make and that is not possible. Limiting frameworks makes it possible to have knowledge (in finite time and space). At the same time, having limits means something is excluded, and we cannot predict the effects of that exclusion. Knowledge is a fragile and, above all, contingent thing (see also Barrow 1999; Luntley 1995: 136– 149). The notion that limits and constraints are necessary conditions for knowledge has an important corollary in the complexity debate. It has been argued above that meaningful structure can only develop in a complex system if there are constraints in place. The fact that a system has many degrees of freedom is in itself no guarantee for complex behaviour. It is only when this freedom is constrained that structure can arise. Such structure is not a priori or externally given, but is developed in response to contingent conditions in the history of the system and has a certain resilience. Complex systems are not balanced on a knife’s edge between chaos and order. They have mostly robust structures, which change over time and enable the system to respond to different circumstances. It is, therefore, incorrect to associate complexity with noise as Taylor (2001) does. If complexity is aligned with notions of chaos, randomness and noise, the accusations of relativism and vagueness will start to hold water. If it is aligned with notions of structure as the result of contingent constraints, we can make claims about complex systems which are clear and comprehensible, despite the fact that the claims themselves are historically contingent. The view from complexity entails that we cannot have perfect knowledge of complex systems. We cannot “calculate” the performance of, for example, complex social systems in their complexity; we have to reduce that complexity; we have to make choices. Normative issues are, therefore, intertwined with our very understanding of complexity. Ethical considerations are not to be entertained as something supplementing our dealings with social systems. They are always already part of what we do. One could attempt to deny that and operate as if one can deal with complexity in an objective way – as if we can calculate everything – and thereby avoid the normative dimension. But this denial of the



Complexity, deconstruction and relativism 

 151

ethical becomes an avoidance of responsibility and is, of course, ethical in itself, albeit a negative (and much too prevalent) ethics. Furthermore, the claim that our understanding of complex systems cannot be reduced to calculation means that there will always be some form of creativity involved when dealing with complexity. “Creativity” should not (only) be understood in terms of flights of fancy or wild (postmodern) abandon, but also in terms of a careful and responsible development of the imagination. Imagining the future will involve risk, but the nature of this risk will be a function of the quality of our imagination. It is important that we start imagining better futures, and for that we need better imaginations. Reading books, listening to music, appreciating art and film is not a form of entertainment to be indulged in after we have done our serious work. These creative activities stimulate the imagination and thereby transform the frameworks we apply when apprehending the world. If we do not foster the creative arts, we will end up in the well-managed dystopia of the brave new world. The view from complexity argues for the necessity of modest positions. In order to open up the possibility of a better future we need to resist the arrogance of certainty and self-sufficient knowledge. Modesty should not be a capitulation, it should serve as a challenge – but always first as a challenge to ourselves.

References Barrow, J.D. 1999. Impossibility: The Limits of Science and the Science of Limits. London: Vintage. Beller, M. 1998. The Sokal Hoax: At Whom Are We Laughing? In: Physics Today, September: 29–34. Cilliers, P. 1998. Complexity and Postmodernism. Understanding Complex Systems. London: Routledge. Cilliers, P. 2000. Knowledge, Complexity and Understanding. In: Emergence 2(4): 7–13. Cilliers, P. 2001. Boundaries, Hierarchies and Networks in Complex Systems. In: International Journal of Innovation Management 5(2): 135–147. Cilliers, P. 2002. Why We Cannot Know Complex Things Completely. In: Emergence 4(1/2): 77–84. Dawkins, R. 2002. A Devil’s Chaplain: Selected Essays by Richard Dawkins. London: Weidenfeld & Nicholson. Dillon, M. (2000) Poststructuralism, Complexity and Poetics. In: Theory, Culture & Society 17(5): 1–26. Derrida, J. 1988. Limited Inc. Evanston, IL: Northwestern University Press. Derrida, J. 2000. Performative Powerlessness – A Response to Simon Critchley. In: Constellations 7(4): 466–468. Ellis, J.M. 1989. Against Deconstruction. Princeton, NJ: Princeton University Press. Emmeche, C. 1997. Aspects of Complexity in Life and Science. In: Philosophica 59(1): 41–46.

152 

 Paul Cilliers

Fleming, M. 1996. Working on the Philosophical Discourse of Modernity. Habermas, Foucault and Derrida. In: Philosophy Today Spring: 169–178. Greenough, P. 2003. Vagueness: A Minimal Theory. In: Mind 112: 235–281. Guillory, J. 2002. The Sokal Affair and the History of Criticism. In: Critical Inquiry 28: 470–508. Habermas, J. 1987. The Philosophical Discourse of Modernity: Twelve Lectures. Cambridge, MA: MIT Press. Haworth, A. 1999. Only One Cheer for Sokal and Bricmont: Or, Scientism Is No Response to Relativism. In: Res Publica 5: 1–21. Holland, J.H. 1998. Emergence. From Chaos to Order. Oxford: Oxford University Press. Kauffman, S.A. 1971. Articulation of Parts Explanations in Biology and the Rational Search for Them. In: Buck, R.C. & Cohen, R.S. (eds.). Boston Studies in the Philosophy of Science Vol. 8. Dordrecht: PSA/Reidel, 257–272. Luntley, M. 1995. Reason, Truth and Self: The Postmodern Reconditioned. London: Routledge. Najmanovich, D. 2002. From Paradigms to Figures of Thought. In: Emergence 4(1/2): 85–93. Norris, C. 1997. Against Relativism: Philosophy of Science, Deconstruction and Critical Theory. Oxford: Blackwell. Nuyen, A.T. 1989. Derrida’s Deconstruction: Wholeness and Différance. In: The Journal of Speculative Philosophy III(1): 26–38. O’Neill, J. 1995. The Poverty of Postmodernism. London: Routledge. Richardson, K. & Cilliers, P. 2001. What is Complexity Science? A View from Different Directions. In: Emergence 3(1): 5–23. Sokal, A. & Bricmont, J. 1998. Intellectual Impostures. London: Profile Books. Sweetman, B. 1999. Postmodernism, Derrida and Différance: A Critique. In: International Philosophical Quarterly XXXIX(1)/153: 5–18. Taylor, M.C. 2001. The Moment of Complexity: Emerging Network Culture. Chicago, IL: Chicago University Press. Wheeler, S.C. 2000. Deconstruction as Analytic Philosophy. Stanford, CA: Stanford University Press. Wood, D. 1990. Philosophy at the Limit. London: Unwin Hyman.

Paul Cilliers

On Derrida and apartheid How do we talk about apartheid? Certainly there is much to be condemned and much to be confessed, but is it possible to position oneself clearly in terms of these two options? Does it make a difference whether one speaks from within South Africa or from outside? Where was (and is) apartheid situated, what is the extent of our complicity in this evil, and who are “we”? These are some of the questions behind this paper. They will be investigated by looking at Derrida’s paper “Racism’s Last Word” (Derrida 1985) and its reception. Simultaneously I will also be concerned with a more philosophical problem: on what grounds is it possible to criticise a position from within the post-structural discourse? Is it possible to criticise Derrida, and more importantly, does his reaction to criticism remain within the parameters of the “ethics of discourse” on which he insists? As always, the theoretical and the ethical/political dimensions are interwoven. The discussion as a whole is part of an attempt to give meaning to the notion of responsibility in contemporary society. Since I will be critical of Derrida’s reaction to apartheid, the impression may arise that I am covertly trying to defend aspects of apartheid. Nothing could be further from the truth. Apartheid is evil. I do, however, want to argue that some intellectuals in the (North) West has often followed a strategy whereby they externalised the problem of apartheid. Apartheid was placed firmly “out there” in South Africa. This can be interpreted as an effort to displace their own responsibility for apartheid – irrespective of where it turns up. If we do not realise that we are all accomplices in the crime, it will not be possible to resist racism, in whatever form, effectively.

1 Words about Derrida’s “last word” In November 1983 an art exhibition, assembled by the Association of Artists against Apartheid, in cooperation with the United Nations Special Committee against Apartheid, was opened in Paris. Eighty-five of the world’s “most celebrated” artists contributed towards the exhibition. The aim was to form a “future museum against apartheid” (Derrida 1985: 290, translator’s note). The exhibition was to be “presented as a gift to the first free and democratic government of South Africa to be elected by universal suffrage”. Until such time the Association of Artists against Apartheid would assume, “through the appropriate legal, Originally published in the South African Journal of Philosophy, 1998, 17(1): 75–88. © South African Journal of Philosophy.

154 

 Paul Cilliers

institutional and financial structures, the trusteeship of the works”, which would meanwhile “be presented as a travelling exhibition to be received by museums and other cultural facilities throughout the world” (290). The condition set by Artists Against Apartheid has apparently been met since this gift has been handed over to the South African Government. As a matter of fact, it formed part of the opening of Parliament in February 1996 and it was exhibited in the House of Parliament for six months. A number of writers and scholars were asked to produce texts for the catalogue of the exhibition. Derrida contributed a text under the title Le Dernier Mot du Racisme. This was published, in a slightly modified version, as “Racism’s Last Word” in Critical Inquiry in 1985. The journal published a response by McClintock and Nixon a few months later, to which Derrida in turn responded. Now, after ten years, we can return to these texts with renewed interest. The event  – and should one not repeat that it was a miraculous event? – for which the Association of Artists against Apartheid hoped, has taken place. South Africa has a freely elected, democratic government. How much influence the Association of Artists against Apartheid had on the event is an interesting and important question that will not be investigated here.1 I am concerned here with a certain way of thinking about apartheid, not only ten years ago, but also now. What is at stake when someone attempts to speak the last word on apartheid, or at least, when someone hopes that the last word on apartheid can be spoken? How can we, each of us, wherever we are, speak about apartheid in a responsible way? I am sure that these were the same problems that concerned Derrida. In the paper under discussion, Derrida is concerned with racism as a global problem, but I am nevertheless convinced that he missed an opportunity. His analysis of the term “apartheid” is so specifically situated that it allows a reading of his text that marks apartheid as a strictly South African phenomenon. Derrida certainly did not intend the possibility of such a (mis)reading, but in his response to criticism of “Racism’s Last Word”, he increasingly opens up the possibility of such a reading. What is more, I will argue that in his eagerness to defend his “last word”, Derrida does not play by his own rules. Derrida’s short paper revolves around an exhortation: may apartheid remain the last name for the ultimate racism in the world. It is clear that he hopes that a time will come for apartheid to finally be abolished. He analyses the “singularity” of apartheid, showing how it took form in South Africa, and then shows how the West was not only intimately involved in its formation, but also in maintaining it.

1 As far as cultural influence is concerned, it would appear that the sports-boycott had a greater effect than any boycott or action on an artistic or academic level.



On Derrida and apartheid 

 155

He concludes by arguing that South Africa could serve as a kind of “computer” on to which the problems of Europe could be projected, and that what is at stake is nothing less than world peace. At first glance then, it is quite clear that Derrida sees apartheid as a serious problem of global concern. I will commence the discussion of Derrida’s paper by looking at McClintock and Nixon’s (1986) response to “Racism’s Last Word”. At the time they were students (from South Africa) participating in a Ph.D. programme at Columbia University. In their critique they take Derrida to task for making the word “apartheid” do too much work. By focusing on this single word, they claim, Derrida loses sight of the different historical circumstances in which the word operated and of the way in which it changed. Many other words were used by the Nationalist government, and since they claim that one cannot develop an effective strategy against racism in South Africa without taking all these shifts into consideration, they proceed to give Derrida a history lesson. It soon becomes clear, and is made explicit in the last section of their response (153–154), that their main objection is not to Derrida’s specific arguments in “Racism’s Last Word”, but to post-structuralism in general. They refer to it as a “method” (153) consisting of “monoliths” like logo-centrism and Western metaphysics and containing “bulky homogeneities” (154). Throughout their argument they perpetrate a frequent, but totally misguided criticism of deconstruction, namely that it is a-historical. Derrida (1986) reacts to all this with righteous indignation. His reaction takes the form of an open letter, a genre that he has used before. He makes no effort to hide his resentment and often employs a condescending, even insulting tone.2 He responds to McClintock and Nixon on three levels. In the first place he claims that they did not understand “Racism’s Last Word” properly because they did not take the context in which it was written seriously, and therefore confused the specific

2 Some examples: “you are not always serious” (156, 162); “I had no illusions in this regard [that the strategic value of Derrida’s call to action is limited] and I didn’t need to be reminded of it by anyone” (157); that they “could not or would not” respect the grammatical, rhetorical, and pragmatic specificity’ of Derrida’s utterance, referring to the first sentence of “Racism’s Last Word” (158) – a very strange claim for somebody that we have always understood to insist that utterances have a surplus of meaning. There are several more examples, like his instruction to them to check their facts (161), or sarcastic remarks like “thanks all the same” (162) and “you have the nerve [...] to write” (165). From the information provided with McClintock and Nixon’s response, Derrida would have known that they were graduate students, and one cannot help thinking that this kind of language sounds like a professor condescending to a student. The fact that this condescension took place in the pages of a prestigious academic journal has a distinctly unethical flavour. One feels that Derrida should either have ignored them (which they probably deserved), or that he should at least have respected other readers of his open letter by not being so self-righteous.

156 

 Paul Cilliers

grammatical and rhetorical devices it employs. In the second place he points to a number of more specific mistakes in their text – not all, “just the most serious and spectacular ones” (160). In the third place he reacts to their general criticism of deconstruction. Let us take these one by one. First, Derrida insists that McClintock and Nixon did not take proper account of the context in which his text appeared, an inexcusable mistake for somebody emphasising the importance of history. “Racism’s Last Word” was an eight-page text accompanying an art exhibition and he was not free to choose its dimensions. In this context the text “couldn’t be a historical or anthropological treatise, [...] it could only be an appeal” (157). Later on (167) he advises them, as he has advised others,3 to read his other texts to clarify all the issues that cannot be addressed in the present text. However, their insensitivity to the detail of this text and its context, Derrida claims, leads to another and even more unforgivable error: “If you have paid attention to the context and the mode of my text, you would not have fallen into the enormous blunder that led you to take a prescriptive utterance for a descriptive (theoretical and constative) one” (158). The utterance at stake is the first sentence of “Racism’s Last Word”: “Apartheid – may that remain the name from now on, the unique appellation for the ultimate racism in the world” (Derrida 1985: 291). This, Derrida (1986: 158) claims, “is an appeal, a call to condemn, to stigmatize, to combat, to keep in memory; it is not a reasoned dictionary of the use of the word apartheid or its pseudonyms in the discourse of South African leaders”. Derrida is not busy with an academic discourse, but with an ethical one. Since his two interlocutors “confused two verbal modes” (160), the history lesson is not only wasted (since Derrida seems to suggest that they did not teach him anything about the history of South Africa that he did not already know), but totally beside the point. In the next section Derrida points out a number of specific errors made by McClintock and Nixon. Most of this section is an argument about semantics (in the colloquial sense of the word) where McClintock and Nixon’s sloppiness is matched by Derrida’s fussiness. The central issue at stake here seems to be the meaning (yes, singular) of the word “apartheid”. McClintock and Nixon (1986: 141) insists that one should follow the word through all its shifts and euphemisms. It is strategically inappropriate to focus on a term that, already by the mid-fifties, “had become sufficiently stigmatized to be ostentatiously retired”. “So what”, Derrida (1986: 161) replies: “It is the thing and the concept they should have retired, not just the word”. Apartheid, called by whatever name, will remain despicable. Despite their rhetorical games, “those in power in South Africa [...]

3 Like Searle and Habermas (Derrida 1988: fn. 9, p. 156, 157), see below.



On Derrida and apartheid 

 157

have refused to change the real, effective, fundamental [sic] meaning of their watchword: apartheid” (163). This is why Derrida insists that “apartheid” is racism’s last word. As a “state racism” it “is famous, in sum, for manifesting the lowest extreme of racism” (1985: 293). Finally Derrida (1986: 167–170) raises the issue of McClintock and Nixon’s misunderstanding of deconstruction in general, and, as is often the case with these misunderstandings, the argument once again revolves around the notion of “text”. They seem to imply that it is not enough to “prize open certain covert metaphysical assumptions” but that one should “point to something beyond the text” (McClintock & Nixon 1985: 140). Derrida interprets this, correctly, as an allusion to his famous statement “there is nothing beyond the text”. A deliberately literal interpretation of this statement is no longer excusable. Derrida has discussed the meaning of this statement often enough, and he is quick to point this out: “an hour’s reading, beginning on any page of any one of the texts I have published over the last twenty years, should suffice for you to realise that text, as I use the word, is not the book. No more than writing or trace, it is not limited to the paper which you cover with graphisms” (Derrida 1986: 167). In sum, Derrida claims that McClintock and Nixon “quite simply did not read my text in the most elementary and quasi-grammatical sense of what is called reading” (157).

2 Responding to critics This “you did not read me (correctly?)” has become Derrida’s standard reply to his critics. In the case of Searle this accusation is certainly justified. Searle does not have a clue of what post-structuralism or deconstruction4 entails. He certainly has not read or understood Derrida properly. He is confident enough in his ignorance to thoroughly deserve the roundabout and ironic games that Derrida plays with him in Limited Inc. (Derrida 1988). In the Afterword to Limited Inc. Derrida makes the strong claim that Searle’s refusal to read him properly amounts to an unethical position, that his unwillingness to respond in a responsible way violates the “ethics of discourse”, that this violation contains a certain violence, sometimes directed against Derrida personally (see e.g. fn. 12. p 158). The points that Derrida make here are of extreme

4 Whatever the name for this body of thought is. Derrida is very cagey about these words. In his response to McClintock and Nixon (1986: 167) he claims that he has never used the term “post-structuralism”, and that it is the first time that he uses the term “deconstructionist”, only to “go quickly” (169). Since I cannot think of anything better. I will continue to use these terms.

158 

 Paul Cilliers

importance: you have to treat your interlocutors in a responsible way. This would at least involve the attempt to try to understand what they are saying to the best of your abilities, and to recognise that your involvement in a certain discourse cannot be separated from a certain commitment to it. A stance like this underscores the seriousness of Derrida’s work and shows that his approach is ethical through and through. Derrida levels the same reply – “you did not read me” – against Habermas. Habermas (1987: 185–210) accuses Derrida of making the “performative contradiction”: he has to make use of what he denies in order to deny it. For example, it is sometimes argued (incorrectly) that Derrida claims that “words do not have a determinate meaning”. If this is true, then the claim itself cannot be coherently understood. Derrida defends himself, adequately in my opinion, against this kind of argument in the Afterword to Limited Inc., and will not repeat his arguments here. What is at stake here is Derrida’s claim that Habermas did not read him. This is how Derrida (1988: fn. 9, p. 157, 158) formulates it: With stupefying tranquillity, here is the philosopher of consensus, of dialogue and discussion, the philosopher who claims to distinguish between science and literary fiction, between philosophy and literary criticism, daring not only to cnticize without citing or giving a reference for twenty five pages, but, even worse, justifying his non-reading and his atmospheric or hemispheric choices by this incredible alibi: “Since Derrida is not one of those philosophers who like to argue, [...] it is expedient to take a closer look at his disciples in literary criticism within the Anglo-Saxon climate of argument to see whether this thesis really can be held. [...] Such procedures still surprise me, and I have difficulty believing my eyes, in my incorrigible naïveté, in the confidence I still have, in spite of everything, in the ethics of discussion (in morality, if not moralism), in the rules of the academy, of the university, and of publication. For if Habermas had taken the slightest care to read me, or made any attempt to cite me, he would have seen (etc.) [...] [...] Is there a ‘performative contradiction’ more serious than that which consists in claiming to discuss rationally the theses of the other without having made the slightest effort to take cognizance of them, read them, or listen to them?

Derrida’s indignation is once again apparent, but by and large he is correct.5 Habermas should have known better than to rely on secondary sources when dismissing an important thinker, whatever his opinion of that thinker’s thought might be. The result is that Derrida can, correctly in my opinion, dismiss Haber-

5 The chapter in The philosophical discourse of modernity under discussion does contain a reference to Signature event context and to Limited Inc. but Habermas (1987: 194) finds the argument between Derrida and Searle “impenetrable” and relies on Jonathan Culler to explain it to him.



On Derrida and apartheid 

 159

mas on ethical grounds, and an important confrontation between two influential strategies of thinking does not take place.6 Is Derrida’s “you-did-not-read-me” reply as adequate in the case of McClintock and Nixon as in the case of Searle and Habermas? On first glance the answer is certainly yes. It is never really clear what McClintock and Nixon try to achieve with their argument other than to dismiss the post-structural approach, an approach they clearly do not understand fully. But I am not always sure if Derrida should get away with the way in which he dismisses them. Let us look at his arguments a little more closely in order to see if Derrida remains faithful to an “ethics of discussion”. Derrida’s first reprimand is that they did not take note of the context of his text – a few pages in the catalogue of an art exhibition. “By reason of its context and its dimensions (which I was not free to choose), by reason also of its style, it could only be an appeal” (Derrida 1986: 157, my emphasis). Now we all know that the context of a text is vitally important, but Derrida has been the one to remind us that the context can never fully determine a text and its meaning. This is one of the most important insights generated by his reading of Austin. Here is an important formulation: Above all, this essential absence of intending the actuality of the utterance, this structural unconscious, if you like, prohibits any saturation of the context. In order for a context to be exhaustively determinable, in the sense required by Austin, conscious intention would at the very least have to be totally present and immediately transparent to itself and to others, since it is a determining centre of context. The concept of – or the search for – context thus seems to suffer at this point from the same theoretical and ‘interested’ uncertainty as the concept ‘ordinary’, from the same metaphysical origins: the ethical and teleological discourse of consciousness (Derrida 1988:18)

Derrida’s argument, as I understand it, is that the idea of a determinate context is linked to the idea of a fully present consciousness. Neither of these can determine that a text can “only be” this or that. What is more, in the Afterword we read that “the simple recalling of a context is never a gesture that is neutral, innocent, transparent, disinterested” (131). Why would Derrida then insist that the context of this specific text – “Racism’s Last Word” – is so important? Let us be charitable and grant that he is merely making a straightforward pragmatic point: he was given so much space in the catalogue, and he had to write something that made sense in a catalogue. I am afraid, however, that this would not do. The context in which McClintock and Nixon, as well as the rest of us, encounter “Racism’s Last

6 Others do pursue the issue a little further, e.g. Fleming (1996).

160 

 Paul Cilliers

Word” is not in the catalogue, but in the pages of Critical Inquiry, an academic journal, one that is perceived to be prestigious. What is more, there is a note at the bottom of the first page – and we have been taught to be most attentive to marginalia. It reads: “A somewhat modified version of ‘Racism’s Last Word’ was originally published in the bilingual catalog of the exhibition”. Whatever “somewhat modified” means, we are not even dealing with the same text as the one whose context was the catalogue of the exhibition. If we are to take the context of the version of “Racism’s Last Word” before us seriously, then we must read it as a journal article and it should, according to the ethics of the game played here, conform to “the rules of the academy, of the university, and of publication” (157). Does this mean that I insist that Derrida should have produced a “historical or anthropological treatise”? Not at all. There is nothing wrong with an academic article, especially one with an ethical focus, in the form of an appeal. The point here is that Derrida was, by his own lights, a little more than unfair to chastise McClintock and Nixon for not realising what the correct context of his text was – a context that should apparently have fixed their reading of the text. Perhaps the context issue is less important than the way in which McClintock and Nixon actually read, or according to him, failed to read, Derrida’s text. Derrida is exasperated by the fact that McClintock and Nixon cannot even follow the simple grammatical structure of his text, that they confuse a prescriptive utterance with a descriptive one. Derrida refers, as we have seen, to the first sentence of “Racism’s Last Word”, a sentence that did not receive privileged attention in McClintock and Nixon’s response. Let us nevertheless grant Derrida that it is an important sentence, and look at it closely. “Apartheid – may that remain the name from now on, the unique appellation for the ultimate racism in the world” (Derrida 1985: 291). This is certainly a prescriptive statement concerning the name of the ultimate racism in the world, but is it only a prescriptive statement? Can we disentangle prescription and description with certainty? I thought that deconstruction has shown us just how complex this relationship is. For example, if one concedes that there is an equivalence between the prescriptive/descriptive distinction and the performative/constative one, it is, once again, Derrida’s reading of Austin that problematises these distinctions. As a matter of fact, Derrida acknowledges that Austin himself has already denounced the value/fact opposition (Derrida 1988: 15).7 Does the statement under scrutiny have this complex character? In my reading of it, certainly. It is prescriptive concerning the name of the ultimate racism, but the statement

7 Although I am quoting from the anthologised version of Limited Inc., these articles originally appeared long before “Racism’s Last Word”.



On Derrida and apartheid 

 161

is descriptive of what the ultimate racism is, namely Apartheid, that specifically South African thing, perpetrated by the Nationalist Government, its functionaries and accomplices. If you feel that this reading is too deliberate, just wait for a few paragraphs. On the next page (Derrida 1985: 292) we find the following sentence: “Apartheid is [my emphasis] famous, in sum, for manifesting the lowest extreme of racism”. If this is not primarily a descriptive statement, then I am incapable of reading any text, let alone Derrida’s. Is all this important? Yes. In the first place, Derrida again expects his interlocutors to be clear about a distinction that he knows is problematic. In the second place, it compromises Derrida’s own stance in a very specific way: the evil of apartheid is placed “out there in South Africa”. I will return to this issue presently. Let us turn briefly to Derrida’s insistence that the word “apartheid” be maintained as such. On a first level I think Derrida is correct. To say “Don’t say apartheid any more, but know that since 1948 there have been ‘three phases’ of racial policy in South Africa”, does miss the point. Apartheid was, and remains, evil under any name, and deserves “unconditional” rejection. But does that mean that we should not follow the word through all its historical transformations? Does deconstruction not teach us that the different ways in which a word is used, especially when trying to repress its meaning, can be marvellously revealing? The way in which the Nationalist government fumbled with words was at times so inept as to be comical (would the term “black humour” be racist in this context?), but more so, it was full of contradictions, especially when one compared what was being said for consumption inside South Africa to that which was said to the outside. Now, is one of the important strategies of deconstruction not exactly to look for such contradictions and then to use them in attempts to dismantle the system that imposes its hierarchies on our language and on our behaviour? To be able to do this, we should be extremely attentive to the masters’ language. What are we to do? Must we insist on the “fundamental meaning” of the word “apartheid” (let me repeat my discomfort with this formulation of Derrida’s), an untranslatable name for the evil perpetrated by them, “over there in South Africa”, or must we closely follow all the shifts, all the veils thrown over the word, all the little scrambles? Once again, Derrida has provided the answer himself: we must do both. Derrida’s general notion of “double writing” implies that our attempts to intervene contain the contradiction that our intervention can only be structured according to the terms used by that which should be dismantled. When this strategy is applied to a specific situation, “it must partition itself along two sides of a limit and continue (up to a certain point) to respect the rules of that which it deconstructs or of which it exposes the deconstructibility. Hence it always makes this dual gesture, apparently contradictory, which consists in accepting, within certain limits  – that is to say, never entirely accepting  – the

162 

 Paul Cilliers

givenness of a context, its closedness and its stubbornness. But without this tension, or without this apparent contradiction, would anything ever be done? Would anything ever be changed?” (Derrida 1988: 152). The above quote, surprisingly enough, is part of Derrida’s response to a question posed to him (in the Afterword to Limited Inc.) on his position in “Racism’s Last Word”. The question results from “American commentators” referring to the implications of expressions like “to call a thing by its name”, “massively present reality”, “grammatical specificity”, the specificity of context, etc. These issues are of course closely linked to the discussion above, but I think Derrida answers them on a different level. At this point he deserves to be quoted at some length: I have never ‘put such concepts as truth, reference, and the stability of interpretive contexts radically into question’ if ‘putting radically into question’ means contesting that there are and that there should be truth, reference, and stable contexts of interpretation. I have – but this is something entirely different – posed questions that I hope are radical concerning the possibility of these things, of these values, of these norms, of this stability (150). The ‘commentators’ whom you evoke would, as you suggest, have totally ‘mistaken’ the ‘implications’ of my discourse in general and what I have said of apartheid in the particular context to which you refer. They commit the same “mistake” as those to whom l respond in Critical Inquiry. I consider the context of that discussion, like this one, to be very stable and determinate (151).

Derrida’s reply here is aimed at the general (mis)understanding that deconstruction asserts a total free-play of meaning, an endless instability that has to imply that anything goes. That has certainly never been Derrida’s position. This caricature of deconstruction has caused much damage and should be dismissed emphatically. Derrida has al ways argued that while meaning or context is never saturated, there is always meaning and context, and that in some cases, like the natural sciences, they are quite determinate. This determination results from “a context which is extremely vast, old, powerfully established, stabilized or rooted in a network of conventions (for instance, those of language) and yet still remains a context” (136).8 These networks are nevertheless never neutral, they can never escape the dynamics of power and enforcement. As far as his writings on apartheid are concerned, Derrida claims that the context was determined enough “that one might count on ties that are stable, and hence demonstrable, linking words, concepts and things, as well as on the difference between true and false. And

8 The whole section surrounding this quote provides important clarifications on the misunderstanding that deconstruction entails relativism.



On Derrida and apartheid 

 163

hence one is able, in this context, to denounce errors, and even dishonesty and confusion” (151). Does this suffice to justify Derrida’s response to McClintock and Nixon? It goes a long way to justify his answer to them. He is not merely busy playing with words, but taking a firm ethical position on an abomination. However, I am not so sure that the way in which he does it is a responsible one, especially as far as all the other readers of his open letter are concerned. Let me be forthright. A close reading of these documents creates the impression that Derrida was so eager to dismiss his critics that he scattered landmines all over a territory that many of us, sympathetic to the strategies of deconstruction, often find difficult to defend. l think that it is possible, if not for McClintock and Nixon, then at least for some other readers, to be a little more than unhappy with how Derrida did what he did in this case. In sum, his reaction was not responsible. I suppose that this would have been, rhetorically, a nice moment to stop, but I feel that I have a responsibility to justify my accusation that Derrida does not fully accept his responsibility, or at least, that he is reluctant to walk the second mile. Let us say that McClintock and Nixon deserved the reply they got, that Derrida adequately defended his response to them, that his response was an occasional piece within a specific context and that we have many other works by him to read (if we can read) if we really want to know what his position is. Would that mean that the last word on “Racism’s Last Word” has been spoken? I am afraid not. Linked to this fear is the nagging doubt that it may not be possible to criticise Derrida at all. What would one have to do to prove him wrong, or rather, how should one attempt to criticise him; not refute him in general, just show that somewhere, in terms that he would accept, he got it wrong or that his perspective was limited? Why would one want to do that? To have the intellectual pleasure of deconstructing Derrida, like some kind of party trick? To prove that your own insights are more important? To free oneself from the stigma clinging to white South Africans? To make philosophical progress? To produce another paper under the pressure of the academic system? Or perhaps as a kind of Freudian patricide to free yourself from somebody from whom you have learned so much? I will not declare my innocence on any or even a combination of these factors, but there are further considerations. The issue of apartheid was, and is, too important not to be taken seriously. Since the last word on apartheid has not been spoken, and probably never will be, I will return to “Racism’s Last Word” for one more reading. Some aspects of this reading may seem too deliberate, may be revealing of my own indignation, and are perhaps at times even unfair. Derrida could  – with some justification – claim that he never intended such a reading of his text. However, committing oneself to a certain (ethical) position, involves a certain

164 

 Paul Cilliers

risk. My target here is not the person Jacques Derrida, but those who preferred to see apartheid as something perpetrated only by a specific group of (white, South African) people.

3 Another reading of “Racism’s Last Word” “Apartheid  – may that remain the name from now on, the unique appellation for the ultimate racism in the world, the last of many”. Derrida’s appeal is clear: may apartheid remain the name, the last name, of the ultimate racism in the world. The question that kept bothering me as I read the article was the following: what exactly is “apartheid” the name of? One would assume that it stands for the general concept of racism, for an evil that should be fought wherever it is found. Nothing in the first section contradicts this assumption, as a matter of fact, a question like “But hasn’t apartheid always been the archival record of the unnameable?” (Derrida 1985: 291) seems to support it. However, it becomes clear as one continues that Derrida is only talking about the specific form of racism adopted by the Nationalist Government in South Africa – that that is the “ultimate racism”. One could (perhaps should) analyse the specific examples of racism in South Africa which he chooses in order to illustrate the extent of the evil of that system. Here is one of several citations that could be used to show that he is talking about a “singularity”, that apartheid is a very specific concept belonging only to the South African context: [The exhibition by Artists Against Apartheid] gives warning: Do not forget apartheid, save humanity from this [my emphasis] evil, an evil that cannot be summed up in the principial and abstract iniquity of a system. It is also daily suffering, oppression, poverty, violence, torture inflicted by an arrogant white minority (16 percent of the population, controlling 60 to 65 percent of the national revenue) on the mass of black population (293).9

At this point Derrida could object strongly. He could insist that he was explicit about the fact that racism is a “Western thing”, that he went to great pains to show how apartheid could only be maintained with the complicity of the West, or even stronger, that it was made structurally possible by the theologico-political

9 I will not comment on the possible, but grammatically correct, reading of this sentence that would imply that the white minority (all of them) physically tortured the black population (all of them).



On Derrida and apartheid 

 165

history of Europe (294–296), that it also involves contradictions within Europe. This is however, not the issue. Despite this structural involvement, Derrida is still only talking about the complicity of the West in maintaining apartheid out there in South Africa. Symbolic condemnations [of apartheid], even when they have been official, have never disrupted diplomatic, economic, or cultural exchanges, the delivery of arms, and geopolitical solidarity. Since 1973, apartheid has been declared a ‘crime against humanity’ by the general Assembly of the United Nations. Nevertheless, many member countries [...] are not doing all that’s required [...] to put the Pretoria regime in a difficult situation or to force it to abolish apartheid. This contradiction is sharpest no doubt in today’s France, which has provided more support for this exhibition than anywhere else (295).

The contradiction referred to here is that France were fighting and maintaining apartheid in South Africa at the same time. The contradiction about which I am concerned is that France (and others) were fighting and condemning apartheid in South Africa while it was rife in their own midst. At this point it would be possible to object. Apartheid, it could be argued, is the name of that specific state-racism perpetrated by the Nationalist government in South Africa, and should not be equated to the notion of racism in general. Nobody denies that there is racism in France, but that is not the same as apartheid. Derrida seems to support this argument: THE LAST, finally, since this last-born of many racisms is also the only one surviving in the world, at least the only one still parading itself in a political constitution. It remains the only one on the scene that dares to say its name and present itself for what it is: a legal defiance taken on by homo politicus, a juridical racism and a state racism (292).

Although l agree that the two words – racism and apartheid – do not have the same meaning, I wish to maintain that there is a serious danger involved in reserving the notion of apartheid for that specifically South African thing. If “apartheid” refers to racism formally structured in a social system, would it be possible to say that since a certain community does not have explicit laws excluding the racially other, that it is free from apartheid? My answer to this question is no. In the first place, what would such a distinction really mean? Is it meaningful, even possible, to say “we are racists, but we do not have any apartheid”? In the second place, the systematic exclusion of the racially other through (explicit or implicit) structures in society forms part of many cultures, especially in Europe. If pressed for examples of cultures which close ranks against those who are foreign, many would choose France or Germany. In the third place, an analysis of the relationships between Europe and her colonies, or a study of current European immigra-

166 

 Paul Cilliers

tion laws, would probably expose a fair amount of legal machinery working to keep different peoples apart. Apartheid, as a modernist strategy to structure and control, was never confined to South Africa. What is even more disconcerting than reserving the notion of apartheid for this specific instance of racism, is the way in which it was condemned, as if from a moral high ground. One of the most telling statements in “Racism’s Last Word”, the one that revealed the extent to which blame was transferred from the inside of Europe to the outside of South Africa, was the following: “The white resistance movement in South Africa deserves our praise” (297). In the first place, who was (or is) the “white resistance movement in South Africa”? Those public heroes selected by the international media for glowing adoration,10 or those who fought the nitty-gritty injustices of apartheid in their day-to-day existence? More importantly, who are these people that find themselves in a position to hand out praise in this matter? Is the “our” in the statement a pluralis majestatis? Does it refer to the average Parisian art-lover, strolling through the exhibition with a glass of champagne in hand (which is what the specific context on which Derrida insists for this text – part of the catalogue for an exhibition – would suggest)? Or does it refer to the class of intellectuals who have been analysing South Africa from a neutral vantage point, enabling them to decide who are worthy of praise? Perhaps there is a more benevolent interpretation of this statement, but I fail to see a way of reading it that does not imply a divide between those worthy of praise and those who feel that they are in a position to hand out praise. In the context of apartheid, in South Africa or anywhere else, the handing out of praise (which is not the same as declaring one’s solidarity) seems to imply that some have the ability to escape the messiness of interaction with the other, to reach some higher ground where they are morally safe.11 We are still trying to figure out why anyone would refer to apartheid in South Africa as the “ultimate racism” without condemning, referring, or comparing it to any other specific form of racism. Could it be that he knows apartheid was the “ultimate” because one could measure its evil in terms of something like lives lost, or suffering caused to the most people? On these criteria it would not be difficult to find worse cases of slaughter and brutality in the history of the world.

10 The harsher light of post-apartheid reality has confirmed the enormous stature of some of these figures but has revealed clay feet as well. 11 The impression that some form of moral high ground is possible, is in this case reinforced by the art exhibition itself. Only a single, exiled South African artist (Gavin Jantjies) is represented, as if to say that apartheid is best condemned from outside. Furthermore, many of the works are incidental pieces that have nothing to do with apartheid.



On Derrida and apartheid 

 167

This can obviously not be what Derrida means. Slaughter and brutality is not measured on a scale so that one can talk about “bad” brutality, “worse” brutality and “ultimate” brutality. Nor is racism. So the question remains: why, for what reason, would someone transpose the worst evils of racism onto a single term, and then situate that term in one specific context outside of himself, even if he is correct on a purely descriptive level? Once again deconstruction itself suggests an answer. By externalising the evils of racism, the West made South Africa the scapegoat, the pharmakos,12 which can be punished outside the walls of the city as a symbol of the evils perpetrated within the walls. As long as the racism in South Africa was so obviously evil, it was easier to denounce that than to confess all the instances of apartheid forming a structural part of one’s own society. I am convinced that Derrida never intended to perform any such externalisation himself and that he will, and has, denounced racism in any form. It is ironic that “Racism’s Last Word” could be read as if doing just that, but that his defence of that text was presented in terms that made it possible to read it as an example of an attempt by the West to deny their complicity in perpetrating apartheid themselves. Is it fair to say that the West, the rest of the world, for that matter, perpetrates apartheid? I think so. Not on a constitutional level of course, but the interaction between different religious, linguistic, ethnic, racial and sexual groups everywhere is, and was, fraught with cruelty. What is more, the disappearance of constitutional apartheid in South Africa has been accompanied, if anything, by a rise in racial hatred in the rest of the world,13 also in Europe and I am not only referring to the situation in Eastern Europe. Dealing with this universal problem responsibly is immensely difficult. How do you celebrate difference and diversity without introducing domination, superiority and suspicion? Certainly not merely by insisting that somebody else behaves worse than you do. By maintaining that apartheid is a universal problem, I do not intend to exonerate anybody in South Africa, nor do I wish to create the impression that what happened here – and that what is happening here now – is not important. I am in full agreement with Derrida as far as the concluding section of “Racism’s Last Word” is concerned (Derrida 1985: 297–299). Here he describes South Africa as a place where world history has been “concentrated” in such a way that it serves as a kind of laboratory (“computer”) for the rest of the world. The problems of heterogeneity in South Africa are no different to those that the rest of the world has to face, but they are extreme. If the South African situation remains hopeless, it

12 This figure is discussed in “Plato’s Pharmacy” (Derrida 1981). 13 For a discussion of the rise of international apartheid, see Richmond (1994).

168 

 Paul Cilliers

will communicate a depressing message to all those studying this “computer simulation”. What then, has happened in South Africa since “Racism’s Last Word”? Now, ten years later, that which Artists against Apartheid, Derrida, and many of us have hoped and worked for, has happened. South Africa no longer has legally entrenched apartheid. The artists’ gift has been handed over to a freely elected, democratic government. Does this mean that here, in South Africa, the last word on apartheid has been spoken, that apartheid is dead? Unfortunately not. It may be legally dead, but its spectre is still haunting us. Racism continues to abound, as it does everywhere else. It is also not surprising that inverse forms of racism flourish in the shadow of the ghost. But South Africans are also speaking some new words on apartheid. They are struggling with diversity, and sometimes finding ways of coping with it. We are talking to – and arguing with – each other. There is a battle to find ways of coming to terms with our terrible history. Perhaps, somewhere in this struggle, there are the beginnings of a gift that South Africa may eventually have to offer to the rest of the world. And this exhibition, this gift that has now reached its destination, that has been accepted and unwrapped, that has been taken up in the economy of giving and taking?14 Perhaps it would have been better if the South African Government did not accept it, if it remained an appeal that travelled around the world, visiting many places, reminding them not only of that specific evil that was South Africa, but also of the continued evils of prejudice and hate right there where the exhibition happens to find itself. In that capacity it would still have had to return to the new South Africa often enough.

References Derrida, J. 1981. Dissemination. Chicago: Chicago University Press. Derrida, J. 1985. Racism’s Last Word. In: Critical Inquiry Autumn: 290–229. Derrida, J. 1986. But, beyond ... (Open letter to Anne McClintock and Rob Nixon). In: Critical Inquiry, Autumn: 155–170. Derrida, J. 1988. Limited Inc. Evanston: North Western University Press. Derrida, J. 1992. Given time: I. Counterfeit money. Chicago: Chicago University Press.

14 For his reflections on the logic, economy and responsibility of the gift, see Derrida 1992 and 1995. For the catalogue prepared for collection to be exhibited in South Africa, Derrida again wrote a brief piece entitled “Luminous sign” (Derrida 1996). It does not throw new light on any of the issues discussed in this paper, except to confirm that the gift has reached its final destination. He claims that “the access to democracy and universal suffrage in South Africa mark an indispensable and irreversible step forward” (10).



On Derrida and apartheid 

 169

Derrida, J. 1995. The gift of death. Chicago: Chicago University Press. Derrida, J. 1 996. Luminous sign. Art against apartheid/78 artists in the 80’s, Association Francaise d’Action Artistique, Ministère des Affaires Etrangères, 1996, 10–11. Habermas, J. 1987. The philosophical discourse of modernity. Cambridge: Polity Press. Fleming, M. 1996. Working in the philosophical discourse of modernity. Habermas, Foucault, and Derrida. In: Philosophy Today, Spring: 169–178. McClintock, A. & Nixon, R. 1986. No names apart: the separation of word and history in Derrida’s ‘Le Dernier Mot du Racisme’. In: Critical Inquiry, Autumn: 140–154. Richmond, A.H. 1994. Global apartheid: refugees. racism, and the new world order. Oxford: Oxford University Press.

Paul Cilliers, Willie van der Merwe & Johan Degenaar

Justice, law and philosophy An interview with Jacques Derrida

Jacques Derrida visited the University of Stellenbosch on 12 August 1998. On this occasion he was interviewed by Paul Cilliers, assisted by Willie van der Merwe and Johan Degenaar.1 Paul Cilliers: Before we start, a brief remark on the audience that would read this interview. The South African philosophical community is fairly heterogeneous. There used to be a quite strong divide between the continental and the analytical traditions. This division seems to have become less marked, but there will be readers that know quite a bit about your work, and also some who know nothing; there will be those who read favourably, and those who don’t. My questions will thus have a certain generality – although I promise I will not ask “do you really mean that the world out there is a book”. Jacques Derrida: I would answer that! Paul Cilliers: Perhaps we already have your answer on that. For now I have three questions: one on justice, one on law, and one on philosophy. May I start with the first one? Deconstruction is generally perceived to be an intervention that does not lead to final or completed results. That which has been deconstructed can always be deconstructed again. Moreover, for something to be intelligible, to have meaning, it must in principle be deconstructable. Yet you refer, in “The Force of Law”2 for example, to justice as something that is not deconstructable. Could you explain what you mean there? Does such a position not reintroduce a kind of metaphysics? Jacques Derrida: Thank you. First of all, I would like to thank you again for your hospitality, for the opportunity you give me, and for discussion with you

1 The recorded interview was transcribed by Yvonne Malan. The text was edited lightly in order to produce a written text from a spoken one. [Editor] 2 Paper delivered at a symposium entitled “Deconstruction and the Possibility of Justice” held at the Benjamin N. Cordozo School of Law, Oct. 1–2, 1989. Published in the Cardozo Law Review, vol. II, July/Aug 1990; and in Deconstruction and the Possibility of Justice, Cornell et al., Routledge 1992. [Editor] Originally published in the South African Journal of Philosophy, 1999, 18(3): 279–286. DOI: 10.1080/02580136.1999.10878189 © South African Journal of Philosophy.

172 

 Paul Cilliers, Willie van der Merwe & Johan Degenaar

and South African philosophers. Before I try to answer these difficult questions, I want to point out that the heterogeneity that you mentioned is not only a South African feature, not a particular feature, I reckon, of this community, only of this community. Everywhere in the world now we have this heterogeneity between at least two traditions, each one being in itself, already differentiated. So, one of the tasks, of the duties we have as philosophers, is precisely to try very hard – not to bridge the gap – but to understand what is going on in this quasi-war, what a philosopher’s translation of the traditions would be. I myself, as you know, am rather on one side – by accident, by culture, but by the fact that I received my training in France, and so on and so forth. So, I must confess that it is very difficult for me to have an important access to another philosophy. And it is not a question of content, of theses, of political positions, a question of style – which they do of course have – but a way of arguing, a way of writing, a way of asking questions that is very difficult to locate. It is not a matter of national language, but of language in a more obscure way. So, I am grateful to anyone who makes an effort to try and change the situation. And when I confess to being power less, it is just a confession, one of my many limits. Now, trying to address the questions you asked, I would say this. I do not doubt the definition of deconstruction that you gave: that it is a way of intervening, of intervention. It is not simply a doctrine, not a system, not even a method, but something which is tied to the event. When I have to summarise very briefly what deconstruction is, and should not be, I often say: Deconstruction is quite simply what happens. It is not simply the theoretical analyses of concepts, the speculative desedimentation of a conceptual tradition, of semantics. It is something which does something, which tries to do something, to intervene and to welcome what happens, to be attentive to the event, the singularity of the event. That is why deconstruction happens as soon as something happens. It did not appear in the twentieth century, nor as a modern movement in the academy in the West. No, I think in every event, not only philosophical, in every cultural event there is some deconstruction at work, something which displaces and opens a structure, a set of actions, to singularity, to something other, to some alterity, to some unpredictable future. This comes at the same time as a threat and as a stroke of luck. You mentioned the word “intervention”, a way of trying to transform a situation. (I am trying to keep your own words when I answer you.) Deconstruction is more and more tied to what is referred to in speech act theory as activity, as the performative, something pragmatic, to the idea of work. From the beginning the concept of trace had to do with working – something which produces a new goal, a way of making and doing something. However, when you define deconstruction as something that does not necessarily produce a result, I would say



Justice, law and philosophy 

 173

yes and no. I would say yes to the extent that the result, the effect, could not be programmed. There is no criterion for the fact of a decision, of the act of writing, thinking, production. To that extent deconstruction is more and more a way of thinking what responsibility and decision should be. Decision and responsibility worthy of these names should not be controlled by previous knowledge, it should not be programmed. That does not mean that we have given up knowledge. Not at all, on the contrary, we have to know all that we can know. But we should also know and think that between the act of knowledge, between science and the act itself, the decision, there will be a gap, there is a heterogeneity between knowing and doing. So when I make a decision, when I take responsibility, to some extent it must be in the night. I have to prepare a decision to know where I can go as well, as consciously as possible, but I should acknowledge that between the accumulation of knowledge and the moment I make a choice, I take a responsibility, I make a decision, there is an infinite abyss because of the heterogeneity of these moments. That is why I also constantly insist on the undecidability, which does not mean that you are simply paralysed and neutralised because you do not know what to do. Simply, in order for a decision to be a decision it has to go through a moment where, irrespective of what you know, you make a leap into the decision. This leap into the responsibility is an infinite one and you take a decision only in a situation when there is something undecidable, when you don’t know what to do. You don’t know. That is, if you knew what to do, there would be no decision, you would have already done ... Paul Cilliers: You would have calculated ... Jacques Derrida: You would have already known. So, this is true for ethical, juridical and political decisions, even for existential decisions. You have to go through an ordeal of undecidability in order to decide. So, to that extent the result, by definition, is unpredictable, unknown  – unpredictable if by predictability you mean knowledge, you mean calculation. Something must remain incalculable for a decision to be a decision. That is why it is an intervention which has – because it is not linked essentially to knowledge – something obscure, something even mystical. I have no objections to people who define this decision as something mystical. Paul Cilliers: That looks like another question that we will not have time to ask you today. Jacques Derrida: I use the word cautiously, to mean something that is not controlled by a scientific knowledge, nor even by an ontological knowledge. The threat of the leap I mentioned, not only threatens a break with science in the strict sense, but with philosophy as ontology, as knowledge. So, you see that the

174 

 Paul Cilliers, Willie van der Merwe & Johan Degenaar

stakes are very high, that this is a break with ontology itself. Saying this, I know how enormous it may sound, but I take my responsibility. Now, it has even become the case that what is called deconstruction, in its many, many forms, appears on the stage as a set of theories, philosophies, concepts to be manipulated, and taught as a form of knowledge. Of course, to that extent they are subject to deconstruction too. To this process you can never put an end, you must keep on asking these necessary questions. But this endlessness does not mean what Hegel would call the dialectical. Even though deconstruction is endless, the injunction to intervene, to take responsibility is here and now absolutely urgent. You can’t wait. So, deconstruction is endless, but you have to respond here and now to the leap, to the gap I mentioned earlier. Let us refer what I have just said to the question of justice. Since everything in philosophy is deconstructable, is offered to deconstruction, subject to deconstruction, why did I say that justice is undeconstructable? We could have a whole interview on that, I will try to be brief. I could of course justify this, as I do in the text you refer to [“The Force of Law”], by appealing to let’s say, contextual support. I was speaking in the context of legal theorists, a conference on law and I wanted to respond to the common objection that those sympathetic to deconstruction are nihilistic or relativistic, that we could not care less about justice. So I insisted that it is exactly the opposite: deconstruction is on the side of justice, not on the other side. But, in order to say so, I had to do justice to the concept of justice. Therefore, I had to distinguish between justice and the law, the positive set of jurisprudential concepts. My assertion was that the history of legal concepts, the history of the law – which is of course moved by some idea of justice – is never adequate to justice itself. That is why there is a history of law, a history of legal systems, of institutions, of constitutions, transformations, political transformations, of legal systems, of national and international law. So, I had to account for the history of law, and in order to account for the history of the law you have to assume that the idea of justice – which is at the horizon of this history – is distinct from the law. Nevertheless, I did not want to accept this distinction in that way, that is to say in a passive way which suggests that there is this regulative idea of justice and that we make an elaborate effort to adjust, to get closer and closer to this regulative idea of justice. No, I was saying that this idea of justice is not the infinitely remote idea of a goal to be reached, but it is something which, here and now, gives us orders beyond any given set of legal concepts. Why did I say at the time that deconstruction is justice? It is difficult for me to reconstitute the argument here. At the time I was beginning to meditate on the origins of the idea of justice, both the Greek and the biblical idea of justice. I read what Heidegger said about dike: dike as harmony, as “Versammlung”, as a gathering of something, something which gathers in a form of harmony. I tried to



Justice, law and philosophy 

 175

question this interpretation by Heidegger, and to propose the idea that justice is not necessarily this achievement of gathering. Since justice supposes the relation with the infinite alterity of the Other, it implies itself a break, an interruption, a disassociation. So, I pose the idea of disassociation, and of difference, if you want, to Heidegger’s interpretation of dike as “Versammlung”. However, to define justice as disassociation is very uncomfortable, difficult, because you are saying that in order to be just, you must not integrate, but be open to the infinite rupture, break, the distance between the singularity of the Other and another Other – which means that justice in itself remains necessarily inadequate to any knowledge, any concept, any criterion, that in itself, it is deconstructive. Justice is deconstructive. It is in the name of justice that we deconstruct. Because there is an infinite gap, an infinity that is an infinite interruption – what Blanchet or Levinas would call “un rapport sans rapport”, a relation without relation – you have to accept the idea that justice implies some interruption, that it has in itself a deconstructive force, if you will, an energy, or a pulse, a drive. Now, since any drive to deconstruction must be inspired by something, I will call this justice. It is in the name of justice that we do what we do when we deconstruct. Why do we do it? What is our interest in deconstruction if not a respect for precisely what we call justice? That is why there are revolutions, that is why we criticise and deconstruct the given systems of norms in legal systems, in politics, in ethics, in social structures and so on and so forth. So, to say that deconstruction is justice is not a quiet equation. It does not mean that deconstruction is just. Justice is deconstructive, deconstructing. So, it is a rather uneasy status. Paul Cilliers: Yes, I think to make it deconstructing, to make it a verb instead of a noun, helps to understand the claim. Willie van der Merwe: I would like to return to the notion of the mystic moment in decision ... Jacques Derrida: The subtitle of the text which I mentioned [“The Force of Law”] is “The mystical foundation of authority”, a quote from Pascal ... Willie van der Merwe: It is about “beyond” and about the status of “beyond”. On Monday3 you coined the term of a hyperbolic ethics. It seems to me, more explicitly in your later work, that you are trying to design, to work out – not find, because that would be a contradiction – an ethics beyond ethics. Various notions play a part in this beyond: the unforgivable, the ineffable, silence. It seems to me

3 On Monday, 10 August 1998, Derrida gave a lecture at the University of the Western Cape entitled “Forgiving the Unforgivable”. [Editor]

176 

 Paul Cilliers, Willie van der Merwe & Johan Degenaar

to be beyond the limits of language, if there is something beyond the limits of language, if one can say what is beyond the limits of language. So, this notion of transcendence, in a sense, which I again hear when you speak about justice as being that in the name of which we deconstruct ... I would like you to say something more about the status of this. The last part of Paul’s question was about metaphysics. Perhaps he can put the second question now, keeping this issue in mind ... Paul Cilliers: Well, in a certain sense you have already touched on the second question, since it concerns the relationship between law and justice. I understand when you say that the law cannot guarantee justice. But what does that say about our attitude towards the law? Perhaps if I can be a little more specific: South Africa has a new constitution. We are all proud of it. It seems to be generally accepted as a serious response to the call of justice. Yet, our country is also suffering from an extremely high level of crime and violence. How does one think about the status of the law, from the perspective of deconstruction, under these conditions? Willie van der Merwe: And also, what role is played in this by metaphysics – that is the link with my question. Jacques Derrida: Let me put it this way: as you know, there are many definitions of metaphysics: it might mean, for instance, being as being, being as presence, as ontology, and so forth. In that case, what I am trying to make is a deconstruction of metaphysics. Now, if you understand metaphysics the way Levinas does, as something which is precisely beyond ontology, defining ethics as counter move against ontology, then what I am doing would be closer to, but not identical, to what Levinas does. I will leave aside this question of terminology, and try to answer your question about the possible return of some metaphysics in the proposition according to which justice is beyond the law. Here I would say no. I would say no, because it is not, for the reasons that I have just given, the reconstitution of a drive towards ontology. This is very difficult to discuss – there is no such a statement in Levinas as such, and the way he would define justice is different from the way I do it. But if metaphysics means in both cases a philosophical gesture, then deconstruction, in its strange relationship to justice, is not metaphysical as such. It is not, it cannot produce any metaphysical system. What I have said about deconstruction and justice has nothing to do with the production of a consistent set of systemic propositions. So, I would not be against metaphysics, of course, but I would not say that this gesture, this transcendence of the justice in relation to the law is in fact metaphysical. You can of course do that if you wanted to, you can write a treatise of metaphysics around this axiom. I would not be interested in



Justice, law and philosophy 

 177

that, but I can imagine a book, an interesting and consistent book on “Metaphysics of Justice”, “Deconstructive Metaphysics of Justice”. Why not? This is possible, but it is not my main interest. Paul Cilliers: But you use the notion of quasi-transcendental. Isn’t that a gesture in the direction of metaphysics and then standing back again? Jacques Derrida: Quasi ... that is quasi-metaphysical. Paul Cilliers: Yes. Jacques Derrida: We would have to spend a long time around this, the way I use the word quasi, quasi-transcendental, quasi this or that. This is of course a central point, but we are going too fast ... I would rather like to say something about the idea of foundation. There is no foundational gesture in this case. What appears here is an abyss rather than a ground. The decision and responsibility has to be taken in experiencing the abyss that is infinite, and unpredictable. Now, an important point as to the urgent injunctions of our work in these times, and especially in this country. When I said that there is a difference, a heterogeneity between justice and law, I would add this point, which I think is decisive: this distinction is not a distinction between two terms, between two poles, as if we had on one side justice, and on the other side Jaw. No, they are two, but they are one. It is impossible to think justice without including in it the injunction to determine justice by the law, that is, to produce just laws. These two poles are infinitely removed from one another, infinitely heterogeneous, but the law must be inspired by justice, it is part of its concept, and justice must command the production of determined laws. So they are linked, they are indisassociable: infinitely different, yet indisassociable. That is why it is important to have constitutions. As you know, we have more than one. In France we have had I do not know how many  – I think we have the record number of constitutions. And here in South Africa you first had a previous constitution, then a provisional constitution, then a last provisional constitution, and you will have to improve the one you are producing now. This means that there is a history of the law, and hopefully progress – and Kant will say there is progress. Even if the revolution does not succeed, even if there are regressions and so forth, the idea that there is a revolution is a sign of the possibility of irreversible progress. I think that the abolition of Apartheid is, of course, not the end of things, but the beginning of something else, a process, an endless process. It is the sign of an irreversible progress, I would say, the possibility of an irreversible progress. Of course, it is in the name of justice that you are improving your constitution, and it is because these constitutions are inadequate to justice that you will have to improve them, to adjust them to the social progress of this country. The legal abolition of apart-

178 

 Paul Cilliers, Willie van der Merwe & Johan Degenaar

heid is not everything. South Africa is a deconstructing country, it is a country which is perhaps deconstructing, more than any other country. Here you have a major example of something with a certain rhythm – this is very important, I take it very seriously – that deconstructs the current state law, state racisms and produces new law – which is, of course, inadequate to justice – and then there is a new social structure with a number of problems of its own, that we cannot deny, and which will require new adjustments and new constitutions and so on and so forth. You have here the experience of events – as a deconstructive force – which compels you, compels us, to adjust the law, adjust the constitution, to do justice, to experience the infinite gap. We can experience this every day, we just have to drive in the country to see what remains to be done. So, I will insist on this separation of law and justice; they are two and one ... Johan Degenaar: In your terminology you would not say that there is a dialectic? Jacques Derrida: No. Johan Degenaar: Do you oppose the use of the word dialectic? Jacques Derrida: Yes, in the majority of the contexts, unless I really find the appropriate context. The common use of the notion I would avoid, exactly because of this infinite gap. But there is a point for a reconciliation to take place: you have to have an infinite difference, otherwise you would not need a reconciliation. Just one word on the notion “hyperbolic”. I often use this word for my own purposes, but it is a quote from Jankelevitch.4 I find that I agree with his use of this word “hyperbolic”. Paul Cilliers: Since we are running out of time, a last question that will perhaps bring things to a close for now: as you know, deconstruction is often criticised for not giving proper practical guidelines for dealing with the messiness of the world. I suspect that this kind of criticism is in many ways a criticism of philosophy in general, rather than of deconstruction specifically. Given your involvement in the promotion of philosophy and its education, how would you respond to the argument – an argument often encountered in the South African context – that philosophy is a luxury, that philosophy, and the teaching of philosophy, cannot be a priority? Jacques Derrida: Two or three points, very briefly. First, you are right, philosophy, as well as deconstruction, is often charged with doing nothing, with giving

4 In Le Pardon (1967). See also “Should We Pardon Them?” and the preceding introductory remarks in Critical Inquiry 22(3), 545–572, Spring 1996. [Editor]



Justice, law and philosophy 

 179

no practical advice. I would say only this: if it were true about deconstruction, why have so many people been so anxious and angry about it for such a long time? They could have said “well it’s OK, it’s something good for the library, for the university”. But I think that they are so nervous about it because they realise that there is some practical injunction here  – an injunction that is sometimes difficult to understand, sometimes difficult to translate, but that makes everyone uneasy. I think this is very practical, very political, and if we had time, I would show some of the practical effects of this uneasiness. As to the teaching of philosophy, well, I did my best. I tried to communicate in my own country, in a very disappointing way, the need to extend, to transform the teaching of philosophy. So, that is my personal disappointment. As to South Africa, I would say that philosophy, provided that it is taught in a certain way, is the most urgent thing today in this country. It should not be simply a secluded, elitist research – it has to be that too, I am in favour of a professional discipline and a specialised and hard way of doing philosophy. What we need in South Africa is a mediation between this sophisticated research and popular teaching, teaching of philosophy in the high schools, which would be connected to the necessary training for citizenship. To know what the constitution is, and what it should be, is philosophy, an act of philosophising. You cannot understand what the constitution is without this act of philosophy. You cannot understand the principle of the constitution without being a philosopher, without philosophy. From that point of view, I think that in a turbulent country, full of change, of transformation of the political and legal structures, philosophy is needed – more than anything else. I think that the politicians who are in charge should understand that it is in their own interest that philosophy is taught as widely as possible, of course in the universities, but also in the high schools. It is easy to say, it is difficult to do. I know this from my own country. I can imagine how difficult it might be, but it should be a task, it should be defined as a priority. Either in the form of philosophy teaching properly, or through other disciplines, that some philosophy be introduced in other disciplines. That’s what we recommended in France, not as interdisciplinary, because I do not believe in interdisciplinarity as such, but in other ways of grafting disciplines, to produce a new form of teaching in which philosophy would be present in history, in literature, in law for certain, in medicine ... Paul Cilliers: In science? Jacques Derrida: ... in anthropology, in social sciences and in, of course, the hard sciences. This is very difficult of course, not an easy programme as such, but one I would favour.

180 

 Paul Cilliers, Willie van der Merwe & Johan Degenaar

Paul Cilliers: Thank you so much for your time. Jacques Derrida: Thank you.

Paul Cilliers

Complexity, ethics and justice Although it is not possible to conceive of the Law without incorporating some notion of justice, justice itself is not a legal concept. Justice can be understood, if somewhat elliptically, as achieving and maintaining nonexploitative relationships among the members of a society without destroying the differences which constitute the society. Law is necessary to achieve this, but justice does not reside in the law itself, it manifests itself in the nature of the contingent relationship between people, and is therefore an ethical concept.1 The problematic relationship between justice (ethics) and the law needs further investigation. Furthermore, justice does not maintain itself once it has (hypothetically) been achieved. Society is always in flux, and the relationship between its members shift continuously. Justice must therefore be perceived as a process. This has important implications for our understanding of the laws that stand in the service of justice, especially in the case of something like a constitution that is by its very nature fairly fixed. From these properties it should be clear that the notion of justice is a problematic one. In this article the focus will be on the status of the rules and laws used to regulate and control society. This will be done by looking at society as a complex system and employing notions from complexity theory to say something about the adequacy of a set of rules, like the Law, to guarantee justice. Since important debates in jurisprudence make reference to postmodern theory (e.g. Douzinas et al. 1991; Cornell et al. 1992), the way in which the terms “modern” and “postmodern” will be used here must be clarified. Because the term “postmodern” is used for a variety of positions, many of which are controversial, it has become difficult to use it without triggering responses based on preconceptions as to what postmodernism entails. One could, however, give substance to many of the so-called “postmodern” claims by analysing them from the perspective of complexity theory, as will be done below.

1 In that sense the term “justice” should not be confused with related terms like fairness, rectitude or redress. Redress for past injustices does not follow with any logical necessity from the notion of justice. In many cases redress may indeed be necessary in order to achieve justice, but the motivation for this would depend on the specific circumstances, not an a priori necessity. It is not difficult to think of instances where redress can lead to new injustice. This is not an argument against restorative justice or redress as such, but an argument against equating redress with justice, as well as a way of emphasising the contingency of justice. Originally published in the Journal for Humanistics (Tijdschrift voor Humanistiek), 2004, 5(19): 19–26. © SWP Uitgeverij BV Amsterdam.

182 

 Paul Cilliers

One aim of this strategy is to dispel two general objections held against postmodern or related positions, the first that it is relative and the second that it is irrational. The rejection of a universal and a historic form of rationality would only imply relativism to someone who is committed to the modern dream to the extent that the absence of absolute points of reference would render the whole world meaningless. A complete relativist is in a way nothing but a disillusioned foundationalist. One could, however, deny the existence of absolute points of reference, without slipping into relativism. From the structuralist and poststructuralist perspective, meaning, whether conceived linguistically or socially, is generated through relationships of difference in a complex network of interaction (see Cilliers 1998: 37–47). Meaning conceived in this way is neither arbitrary nor per definition unstable. I take this to be one of the central arguments of Derrida’s position.2 In the second place, a postmodern position does not imply a total dismissal of rationality, it just points out the limitations of an abstract, instrumental kind of rationality that is not sensitive to the complexities of specific contingent situations. Toulmin (1990: 31, 32) refers to this kind of rationality as “Platonic” and argues for a more Aristotelian rationality instead – a rationality that incorporates a wider spectrum of human capabilities than pure logic. The argument can be made stronger: the complexity of the world we live in cannot be described fully in terms of a closed set of rules (see below). If the notion “rational” should refer only to a set of coherent, logical principles that are universally valid, the postmodern position would dismiss it. However, it could also refer to the strategies we use to cope with the messiness of the world as best we can, for example, in the effort to behave ethically. In short, the term “postmodern” can be used to refer to a style of thinking that denies the fact that we can provide single descriptions of the world. Such an approach does celebrate difference and diversity, but it is not irrational or relativistic. It acknowledges the complexity of the phenomena we deal with.

1 Modern and postmodern ethics The choice between a modern and a postmodern style of thinking dearly has implications for our understanding of ethics. Modern ethicists would argue for universal ethical principles, principles that would always apply to everybody. A good example of such an ethics  – at least in the way it is popularly under-

2 See Derrida, 1988: 136.



Complexity, ethics and justice 

 183

stood – is Kant’s categorical imperative. The failure of such principles to provide criteria for proper ethical behaviour in actual, practical circumstances is not the result of a lack of moral integrity on the part of modern thinkers, but of the limitations of modernist rationality. The framework within which they thought was not concerned with specific, contextualised instances, but with the general and the abstract. If one’s understanding of ethics involves the combatting of real, contingent injustices, one can argue that the modernist position is actually a way of circumventing ethical responsibility. The effects of modern ethics are unfortunately not confined to philosophical positions. The functioning of the state and the legal system are governed by similar values. From a modernist perspective the law is conceptualised as a universal set of principles or rules – at least within the boundaries of a state. The citizen has no choke in the matter. The proper functioning of the state and of civil society depended upon a universal acceptance of the rule of law. One can ask, however, whether the following of a strict set of rules, irrespective of the institutions which impose them, involves any ethical behaviour whatsoever. “When following a rule, the moral agent merely has to identify the relevant rule and from that calculate the appropriate response. If no moral decision needs to be made, it is clear that the person itself cannot be held responsible for the consequences of her behaviour. The fault must then lie with the rule itself, or with the institution which legitimises it – usually an anonymous structure. There are few better examples of the iniquities possible under such circumstances than what happened during apartheid in South Africa. By uncritically remaining within the confines of the law, justice was violated in a terrible way, particularly by those (judges and law-enforcers) normally depended on to protect justice. It is exactly against this blind following of a rule that postmodern ethics reacts. The argument is that there can be no question of ethics without involving moral responsibility. The moral agent has to accept the responsibility for his decisions, a responsibility that cannot be shifted onto those enforcing the law.3 Ethical behaviour involves an acknowledgement that, given the complexities of our social circumstances, generalisation is impossible, and moreover, that contradictory demands are made on us. We cannot avoid choice. In general, the post-

3 From a postmodern perspective, the way in which moral decisions affect the subject have radical implications. Ethical behaviour is not something performed by a subject amongst other activities, nor is it the result of rational thinking. The ethical dimension is prior to everything else since the subject is constituted by its ethical behaviour. One is not a thief because stealing is prohibited by a law, or when you are caught out. One becomes a thief through stealing, whether someone else knows or not. You are constituted as a thief by stealing just as you are constituted as a person of integrity by what you do, irrespective of the recognition you get.

184 

 Paul Cilliers

modern argument claims that there are no a priori or universal principles that can be used to ensure ethical behaviour under all circumstances. Each contingent situation demands its own consideration. There is, however, one important sense in which this argument remains problematic. We are uncomfortable with calling the blind following of rules ethical, but we should also be uncomfortable with a purely private morality. An ethical position entails more than only responding to a specific situation. We feel that there should be reasons for acting in a certain way, and furthermore, that those reasons should be of such a nature that they would convince others to act similarly under similar circumstances. But how does one do that without propagating a general position, exactly that which is argued against? The problem is the following: we cannot base justice on a set of rules, yet we cannot have a legal system without a relatively fixed and general set of rules. Do we have to accept this conclusion – that a universally just law is in principle not possible – and take the pragmatic option of saying that a law that is not always just is better than no law at all? I would say no. There is something profoundly unphilosophical, if not unethical, in the insistence that we should proceed only from practical constraints when determining our choices. Although we can never avoid reality, we have to resist what has been called “the tyranny of the real”. The philosophical position demands that we also reflect on the principle of the matter, and not only work within the constraints imposed on us by the practicalities of the situation. The problem of the “impossibility” of justice will be pursued further by turning to the notion of complexity.

2 Complexity and complex systems A central strategy in the modernist approach to science is the analytical method. Since it is too difficult to deal with complex systems as a whole, they are divided into separate, less complex components that can be analysed through traditional (mathematical) methods. This approach works well under many circumstances, but not when dealing with complex systems. The problem is that the process of analysis destroys the so-called “emergent properties” of the system. These are the properties that result from the interaction among the components, properties that cannot be reduced to some characteristic of the components themselves. (A good example of an emergent property is consciousness; something that arises from the interaction between a number of simple neurons.) Emergent properties – and it must be emphasised that there is nothing strange or mystical about them – are often exactly the characteristics of a complex system we want to study.



Complexity, ethics and justice 

 185

As a result, when we deal with complex systems like living organisms, natural language or social systems, we have to deal with the system as a whole. Moreover, since such systems interact with their environment, it is often extremely difficult to decide what should be included as part of the system and what not. During the past decade or so complex systems have received a lot of scientific attention, especially in terms of the following questions: What is the structure of a complex system? How does this structure develop? How does the system represent and process information? Can we model the behaviour of a complex system? The interest in complexity has been fuelled by the popularisation of chaos theory, but it should be stressed that chaos is not what is at stake at all. The subject here is systems with a complex structure, a structure that is maintained and developed through a process of self-organisation. These systems also have to survive under difficult circumstances. In order to do that, they have to be pretty robust. It is therefore incorrect to think of them in terms of either chaos or determinism. The nature of complex systems can briefly be described in the following terms:4 A complex system has a large amount of components that could by themselves be fairly simple. These components are richly interconnected so that they can interchange energy and/or information. It is important to note that these interconnections are nonlinear. The characteristics of the system (especially the emergent properties) are not primarily a result of the nature of the components, but of the pattern of interconnection. As examples of such systems, think of the brain or of an economic system. The brain consists of a large number of neurons connected through the synapses. Although the neurons are, at least in terms of their information processing capabilities, fairly simple, the capacities of the brain as a whole are striking. An economic system is constituted by a large number of individual economic agents, often clustered together in groups. The macro-features of the economy, like GDP or inflation, are not under the direct control of individuals, but emerge as the result of the patterns of interaction between individuals and groups of individuals. For our purposes, there are two particularly important aspects of complex systems to consider. Both arise out of the nonlinearity of the interaction. In the first place, nonlinearity means that the law of superposition is not valid. This means that a complex system cannot be replaced with an equivalent system that is simpler: complexity is incompressible. Such a state of affairs has serious implications for the way in which we understand and describe such systems. Since it is practically impossible to grasp the system as a whole at a specific moment, any description of the system will in principle be limited. As soon as an attempt

4 For a more detailed characterisation of complex systems, see Cilliers 1995, 1998.

186 

 Paul Cilliers

is made to say something specific, some other aspect will have to be ignored or suppressed. There is nothing “wrong” with this state of affairs, nor is it something we can avoid. It is impossible to provide a single, complete description of a complex system. In the second place, nonlinearity implies that small causes can have large effects, and vice versa. Not that this will always be the case, small causes can also have small effects, but the possibility remains that something that appears to be insignificant, can be amplified tremendously. Since the interactions are not only nonlinear, but also rich (i.e. each component interacts with many others), it becomes practically impossible to predict the behaviour of the system. What is denied is not that the eventual behaviour of the system is caused by previous conditions, but that the behaviour can be deduced from an examination of the causes alone. Why is this important for our discussion of law and justice? From an ethical perspective, there is no denial that law and justice pertains to the social sphere. Social systems, irrespective of whether we refer to localised societies or more global ones, are complex, and therefore have the general characteristics of complex systems (Cilliers 1995: 129, 130). We cannot talk about justice in the social sphere without realising that we are dealing with complex phenomena. When confronted with a complex system it is not possible to give a single, complete description of it, and we cannot make deterministic predictions about its behaviour. This is equivalent to what postmodern theory claims, at least in the sense in which “postmodern” is used here. Hopefully the discussion so far has shown that we can argue for a position similar to that taken by some postmodern theorists without taking in a postmodern position as such, at least not a relativistic one. In terms of the problem under investigation, the most important point to be made at this stage is that although we cannot give a complete, rule-based description of a complex system, we do not conclude that the behaviour of the system is random or chaotic, or that we can say nothing about the system at all. Furthermore, the system itself is oblivious of our inability to describe it analytically. The problem lies with our description of the system. Our descriptions are only intelligible when it is done in general, coherent, terms, when the complexity is reduced. We try to explain or understand the system in terms of rules, but the system itself does not depend on those explanations, just as a stone does not solve differential equations when it falls to the ground. These considerations have a number of implications for our discussion of the possibility of justice.



Complexity, ethics and justice 

 187

3 Law, justice and complexity We can not formulate the conclusions of the previous section in ethical terms. The argument from complexity claims that a single story, or in the words of Lyotard, a “coherent meta-narrative”, cannot describe any social system fully. Any attempt to give a complete description will distort, disregard or violate some aspect of the system. In the context of society, this means that any attempt to describe or regulate the social system will result in the violation of something or someone that is not (or cannot be) considered in terms of that description. The reason why a certain description is acceptable has less to do with rationality and more with power. We do not have to look hard to find examples of master narratives which oppressed the “other” in the system, whether they be of a different race, religion, gender or sexual orientation. The implications of my argument are, however, somewhat stronger than a mere dismissal of discredited systems of domination. It claims that efforts to include or protect the “other” in a general description of society will not result in justice. Such a description will in its turn violate some other part of the system. Think of the complexities involved in affirmative action, or of the dominating effects of political correctness, an attitude that attempts to fix the relationships in a system in a predetermined way. It is impossible to arrive at a complete and just description of society, not because we lack the intellectual resources, but because the demands made on such a description are contradictary. To provide justice to someone will mean that somebody else is treated unjustly. One cannot begin to think about the problem of justice if one does not accept its impossibility. The realisation of this aporetic position is the first step in accepting one’s ethical responsibility. To put it in a nutshell: to disregard the complexity of the social system is not merely a technical or descriptive error, it is unethical. At this point it may seem as if the argument is that we should dispense with the notion of justice. That is not the case. To merely dismiss the notion would be a postmodern move in the “bad” or relativistic sense of the word. The problem is, however, that if we wish to maintain the notion whilst acknowledging that it is impossible to guarantee it in real situations, we seem forced to understand justice as some kind of transcendental concept. For justice to prevail, it seems as if it has to be guaranteed by something outside the social system, something or somebody with complete knowledge of and control over the system. Have we now come full circle? If we want to retain the notion of justice, are we forced to return to a transcendental or modernist position? Unfortunately this would not help us. A transcendental notion of justice is useless in practice since we do not know what it is. As finite, contingent human beings we do not have access to transcendental knowledge. In empirical terms a transcendental notion

188 

 Paul Cilliers

of justice is as empty as no notion of justice. (Once again we see that the fundamentalist and the relativist positions are closely related.) Exactly the same argument goes for the status of the law. It is in principle impossible to find any empirical rule, no matter how just it seems when conceived, that will not under special circumstances violate justice for somebody. Transcendental rules (like Kant’s categorical imperative), on the other hand, have no empirical content, and therefore cannot help us when we have to make ethical decisions under the messy conditions of real life. Does this lead to the acceptance of a lawless society where we have to trust each individual’s sense of responsibility to do the right thing? Although agree with Bauman (1993: 10) that humans are not intrinsically bad, such a trust is totally misplaced. They are not intrinsically good either. The absence of the law will allow the selfish and powerful to exploit and dominate – an unjust state of affairs. At this point the lawmaker may urge that we have no choice but to make the pragmatic move and argue that although we know that the law cannot be just under all circumstances, we also know that things would be even worse without the law. We therefore have to compromise on a certain amount of justice, she argues, in order to put a practical legal system in place. It is not perfect, but like democracy, it is the best we’ve got. I take this to be the route followed by the founding fathers of most Western legal systems. This response may be practical, but from a philosophical, or rather, ethical perspective it is not adequate. To engage with the problem of justice in a philosophical way does not entail a choice between either a transcendental position or a pragmatic one. Rather, it entails entering into this dilemma and, in a way, to accept both sides of it. A possible way to do this is, which could incorporate the perspective developed from complexity, is suggested by Drucilla Cornell (1992) in her development of Derrida’s position. ln The Force of Law: The ‘Mystical Foundation of Authority’ Derrida (1992) describes the fundamental aporia of justice in the following way: How are we to reconcile the act of justice that must always concern singularity, individuals, irreplaceable groups and lives, the Other or myself as Other, in a unique situation with rule, norm, value or the imperative of justice which necessarily have a general form, even if this generality prescribes a singular application in each case? If I were content to apply a just rule, without a spirit of justice and without in some way inventing the rule and the example for each case, I might be protected by the law [droit], my action corresponding to objective law, but it would not be just (Derrida 1992: 17).

The position presented in this paper acknowledges that we cannot give complete descriptions of complex situations in terms of a finite set of rules. At the same time, it insists that we cannot cope without rules. We need them in order to make sense of the world and to operate in it. This may sound like contradictory



Complexity, ethics and justice 

 189

demands, but that would only be the case if the rules have to be universal and timeless, and that is not how they are conceived of. They should be drawn up as if they were universally valid, but with the proviso that they have to be re-evaluated each time they are applied. In other words, we do the best we can each time we have to make an ethical choice. We gather all possible information and consider all possible options, then make a decision as if we would expect it to be a universally valid decision while we realise that we could not consider all possible options, and that we have to be prepared to reconsider the choice. We cannot escape making a decision, and we can also not escape accepting responsibility for its outcome, even if the outcome was something that could not be foreseen. This as if allows us to retain same notion of justice that is both ideal and practical. We employ the notion as if it is a transcendental one, but do not allow the transcendental emptiness to prevent us from acting and accepting the responsibility for the actions ourselves. (This understanding of justice is what Derrida and Cornell refer to as quasi-transcendental.) One can reformulate this position in complexity terms: One cannot know a system in all its complexity. This does not mean, however that we can say nothing about the system. It means that we have to be careful with the scope of the claims we make about the system. Such claims are always contextualised, but they need not therefore be either random or relative, even if they have to be limited and provisional. We can make proper use of such claims or descriptions, until such time as the shortcomings (which will be there despite our best efforts) become apparent. We will be more attentive to such shortcomings if we accept that they will be there. It is also less disabling to work with a provisional understanding of the system than to depend on some kind of characterisation that should be perfect. Such a position can inform our conceptualisation of the legal system. It means that while we cannot escape the responsibility of drawing up the law, the justness of the law has to be re-established each time an appeal to that law is made in a specific context. In practice this means that although there is a law, it should be seen as a provisional tool, something that is revisable. It also means that those who have to judge cannot shift the responsibility for their judgement onto the law. When we make decisions we have to reduce the complexity of the world, but we cannot escape the way in which these decisions will work out in the complex reality. The responsibility for the judgement, and for the consequences of the judgement – which are never fully predictable – will remain with those who judged.

190 

 Paul Cilliers

References Bauman, Z. 1992. Intimations of Postmodernity. London: Routledge. Bauman, Z. 1993. Postmodern Ethics. Oxford: Blackwell. Cilliers, P. 1995. Postmodern Knowledge and Complexity (or why anything does not go). In: South African Journal of Philosophy 14 (3). Cilliers, P. 1998. Complexity and Postmodernism. Understanding complex systems. London: Routledge. Cornell, D. 1992. The Philosophy of the Limit. London: Routledge. Cornell, D., Rosenfeld, M. & Carlson, O.G. (eds.). 1992. Deconstruction and the possibility of justice. London: Routledge. Derrida, J. 1988. Limited Inc. Evanston: Northwestern University Press. Derrida, J. 1992. The Force of Law: “The Mystical Foundation of Authority”. In: Cornell, D., Rosenfeld, M. & Carlson, O.G. (eds.). Deconstruction and the possibility of Justice, London: Routledge. Douzinas, C., Warrington, R. & McVeigh, S. 1991. Postmodern Jurisprudence, The law of texts and the texts of law. London: Routledge. Lyotard, J.P. 1984. The Postmodern Condition: A Report on Knowledge. Manchester: Manchester University Press. Toulmin, S. 1990. Cosmopolis: The Hidden Agenda of Modernity. New York: Free Press.

Part 1: Single-authored Papers Theme 3: Implications of Complexity Thinking

Paul Cilliers

Difference, identity and complexity If the world we lived in, or more specifically, if the organisations we work in and with, were mostly symmetrical and homogenous, there would be a number of advantages. They would be stable and their behaviour would be predictable. It would also be possible to model them accurately, and thus to understand them fundamentally. “Knowing” them would lead to the possibility of controlling them. The problem is, such a world or such organisations could only be very uninteresting. Living things and complex social systems are by their nature heterogeneous and asymmetrical. Complex systems are made up of a multitude of nonlinear interactions that cannot be simplified.1 They are unpredictable and full of surprises. There are serious difficulties involved in understanding, let alone modelling, them. But perhaps the complex behaviour of such systems is only epiphenomenal. Perhaps, underneath the multifaceted surface, there are general principles to which the seemingly contingent behaviour could be reduced. This would allow us to model the essential behaviour of these systems, and not be distracted by the contingencies. Finding these internal regularities was the hope of what could generically be called Modernism.2 This strategy was governed by the ideal to find universal, ahistorical, and non-contingent principles that would describe complex systems accurately and thus allow for prediction and control. If such an ideal was the guiding principle, diversity would be a problem. It would complicate our understanding and interfere with our planning. It would confront us with the surface of things, not with their essence. It would force us to deal with a countless number of factors, too many to handle. I shall argue, however, that such an understanding of diversity is not only misguided, but dangerous. Diversity is not a problem to be solved; it is the precondition for the existence of any interesting behaviour. The notion of “diversity” is used here in the context of post-structural theories of meaning and of the characteristics of complex systems. These contexts will be unpacked in more detail later, but my general argument is that in a post-structural understanding of language, meaning results from the differences between

1 See my Complexity and Postmodernism: Understanding Complex Systems (London: Routledge 1998), 2–7. 2 See Zigmund Bauman, Intimations of Postmodernity (London: Routledge, 1992). Originally published in Philosophy Today, 2010 (Spring), 54(1): 55–65. © DePaul University.

194 

 Paul Cilliers

all the signs in the system. Sameness does not generate meaning. The richness of the system is a function of the differences it contains. Similarly, complex systems are made up of the nonlinear interactions among large amounts of elements which are not necessarily complex in themselves. These interactions produce the “emergent” properties of the system, the higher order properties that make the system what it is. A good example is the way in which consciousness emerges from the interaction between neurons in the brain. For this to take place, there must be a large number of neurons which are nonlinearly and asymmetrically connected. A small amount of homogenous neurons will just not do it.3 This “necessity” of diversity can also be explained by looking at an organisation. To be able to fulfil its role and to cope with a challenging and changing environment, an organisation needs diverse resources. The functions of the different components of the organisation are not simply interchangeable. The crane operator cannot do the job of the financial manager and vice versa. The more complex the role of the organisation is, the more diversity is required to perform it. The problem to be addressed should now begin to emerge. For an organisation to have vital and dynamic properties, it needs a lot of diversity. If, however, we want to describe, understand, control, or manage such an organisation, the diversity becomes a problem. We cannot reduce rich, nonlinear difference to simple descriptions, but we need descriptions nonetheless. It was the hope of Modernism that such simplified descriptions – descriptions which are accurate and contain the essence of the matter – could be found. The post-structural argument and the argument from a critical understanding of complexity is that such reductive strategies are seriously flawed.4 What then can we say about difference and diversity? Are we reduced to throwing up our hands and saying “things are very complex”, or are we doomed

3 Some knowledge of complexity thinking and of the earlier work of Derrida is required for the argument in this essay. Brief expositions will be provided, but for a detailed discussion of the characteristics of complex systems, see my Complexity and Postmodernism. That text develops the similarities between a critical theory of complexity and deconstruction exhaustively. 4 It should be made explicit that not all forms of complexity theory share this critical sensitivity. Edgar Morin, in “Restricted Complexity, General Complexity,” in C. Gershenson, D. Aerts, and B. Edmonds, eds., Worldviews, Science, and Us: Philosophy and Complexity (Singapore: World Scientific, 2007), 5–29, distinguishes between what he calls “restricted” and “general” theories of complexity. Restricted complexity acknowledges the relational nature of complex systems, but hopes that essential characteristics of these systems can be positively identified. This return to reduction is often encountered in forms of complexity theory which developed out of chaos theory. In particular, this includes many of the traditional positions on complexity associated with the Santa Fe Institute.



Difference, identity and complexity 

 195

to use flawed reductionist descriptions? This essay attempts to move the argument beyond this dichotomy through a philosophical analysis of the notions of “difference” and “diversity” – notions which are, in this context, used somewhat interchangeably. It will be argued that difference is not merely one of the characteristics of such systems, but a precondition for their existence. The relationships of differences constitute complex systems. These differences are not only the observable differences on the emergent level of the system, but also, and perhaps primarily, all the small differences which provide the means for emergence to take place, that which Derrida calls “traces”.5 It will be argued that the identity (or identities) of the system is a result of these differences and interconnectivities, not something which precedes them. Although the notions of difference and identity are intertwined in an inextricable way (as will be argued below), one could say, as a kind of non-foundational ontology, that there are really only differences. Such an analysis will show that we can give much more content to the problem of diversity, that there is more to say about it than to simply acknowledge it. There is an important reason why one should not commence the investigation of difference on the emergent level of higher order phenomena in social systems. The problem is that the differences on this level are already the result of smaller differences. Focusing on the large-scale differences, like, for example, differences in race or gender, tends to underestimate the extent to which these are already divided categories. A superficial understanding of difference can thus lead to an eradication of differences within a certain group.6 An overemphasis on difference and otherness, on the social level, may paradoxically result in a ploy to protect us from the different by generating a discourse which emphasises an incommensurability between heterogeneous groups. To understand the “logic” of difference, we must first look at difference as a necessary condition for meaning at a “low” level, i.e., look at how the conditions for meaning and emergent characteristics are constituted. Although this analysis is a philosophical one, one which engages with the “logic” of the notion of difference on a general and abstract level, the issues discussed here (the relational nature of difference, the necessity for the play of difference to be bounded, and the relationships between difference and identity) have

5 “Nothing, either in the elements or in the system, is anywhere simply present or absent. There are only, everywhere differences and traces of traces.” Jacques Derrida, Positions, trans. Alan Bass (Chicago: University of Chicago Press 1981), 26; see my Complexity and Postmodernism, 41–45 for a detailed discussion. 6 See Christine Sypnowich, “Some disquiet about ‘Difference,’” Praxis International 13 (July 1993): 99–112.

196 

 Paul Cilliers

important implications for large-scale systems like social systems and organisations. These implications will be examined in the concluding parts of this essay.

1 The “logic” of difference In order to understand the general significance and implications of the notion “difference” in the context of complex systems, we need to develop a more nuanced “theory of difference”. This will be developed in three steps.

1.1 The necessity of difference In the first place, the argument that difference is essential has to be substantiated. To merely insist on difference, as if it is necessary in some metaphysical way, is not sufficient.7 Such an argument can be built around the claim that difference is a necessary condition for meaning. For something to be recognisable as being that something, it must be possible to differentiate it from something else. Sameness (not to be confused with the notion of “identity” as it is used here) refers to an absence of difference. The more differences there are, the more distinctions can be made. Meaning is the result of these distinctions, of the play of differences. In a philosophical context, this argument is best made using Saussure’s theory of language as a system of differences. Meaning, for Saussure, is not the result of an essential characteristic of a sign, i.e., some a priori identity, but of the relationships between all the signs in the system. To explain the way in which these relationships work, Saussure uses the example of a train, say the “8:25 Geneva-to-Paris”.8 Although the train itself, its personnel, and its passengers are different every day, the “8:25 Geneva-to-Paris” maintains its identity by its relationships to the “8:40 Geneva-to-Dijon,” the “12:00 Geneva-to-Paris”, or the “0:38 Bombay-to-Madras” for that matter, irrespective of whether it leaves at 8:25 exactly, or reaches Paris in the same state as when it left. The train does not have any identity by itself, its identity is determined relationally. Similarly, a linguistic sign derives its meaning from its relationships to other signs. The signifier “brown” does not have a meaning because it can be identified with a concept that unambiguously contains the essence of “brownness”, but because it can be

7 Ferdinand de Saussure, Course in General Linguistics (London: Fontana, 1974). 8 Ibid., 108.



Difference, identity and complexity 

 197

differentiated from the signifiers “black”, “blue”, “grey”, “hard”, “train”, etc. The sign is determined by the way in which it differs from all the other signs in the system  – “in language there are only differences without positive terms”.9 The sign is a node in a network of relationships. The relationships are not determined by the sign, rather, the sign is the result of interacting relationships. Similarly, Freud, in his early neurological model of the brain, also described neural interaction as a system of differences.10 Freud’s model consists of neurons that interact through pathways which channel the energy in the brain. This energy comes from both outside the body (perception), and from internal sources. Pathways resist the flow of energy, unless it is used often. The characteristics of the brain are determined by the various patterns of energy flowing through it. Two important aspects of this model deserve attention. In the first place the role of memory should be underscored. “Memory” refers here to the physical condition of the brain: which pathways are breached (“facilitated”) and which are not. Memory is not a cognitive function performed by a conscious subject, but an unconscious characteristic of the brain (which is an organ, part of the body). Memory is the substrate that sets up the conditions for all the functions of the brain. The second important characteristic of Freud’s model concerns the role of the neurons. No neuron is significant by itself. Memory does not reside in any neuron, but in the relationship between neurons. This relationship, Freud declares, is one of differences.11 What we have, therefore, is a model structurally equivalent to Saussure’s model of language: a system of differences. Such a system of differences can also be used to describe how a complex system works.12 Such systems consist of a number of components which interact nonlinearly. The complexity of the system does not reside in the components, but is a result of these interactions. If these interactions were ordered, homogenous and symmetrical, no interesting behaviour would arise. There has to be asymmetry. This is another way of stating that the relationships between the components are relationships of difference. Space does not allow for a complete development of a theory of complex systems from a post-structural perspective, but for one important remark. If one sticks to a purely structuralist (i.e. Saussurian) understanding of complex systems, one ends up with a model which argues

9 Ibid., 120. 10 Sigmund Freud, Project for a Scientific Psychology, in Standard Edition, vol. 1 (London: Hogarth Press, 1950), 281–397. 11 Ibid., 300. 12 Complexity and Postmodernism, 1–7.

198 

 Paul Cilliers

that things may be relational and very complicated, but if you work hard enough, with clever enough techniques, you can figure the system out – essentially the general structuralist claim. This understanding would correspond to what Morin calls “restricted complexity”. A “general” understanding of complexity requires a more reflexive and transformative approach. It is exactly in this respect that deconstruction makes a vital contribution. It allows us to describe the dynamic nature of the play of differences. Derrida’s deconstruction of the structuralism of Saussure centres around the concepts of trace and différance. The concept “trace” can be used to refer to the individual differences between the components in a system. Each trace has no meaning in itself, but through their interaction the meaning of a sign emerges. The notion of différance can be used to describe the dynamics of complex networks. The analogy works in the following way: the interaction between a number of components in the system generates a pattern of activity, traces of which reverberate through the whole network. Since there are loops in the network, these traces are reflected back after a certain propagation delay (deferral), and alter (make different) the activity that has produced them in the first place. Since complex systems always contain loops and feedback, delayed self-altering will be one of the network’s characteristics; a characteristic described quite precisely by Derrida’s notion of différance – a concept that indicates difference and deference, that is suspended between the passive and active modes, and that has both spatial and temporal components.13 Difference is therefore not simply the static differences between components in the system; the components themselves are constantly transformed. This basic, dynamic model can also be used to generate an understanding of how an individual comes to be (develops its identity) in a network of relationships with other individuals, or how an organisation comes to be as a result of the relationships between its internal components as well as the relationships with other organisations from which it differs.14 The identity, or “meaning”, of an organisation is not pre-given or complete. It develops and transforms as a result of the play of differences that constitutes it. We will return to this issue later. The argument thus far can be summarised in what one could call the law of meaning: without difference there can be no meaning. If we accept this, it will follow that if we want a rich understanding of the world and of each other (i.e. a

13 Jacques Derrida, “Différance,” in Margins of Philosophy, trans. Alan Bass (Sussex: Harvester Press, 1982), 1–27. 14 These arguments are detailed in my Complexity and Postmodernism. See especially chapter 7.



Difference, identity and complexity 

 199

lot of meaning), if we want resilient and dynamic organisations, then we need an abundance of differences. The point to be emphasised is that an abundance of difference is not a convenience, it is a necessity. Complex systems cannot be what they are without it, and we cannot understand them without the making of profuse distinctions. Since the interactions in such systems are nonlinear, their complexity cannot be reduced. The removal of relationships, i.e. the reduction of difference in the system, will distort our understanding of such systems. A failure to acknowledge this leads to error, an error which is not only technical, but also ethical. When we pretend that we can understand or model a complex system in its full complexity, such pretence is not only hubristic, it is also a violation of that which is being modelled, especially when we are dealing with human or social systems. Trying to understand complex systems involves a certain modesty.15 However, if we merely insist on an abundance of difference which is irreducible, we are not saying enough about how complexity is constituted. A limitless play of difference does not, as some postmodernists seem to argue, lead to the generation of meaning, nor can a complex system function without being constrained in some way. On the epistemological level (our descriptions of complex systems) as well as the ontological level (the functioning of complex systems in the real world), boundaries are required. This boundedness can be examined from two perspectives. The first will be referred to as the economy of difference. The second is concerned with the inescapable presence of some kind of identity. These two perspectives will now be examined further.

1.2 The economy of difference Since the interactions in a complex system involve all the components, and since complex systems are open systems, the play of difference is potentially infinite. If this was actually the case, no meaning could emerge, since the deferral would be absolute. Real systems, however, are bounded. There has to be a boundary in order to be able to identify a system as this system and not another. The boundary

15 A complex system is constituted through the relationships of differences. These relationships are nonlinear. If the complexity is reduced, i.e. some of the difference is removed, it distorts our understanding of the system. Nevertheless, we have to reduce the complexity in order to be able to say something about the system at all. Because of the nonlinearity, the magnitude of the resulting distortion cannot be predicted. Since we know this beforehand, we have to accept responsibility for these distortions. See my essay “Complexity, Deconstruction, and Relativism”, Theory, Culture, and Society 22:5 (2005): 255–267, for a detailed discussion of this point.

200 

 Paul Cilliers

is thus constitutive of the system, it enables the system to be. It does not simply close the system off, it facilitates interactions between the system and its environment.16 However, if the characteristics of a system are a result of the interplay of differences, and these relationships cannot continue to reverberate in an unconstrained way. At some stage they have to be reflected back upon themselves in order to consolidate into a pattern which constitutes some aspect of the system.17 The constraints introduced by the boundary lead to what one could call an “economy” of difference. This notion needs to be unpacked in the context of complex systems. A system does not have the capacity to be complex just because it is multidimensional or has many degrees of freedom. Complexity does not arise as a result of a chaotic free-play with infinite possibilities. Complex systems have structure. It is the structure of a complex system which enables it to behave in complex ways. If there is too little structure, i.e. many degrees of freedom, the system can behave more freely, but this freedom leads to activities which are meaningless, random, or chaotic. The mere “capacity” of the system (i.e. the total amount of degrees of freedom available if the system was not restricted in any way) does not serve as a meaningful indicator of the complexity of the system. Complex behaviour is only possible when the behaviour of the system is constrained. On the other hand, a fully constrained system has no capacity for complex behaviour, either. This claim is not quite the same as saying that complexity exists somewhere on the edge between order and chaos.18 A wide range of structured systems display complex behaviour. Complexity is not simply a function of plenitude, but of interchange and relationships. In order to say more about the nature of these relationships of difference and how they constrain and are constrained in a certain economy, it is necessary to

16 I discuss the nature of boundaries and the way in which they are enabling in “Boundaries, Hierarchies and Networks in Complex Systems,” International Journal of Innovation Management 5 (June 2001): 135–147. 17 A related argument is provided by Anthony Wilden when he distinguishes, in a fundamental way, between the digital and the analogue. For a collection of “differences” to become a “distinction”, i.e. a carrier of meaning, it must become a “discrete element with well-defined boundaries”. System and Structure: Essays in Communication and Exchange, 2nd ed. (London: Tavistock, 1984), 169. 18 This point can also be elaborated from the perspective of self-organised criticality. This perspective helps to resist a too close association between chaos theory and structured complexity. A nonlinear interaction between a few components can produce chaos, but “chaos theory cannot explain complexity”. P. Bak, How Nature Works (New York: Springer, 1996), 31. A complex network of interactions will constrain chaotic behaviour.



Difference, identity and complexity 

 201

realise that “difference” does not mean “opposition”. To say that A differs from B is not to say that B is not-A. There may be a lot of similarities between A and B, they may differ only in some small aspect. As a matter of fact, there has to be at least some common element between them (this point will be returned to below). Furthermore, is it possible to even talk about the difference between only two things? As long as we deal with just the two things, the difference between them is totally unconstrained. A differs from B in everything that B is not, and vice versa. The difference between them is boundless, or to put it differently, it would not be possible to give content to the difference between A and B, if the two are the only entities taken into consideration. Difference is not a function of a binary opposition, but of a network of relationships framed in a certain way. A collection of differences is required to narrow down what is completely open to something that has an identifiable meaning. This is vital for the way in which the components of the system acquire meaning.19 Of the many relationships of difference associated with a component of the system, think of a specific one.20 This relationship does not determine the meaning, or part of the meaning, in any way. Since it is a relationship of difference, it can only minutely indicate part of what the meaning is not, and thereby place a little constraint on the meaning of the relevant component. The meaning of a component at a specific point in the history of the system is, therefore, that which satisfies all the current constraints placed on it through all its relationships in the current context, i.e. as determined by the current boundary.21 It should be clear now that difference does not generate meaning in an unlimited way. Meaning is only possible when there are many differences interacting by constraining each other. Put differently, meaning is only possible if difference is confined. Again, this does not mean that we can now pin down the meaning of a component in the system. If there are only a few relationships associated with a certain component, the associated meaning will have many degrees of freedom. If there are more relationships involved, the meaning is more richly constrained. In other words, if the set of relationships of difference associated with a certain component is underdetermined, the meaning of that component will be fairly

19 A has meaning because of its relationship with B and C and D and F ... Nevertheless, this list cannot be infinite. 20 Such a “one”, specific relationship of the many relationships associated with a component is what I understand under Derrida’s notion of the “trace”. It is, of course, not possible to give conceptual content to a trace, despite the fact that there “is” nothing but traces. 21 It should be kept in mind that the constrained system of differences does not generate meaning in a static way, but that it is a dynamic process which could be described through the notion of différance.

202 

 Paul Cilliers

arbitrary; it will be more open but somewhat lean. If the set of relationships is complex, the meaning of the component will be much more unique, but, and this is the crux of the matter, it will simultaneously be more rich and varied. The fewer constraints, the more possibility, but possibility left empty. The more constraints, the better we can get at the meaning, but the more bountiful it is. To take a social example: the life of a hermit can be fairly unconstrained, but it is difficult to give much social significance to her existence. It will be much easier to say something about the significance of somebody with a rich set of social interactions, interactions that will at the same time constrain that person’s life. Possibility can only be actualised in the presence of constraints. This “economy” of difference should be understood in the Hegelian sense. There is constant interaction between a bounded number of components. It is useful, though, to distinguish between a “restricted” and “general” economy,22 just as we can distinguish between a restricted and a general complexity. The fact that the system is bounded does not mean that the components involved remain the same – in the sense that the components do not change and in the sense that the same set of components remain involved. This would be a restricted economy. From the perspective of a general economy one would underscore the dynamic nature of the system, which includes the fact that elements can and will change and that the boundary can shift. The boundary will also never be complete or exact. It will contain folds and gaps, elements which enable the transformation (or deconstruction) of the system. The notions of constraints and boundaries remain indispensable nonetheless, even if these notions include the working of différance. The necessity of constraining structures is highlighted from the perspective of deconstruction in a different way, as well. For a text to have meaning at all, it must be deconstructable.23 The constraining hierarchies in a text are necessary, even if the meaning which arises is not final. It is the occasional dream of deconstruction, Derrida claims in the Afterword to Limited Inc. to make an attempt to incorporate the absolutely complete context, and thus arrive at an exact meaning.24 However, this is not possible. Only a limited context allows for meaning,

22 See Jacques Derrida, “From Restricted to General Economy: A Hegelianism without Reserve”, in Writing and Difference, trans. Alan Bass (London and New York: Routledge), 317–350. 23 See, e.g. John Caputo, Deconstruction in a Nutshell: A Conversation with Jacques Derrida (Bronx, NY: Fordham University Press, 1996). 24 Jacques Derrida, “Afterword: Toward an Ethic of Discussion,” in Limited Inc. (Evanston: Northwestern University Press, 1988), 136.



Difference, identity and complexity 

 203

and because of these limits, the meaning can be deconstructed. “If things were simple” he says, “word would have gotten around.”25 My argument thus far can be summarised by saying that although difference should be proliferated, it cannot be done in an unbounded fashion. There is a certain economy involved in the process whereby differences generate meaning in a complex system. This economy imposes limits on difference. We can, therefore, reformulate the “law” of meaning stated above: without constrained difference there can be no meaning. This now leads to a discussion of the relationship between the different and the same.

1.3 Difference and identity26 “Were one to write a general philosophical history of the concept of difference, one might be tempted to view it as the progressive emancipation of difference from identity”, writes Rudolph Gasché.27 He continues: “If at the dawn of philosophical thinking difference scarcely left the shadow of identity, identity now barely shows its face.” Some of the reasons for this are not difficult to understand. The postmodern flight from universal principles and unifying meta-narratives resulted in a strong emphasis being placed on the notions of difference and the other. To a large extent this criticism of modernity and the enlightenment is correct, but to think that one could talk about difference without involving the singular or the same is equally problematic. The mistaken opposition between the notions of difference and identity is a further result of confusing the notion of “difference” with that of “opposition”,

25 Ibid., 119. 26 The word “identity” has a number of meanings, often shading into each other. It can refer to something singular (oneness) or to things which cannot be distinguished and thus are “identical”. The notion of “personal identity” has to do with what makes a person identifiable as that person, and not another, with what it is which “makes up” a person (or an institution). In the critical theory of the Frankfurt School, identity thinking refers to the mistake which “aims at the subsumption of all particular objects under general definitions and/or a unitary system of concepts”. D. Held, Introduction t o Critical theory: Horkheimer to Habermas (Berkeley: University of California Press 1980), 202. Particular identities are sacrificed in favour of a universal identity. “Identity thinking” is therefore another example of a modernist resistance to difference. In this essay I use the term to indicate, on the one hand, the complexly interwoven relationship between the different and the same and, on the other, the construction of (personal) identity through relationships of difference. 27 Rudolph Gasché, Inventions of Difference (Cambridge, MA: Harvard University Press, 1994), 82.

204 

 Paul Cilliers

i.e. to think that to say A is different from B, is the same as saying A is not B (see above). In order to recognise a difference between A and B, they must in the first place be identifiable as A and as B (in their singularity), and secondly, they must, even if only slightly, share something that makes a comparison possible (there must be some element of identity). Moreover, as has already been argued, it is not really possible to talk of the difference between A and B if A and B are the only two things under consideration. The difference between apples and pears can only be understood in terms of what they share, e.g. that both are fruit. One can talk of the difference between apples and motorbikes, but this difference is so open that it will only have meaning in terms of a number of other factors that form part of the context of the comparison – perhaps that apples cannot be used for transport or that motorbikes are not nutritious. To attempt to relate two things that are radically or absolutely “other” is something that cannot be done; the comparison will be totally meaningless in the full sense of the word. If we encounter something totally alien we will not be able to recognise it. Gasché formulates this point in the following way: any encounter worth the name presupposes not only encountering the Other in all his or her singularity, but recognizing this singularity in the first place. Paradoxically, even the most radical singularity must, in order for it to be recognized for what it is, have an addressable identity, guaranteed by a set of universal rules that, by the same token, inscribes its singularity within a communal history or tradition, and problematics.28

Let us consider briefly what the implications of this understanding of difference are for our relationship with the other in the social sphere.29 The realisation that differences are constitutive leads to the recognition of the importance of difference. If, as a result of this insight, the notion of difference is absolutised, it may lead one to think that no relationship between the self and the other is possible; that the other is absolutely other. However, in order to be able to recognise the other as other at all, some form of identity between the self and the other is required. As a matter of fact, the claim that the other is completely unknowable is nothing but an inverted insistence on pure identity  – in the sense that the other has an identity which is not breached by any difference – in the same way that relativism is an inverted form of foundationalism.30 Does the insistence

28 Ibid., 2. 29 For a different, more political discussion of this issue in the context of Eastern Europe, see Martin Matuštík, “Derrida and Habermas on the Aporia of the Politics of Identity and Difference: Towards Radical Democratic Multiculturalism,” Constellations 1 (1995): 383–398. 30 This insight can be used to criticise Levinas’ understanding of the Other as something abso-



Difference, identity and complexity 

 205

that the other must always already share something with the self before it can be recognised as other imply that the other can be fully appropriated? Not at all. There is an irreducible difference between the self and the other that will always complicate the relationship. But we are not lost in space. The moment we can recognise the other as other, there must already be a minimal form of identity (some small similarity) to make the recognition possible. The relationship will remain complex, and merely acknowledging this does not guarantee that the other will not be violated. It merely provides a point of departure, from where a relationship, even if it is a tenuous one, with the other can be attempted. Gasché again: if the singularity of the Other requires a minimal universality to be itself and to be recognized as such, then the Other’s point of view, or private fantasies, become repeatable, risk being lost by becoming entirely mine. Yet without this risk no justice can possibly be done to the singular; without it, the very possibility of something singular would remain irretrievably lost.31

To summarise this section: meaningful relationships in a complex system develop through relationships of difference, not through opposition. For meaning to become possible, some form of similarity must already be there. This does not imply that the meaning can be fixed or exhausted in any way. The element of identity inaugurates the play of difference on the one hand, while on the other, it is the result of that very process. We cannot think identity without also thinking différance. We can therefore add a further refinement to the “law” of meaning: without constrained difference and repeatable identity, there can be no meaning.32 If we now want to talk of identity in the sense of “personal identity”, or the identity of an institution or system, the same law holds.

2 Complex identity After discussing some of the general philosophical aspects of difference on a micro level, an attempt can be made to see how they translate to the macro level of persons and groups of persons. The argument is based on the assumption that the general characteristics of complex systems (that they are constituted through

lute, as opposed to Derrida’s understanding of the other as something more richly differentiated. See Drucilla Cornell, The Philosophy of the Limit (London: Routledge, 1992), 68–72. 31 Gasché, Inventions of Difference, 16. 32 This position can also be formulated in terms of Derrida’s notion of “iterability”. See his “Signature Event Context,” trans. Samuel Weber and Jeffrey Mehlman in Limited Inc., 1–24.

206 

 Paul Cilliers

nonlinear interaction, that they operate in a state far from equilibrium, that they have the capacity to self-organise, that they have emergent properties, etc.) are applicable to systems on different scales. Although these characteristics do not allow us to pin down the behaviour of any specific system at a specific time, they do help us to understand some of the dynamics of complex systems, as well as providing reasons for why it is so difficult to model them. The basic claim is the following: if, generally speaking, the meaning and function of a component in a complex system is the result of relationships of difference, this would also hold for social systems. In this context, then, the notion “meaning” can be used to indicate the identity of the system. Thus, the identity of a person or an institution is the result of constrained differences. Identity is an emergent property resulting from the diversity in the system, and not something which exists in an a priori fashion. It is therefore a mistake to think of difference as something that exists in the difference between already established identities. Identity and difference mutually imply each other in an open dialectic. Such a position allows us to say a few things about social identity. In the first place, such an identity could be constructed from relatively few components, but it will then be somewhat lean and shallow. The more diversity there is involved in the construction of the identity, the richer it will be. A “rich” identity does not imply that such an identity is open, general or vague. This is exactly the nature of a lean identity. A rich identity is also richly constrained. It is more specific, and at the same time more nuanced. Take the example of a self-reliant minority. Such groups may tend to derive their identity by recycling internal, well-established differences and by excluding outside influence. This may easily result in a “lean” and static identity. If, however, a minority finds its identity in a rich interaction with other groupings, such an identity will not only be richer and more specific, but it will also be more resilient. More specifically, if identity is the result of diversity, and if differences are constantly being moved around in feedback loops and imposed from outside the system as the context changes, then identity is by definition a dynamic concept. Identity can, of course, be quite stable, but if it gets locked in (by ignoring, for example, important changes in the environment, or by deciding not to interact with other systems), such a fixed identity will most definitely be detrimental to the system. The fixing of relationships within a system, and the closing down of its borders, will introduce a rigidity which leads to senescence or pathology. At the same time, this does not mean that the identity of a system should change indiscriminately. Even if identity is dynamic, there should be an appropriate tempo of change. It is not possible to provide general guidelines to what this tempo should be, since it will differ in different contexts. There is, however, a flag to be waved at this point: many analyses of complex dynamic systems in the social sphere



Difference, identity and complexity 

 207

tend to emphasise adaptability, and therefore argue for rapid change. It could be argued that, as a generalisation, this is wrong. In order to maintain any identity whatsoever, and for it not to merely reflect its environment, a system must change at a slower rate than its environment.33 It can do this, and still cope with a changing environment – or rather, cope with it better – only if it has an abundance of richly constrained diversity. This insight can be expanded by what Peter Allen calls the “law of excess diversity”.34 A system should not only have the “requisite variety” it needs to cope with its environment (Ashby’s law), it should have more variety. Excess diversity in the system allows the system to cope with novel features in the environment without losing its identity – as long as one remembers that identity is now a dynamic concept which is subject to change. What is more, if a system has more diversity than it needs in order to merely cope with its environment, it can experiment internally with alternative possibilities. The capability to experiment may just be another word for being creative. Thus viability, resilience, and even survival, are notions intimately linked with the idea of creativity.

3 Difference, organisation and organisations It is important to reiterate that the constrained play of differences is not simply one of the activities of the system, but that it constitutes the system. A meaningful way to look at the play of difference is to see it as the way in which the system is organised. The relationships constituting the system are not random or chaotic, they are structured. The complexity of the system is not simply a function of the interactions between many components, but of their organisation.35 The fact that some form of structure is necessary does not imply that the organisation of a system is ever static or complete, even if the organisation was initially determined

33 These ideas are developed in my “On the Importance of a Certain Slowness: Stability, Memory, and Hysteresis in Complex Systems”, Emergence: Complexity and Organization 8:3 (2006): 106–113. 34 Peter Allen, “A Complex Systems Approach to Learning: Adaptive Networks”, International Journal of Innovation Management 5 (June 2001): 149–180. 35 The source of this organisation is a complex issue beyond the scope of our discussion here. Although some systems have a certain organisation imposed on them, complex systems can also develop their structure through processes of self-organisation and evolution, independently of an external designer.

208 

 Paul Cilliers

from the outside. The play of difference leads to constant transformation of the relationships in the system. This process of transformation can be fruitfully understood through the notion of deconstruction. Deconstruction never implies the destruction of structure, but the replacement of one structure with another, which can in turn be deconstructed. Deconstruction is also not an action performed on a system, where the deconstructor is active and the system is passive; the process is suspended between active and passive. Interventions from the outside enter into the play of differences always already at work in the system. Derrida sometimes refers to deconstruction as a characteristic of the system itself: it deconstructs. We should, therefore, understand the organisation of the system as simultaneously something stable and something ready for change. It is this double movement which allows the system to maintain and develop its identity in a dynamic way. This understanding should also influence the way in which we think of organisations. Some of the implications, incorporating insights from the rest of the paper and with reference to more detailed discussions, are summarised in the following points: 1. Differences within the organisation should be seen as a resource. It is not possible to predict when a certain difference, previously seen as unimportant, may become vital. 2. The identity of an organisation is the dynamic result of the differences at work in the organisation and of the way in which they interact with the environment. This identity is intimately related to the generation of meaning, and not to predetermined functions or definitions. It should be stable, i.e. it should resist some external influences, but at the same time it should transform (deconstruct) in order to remain vital. There is no objective way of calculating the tempo of change since both the organisation and its environment are irreducibly contingent. 3. Organisations are not chaotic things. They need structure in order to be able to behave interestingly. The constraints necessary for this should be seen as enabling. At the same time, these constraints should be deconstructed continuously if they are to remain meaningful. The argument for the constrained play of difference is not an argument for rigid control or for maintaining restrictive structures. The fact that some structure is necessary does not imply that all structures are good.36

36 See my “Boundaries, Hierarchies and Networks in Complex Systems,” International Journal of Innovation Management 5 (June 2001): 135–147.



Difference, identity and complexity 

 209

4. If the organisation is seen as a complex system, every aspect of it contains normative elements. Ethics is not only something the organisation “does”, the organisation is constituted through normative processes. This is a result of the nonlinear play of differences which cannot be reduced to a final or objective description. Which differences are allowed to flourish, how much structure is required, how the identity of the organisation is conceived, these are all issues which cannot be reduced to calculation. They rely on judgment and choice. An organisation does not simply make choices, it is its choices.37 The argument that an organisation is constituted through constrained difference also has implications for how one should manage a complex organisation. The elements which comprise the organisation should have a certain amount of freedom. This will enable the organisation to develop new differences, and to adapt to changes in the environment. Components of the organisation should also have a space in which to experiment, even if such experiments are not yet demanded by the environment. This will allow the development of “excess” difference, which is an investment in the future of the organisation. At the same time, however, this freedom cannot be absolute, it has to be dynamically constrained. To find the balance between freedom and constraint is the role of the manager. If the management is too rigid, it will restrict the play of difference, which will result in a loss of meaningful identity and capacity. If the management is purely laissez-faire, the play of difference will be directionless. An organisation does not thrive on chaos as such. The responsible manager is one who can identify which structures and constraints are enabling. These structures, nevertheless, do not arise simply as an act of management. They are systemic properties. Management should be sensitive to the conditions from which meaningful structure could emerge. A certain structure allows a certain pattern of meanings. Such structures need to be continually deconstructed, i.e. replaced with structures which are in turn deconstructable, for the organisation to remain a dynamic entity. The translation of the points argued for in this paper into more specific strategies in organisational theory is a task left to specialists in that field.

37 See my “Complexity, Ethics, and Justice,” Journal for Humanistics (Tijdschrift voor Humanistiek) 5 (Oct. 2004): 19–26.

210 

 Paul Cilliers

4 Conclusions The argument in this essay is primarily one which resists an interpretation of deconstruction, and a post-structural understanding of difference, as an absolute free-play. Deconstruction acknowledges the inevitability of structure, and of its transformation. This “double movement” should be central when we think of institutions and organisations. We cannot do without them, but we should be radically critical of what they should be and become. If the upshot of all this is that diversity should be fostered, is this a process without risk? Certainly not. There may be certain differences we want to resist. We would not want to condone racism, for example, just because it constitutes a difference. The fact of the matter is that we can resist something like racism much more effectively from a richly nuanced position, i.e. a position informed by difference, than by simply rejecting it. The difference we want to reject must be subjected to the play of difference in such a way that a resilient resistance to it can be developed. Critique and dismissal is not the same thing. The way in which we conceive of differences and structures will determine the nature of our institutions, and thus of the world we live in.

Paul Cilliers

Complexity and philosophy On the importance of a certain slowness In philosophy the winner of the race is the one who can run most slowly. Or: the one who gets there last (Wittgenstein, Culture and Value).

1 Introduction As a result of a whole range of what one could call “pathologies” in contemporary culture, the idea of “slowing down” has of late been mooted in a number of contexts.1 A few can be named briefly. The “Slow Food” movement, which started in Italy but has a worldwide following, extols the virtues of decent food made from decent ingredients without compromise. The resistance shown to “junk food” is not only based on aesthetic considerations, but also on ethical (and nutritional) ones. The movement promoting “Slow Cities”, also of Italian origin, fosters an understanding of cities that is more humane. Such cities should encourage walking rather than driving, have small shops with local products rather than shopping malls, and, in general, provide opportunities for the community to interact, not to live in isolation. “Slow schooling” is a movement that questions educational processes in a world geared for instant results. It emphasises the contextual nature of knowledge and reminds us that education is a process not a function. On a more personal level, “slow sex” involves attitudes that try to prevent that the values of the marketplace ruling in our intimate relationships. We need to recognise that the journey is more important than the destination, and that takes time. An immediate or perpetual orgasm is really no orgasm at all. There are a number of very important issues at stake in these examples. In what follows, however, the focus will not be on these social movements as such, but on some of the underlying principles that make the debate on slowness an important one. Through an analysis of the temporal nature of complex systems, it will be shown that the cult of speed, and especially the understanding that speed

1 See Honoré (2004) for a discussion of the emergence of several movements that challenge the “cult of speed”. Originally published in Emergence: Complexity & Organization, 2006, 8(3): 106–113. DOI: 10.emerg/10.17357.bd3be2ec507c9e039579778f0452f0a1 © Emergent Publications.

212 

 Paul Cilliers

is related to efficiency, is a destructive one. A slower approach is necessary, not only for the survival of certain important values or because of romantic ideals, but also because it allows us to cope with the demands of a complex world in a better way. The argument will be made initially by briefly analysing current distortions in our understanding of time. These distortions result, on the one hand, from the rational and instrumental theories we have about a modern world, and, on the other, from the effects of certain technologies, especially communication and computer technologies. In order to show why these are “distortions”, or at least to show why these distortions are problematic, the temporal nature of complex systems will be discussed. The relationship between memory and anticipation will be central to this discussion, but attention will also be paid to the importance of delay and iteration. These characteristics of complex systems have important implications for our understanding of the formation of identity, both individual identity as well as the identity of groups.2 In closing, a number of general cultural issues involving the fast and the slow will be looked at. It is important to realise that the argument for slowness is not a conservative one; at least not in the political sense of the word. It is not merely backward looking nor a glorification of what has been. Although it emphasises the historical nature of knowledge and memory, the argument for slowness is forward looking: it is about an engagement with the future as much as with the past. Slowness is in itself a temporal notion, and in many ways the opposite of the notion “static”. In point of fact, it is actually an unreflective fastness that always returns you to the same place. It should also be stated up front that there is no argument against an appropriate fastness. A stew should simmer slowly, but a good steak should be grilled intensely and briefly. The argument is against unreflective speed, speed at all cost, or, more precisely, against speed as a virtue in itself: against the alignment of “speed” with notions like efficiency, success, quality and importance. The point is that a system that has carefully accumulated the relevant memories and experiences over time will be in a better position to react quickly than one that is perpetually jumping from one state to the other. Perhaps “slow” and “fast” are not exactly the correct terms to use. Terms like “reflective” and “unreflective”, or “mediated” and “unmediated” may be

2 The fact that complex systems demand plural descriptions is in itself something that takes time. Although the link between plurality and delay has to my knowledge not been made explicit, it is implied where there are arguments for multiple descriptions and when “process” is emphasised. See for example Richardson (2005).



Complexity and philosophy 

 213

more accurate. Nevertheless, the debate taking place uses “slow” and “fast”, and the terms do have a certain rhetorical significance. If we stay with their use, it is done in a metonymical way. The whole point of this paper is to give them a richer meaning.

2 Living in the present In Time: The Modern and Postmodern Experience, Helga Nowotny (1994) argues for a certain shift in our experience of time. In short, in my paraphrase, and incorporating insights from Bauman (e.g. Bauman 1992), the argument is the following: One of the main aims of the instrumental rationality flowing from the Enlightenment was to create conditions in which we are not controlled by contingency. To achieve these conditions, it is necessary to understand, and preferably control, the future. This demands coordinated and goal-oriented action in the present. Modernism becomes a project that demands our total commitment against the forces of irrationality and chaos. The modernist project has two important effects on our understanding of time. In the first place, our actions need to be coordinated. This can only happen if time is universalised in such a way that we all live in the “same” time. This was achieved mainly through technology – that is, the construction of accurate clocks – and by regulating time globally. Instead of each person or local community living in their own time, it was necessary to synchronise time in such a way that activities in, say, New York and Paris could be correlated. The effects of this, however, go much further than merely synchronising time in different parts of the globe. It also means that private time and public time are synchronised. We have to live our lives according to a generalised and controlled understanding of time. A subjective, or should one say phenomenological, experience of time has to be sacrificed in order to generate a universal temporal framework in which we can operate efficiently. The second effect of instrumental rationality on our understanding of time is a result of the desire to control the future; for the future to be made knowable. This would only be possible if the future, in some essential way, resembles the present. We cannot anticipate what we do not know, and therefore we should do everything in our power to create a future that does not disrupt the steady progress we are making toward a better world. This modernist strategy is perhaps exemplified best in Hegel’s dialectic of history, which is supposed to converge toward an ultimate solution. The actual result of this ideology is to extend the present into the future, causing us to live in a perpetual “present”. This collapse

214 

 Paul Cilliers

of the diachronic into the synchronic allows instantaneous interaction between everybody; it creates a world that is fast and efficient. The sacrifice made in order to achieve this, however, is nothing short of sacrificing the very notion of temporality. Nowotny (1994: 16) calls it “the illusion of simultaneity”. The way in which contemporary society lives in an eternal present, or what Eriksen (2001) calls the “tyranny of the moment,” is made possible, and augmented, by the surge in technology, especially computer and telecommunication technology. We are instantaneously in contact with everybody, everywhere. Not only has the distinction between home and the workplace collapsed, but also the distinction between work time and private or leisure time. It is expected of many of us to be available, always and everywhere.3 This state of affairs may have been less detrimental if it did not also demand instant response. The very reason for mobile phones and email lies in the fact that immediate response is possible. It is in this “immediate” that the main problem lies. There is less and less time for reflection. Reflection involves delay, and in a cult of speed, delay is unacceptable. This move away from reflection to immediate response has profound implications for our understanding of what it is to be human (see Parkins 2004: 376–379), to which we shall return. The “illusion of simultaneity”, the idea that if we live quickly and efficiently in the present we are somehow closer to reality, is nevertheless exactly that: an illusion. We cannot escape our temporal nature, and our persistence in trying to do so can only lead to pathology. The necessity of delay and reflection needs to be re-evaluated. This can be done from a number of perspectives. A Freudian analysis would show that instant gratification is actually a destruction of pleasure. More sublime pleasure can be found only if desire is delayed, anticipated as a memory of something still to come, yet something that should also in principle be able to surprise us. Derrida calls the illusion of living in the present, of thinking that we have access to an objective understanding of reality if we live “in” it, the “metaphysics of presence” (Derrida 1976: 49). He introduces the notion of différance specifically to undermine the metaphysics of presence (62). Différance is a notion that intertwines difference (as a spatial notion, one could say) and delay (to defer, a temporal notion) as the engines of meaning (Derrida 1982). The present consists only as a combination of memory (of what has been) and anticipation (of what is to come). In his novel Slowness, Milan Kundera (1996) uses the metaphor of somebody riding on a motorcycle as being constantly in the present. Speed and the demands

3 When arrangements are made for more flexible working hours to facilitate parenting, for example, this is already a form of resistance to speed and efficiency.



Complexity and philosophy 

 215

of the machine reduce his horizon to something immediate. Someone walking, however, is moving at a pace that allows for a much wider horizon. The stroll unfolds in time in a way that opens up reflection about where we are coming from and where we are going to as we walk. This theme of both the past and the future being present in a meaningful experience of the present could be pursued in much more detail from both a Freudian and Derridean perspective – and several others too – but the argument for a meaningful temporality – that is, something slower – will be made here from the perspective of the dynamics of complex systems.

3 Complex systems, temporality, and memory An important aspect of complex systems, one that certainly complicates our understanding and modelling of such systems, is their temporal nature. Complex systems unfold in time, they have a history that co-determines present behaviour and they anticipate the future. Moreover, as we know at least since the work of Prigogine, the behaviour of complex systems is not symmetrical in time. They have a past and a future that are not interchangeable. This being “situated in time” does not always receive adequate attention in our analysis of complexity. The central notion at stake when we talk of time and complexity is that of “memory”. Memory is the persistence of certain states of the system, of carrying something from the past over into the future. It is not merely the remembering of something in the past as if belonging to that past, it is the past being active in the present. We should therefore not think of memory in abstract terms, but of memory as something embodied in the system. In many respects the system is its memory. If one accepts an understanding of complexity that emphasises the relational nature of the system, it is useful to think of systems as networks where the connections between the nodes are more important than the nodes themselves. The nature of these connections is a result of which states of the network are “retained”, thus the structure of the system is a result of the sedimented history of the system.4 It is important to remember that memory is not merely a cumulative process. The structure in the network of relationships can only develop if certain states of the network are not maintained. Memory is a result of a process of selection. The states that are significant are repeated more often and therefore form more

4 This argument can also be made using the example of the brain, and links with many Freudian arguments in an interesting way. See Cilliers (1998: 45–47, 92, 108) for further discussion.

216 

 Paul Cilliers

permanent links in the network. Less significant states will fade away over time. Memory is only possible if the system can also forget.5 What is important to note at this stage is that memory is not an instantaneous thing, it takes time to develop, it is slow. If one characterises memory as the past being carried over into the future, it follows that the future can only be anticipated in terms of the memory of the system. Anticipation is not, or at least should not be, simply an extrapolation of the present. It is a complex, nonlinear process that tries to find some trajectory, some way of “vaulting” from that which has already been experienced to that which has to be coped with. The quality of the anticipation is a function of the quality of the memory. A more varied, richer, deeper, and better-integrated memory will open up more sophisticated anticipatory capabilities. The obvious question now would be to ask how such a rich memory is formed. This is a complex issue, but for the sake of the argument at stake here, one can say the following: Memory is information from the environment that has been filtered, it is that which has been interpreted – by the memory already sedimented in the system – as significant. The identity of the system is, in some sense, its collection of dynamic memories. The implication is that the system cannot reflect, or act on, everything that is going on in the environment at a given moment. If that were the case, the system would always be merely a reflection of its environment and would have no identity of its own. In order for it to be a system at all, a system that has its own identity, that can react to the environment and not merely mirror it, a certain hysteresis is required. The system must be slower than its environment.6 The notion of hysteresis is an important one.7 It links to the notions of delay and différance discussed above. An event in the environment of the system does not have inherent and immediate significance for the system. Its significance

5 This process is known as the “use principle” or Hebb’s rule. For more detail see Cilliers (1998: 17–18, 93–94). 6 Perhaps this claim needs to be qualified somewhat. Strictly speaking, the system must change at a different rate than its environment. There are many activities within as well as outside the system (without introducing the problem of boundaries here; see Cilliers 2001), thus it is difficult to talk of the “rate of change” as a single measure. Some processes in the system may be faster than the environment, but they are possible only within the context of more general structures that have been formed through iterative processes that involve recycling and delay. One could perhaps say that the system should be slower than its environment “on average”, but this could lead to other confusions since “average” is a problematic notion in this context. At the very least one could say that processes involving delay are crucial in identity formation. 7 Hysteresis is the “lagging of effect when cause varies” (Oxford Concise English Dictionary).



Complexity and philosophy 

 217

is established in time as it is re-enacted in the system and carried over into the future. In a way, the significance of an event has always already been established (in terms of the memory of the system), but never completely or finally, since the significance is always also to be determined by what is still to come. The system has to hang on to some aspects with a certain tenacity: not let go of them too quickly. There is risk involved in this, of course. The system has to invest resources in this process. It cannot maintain everything; it has to select. If too many of the wrong things are carried over it will impair the system’s performance. However, if not enough is carried over, it will also fail. To put it in slightly different terms: The system has to find a way to discriminate between information and noise.8 If it follows every trend in its environment, it will also be following noise. If it reacts too slowly it will only follow the low-frequency trends, which may also be just noise. The system must be stable enough not to be buffeted around by every fluctuation, and it must be flexible enough to be able to adapt when necessary. Where this optimal point lies is not a question that can be answered from an objective viewpoint. The balance between stability and change is a contingent thing that plays itself out in time. What one can say, though, is that merely to be fast will destroy the system. The argument for a certain slowness should start to take shape now. A viable system has to be able to resist some of the dynamics in its environment. There should be a temporal space in which the past is allowed to play itself out in interaction with the present. There must be time for reflection and interpretation. The faster the system becomes, the shallower its resources will be. Ultimately quick behaviour will be no more interesting than Brownian motion. It must be stressed again that the argument for a certain slowness is not a conservative argument. A certain amount of conservation is a prerequisite for a system to maintain itself, of course. The important point, to which we shall return, is that a “slow” strategy is not a backward-looking one. If a somewhat slower tempo allows a system to develop a richer and more reflective memory, it will allow the system to deal with surprises in its environment in a better way. The argument of slowness is actually an argument for appropriate speed. There is no objective or immediate rule for what that speed is. If anything, it is a matter of experience, and experience (as Aristotle urged) has to be gained, it cannot be “given” in an immediate way. It is experience that determines which piece of meat

8 The distinction between information and noise is, for the system, a strategic choice. There is no predetermined criterion by which they can be separated. This, however, does not imply that all noise is information just waiting to be framed in the appropriate way. See the discussion of Taylor below.

218 

 Paul Cilliers

should be fried quickly and which should simmer slowly in the stew. She who fries everything quickly will only have something nice to eat now and then, and then purely by chance.

4 Integrity, identity and reflection In his novel The Discovery of Slowness, Sten Nadolny (2003) gives us a fictionalised account of the life of John Franklin. Franklin, a 19th-century explorer primarily obsessed with finding the Northwest passage, is slow. His advance (in the Royal Navy) is also slow, mainly because being slow is confused with being stupid. Since he is not stupid he is gradually awarded command, and those working with and under him discover the advantages of being slow. Franklin is persistent, dependable and trustworthy. Even in war, thorough reflection pays dividends that are not always immediately apparent. His political career, as governor of Van Diemen’s Land (now Tasmania), ends badly only because he is disgraced by those out for quick and selfish results. His principles are not negotiable. Franklin is a worthwhile human being because he has integrity. There is a substance to his personality that may seem opaque at first, but eventually shows itself as solidity. The nature of his integrity is directly coupled to his slowness. He assimilates, integrates and reflects before he acts. This is sometimes a ponderous process, and he pays a price for it. Under normal circumstances it is easy not to notice someone like this, or to pass him by, but when there is a crisis, it is him people turn to. He can be trusted, he will come up with something. This is most significant. It is exactly when one would think that being fast is what is required that slowness proves its worth. The link between slowness and integrity is also an issue in J.M. Coetzee’s (2005) novel Slow Man. Here we have a character who resists change, despite the cruel demands being made on him. He clings to a set of values that are important to him, and this gives his personality substance. However, he is too stubborn, and eventually he cannot adapt to new circumstances. One has tremendous sympathy for him, but he turns out to be too slow, and pays the price for it. Even so, it is clear that when there is a choice between the loneliness of the slow and the superficial companionship of the quick, the author sides with the slow. Integrity is more important than a certain kind of success. Despite Coetzee’s darker view, there is no reason why slowness should be solitary and sad. Quite the contrary is true. In his novel Slowness, Milan Kundera (1996) shows with great conviction how a certain slowness is a prerequisite for being fully human. What is at stake in this novel is not moral integrity or a kind



Complexity and philosophy 

 219

of Calvinist dependability, but the sensuality of human interaction, the beauty of a relationship that unfolds in time, the ecstasy of a love that has a history and a future. Being human implies having a body, something with its own rhythms and demands. If we reduce all of this to something merely instrumental, to transactions written in legal terms (not in lyrical prose), if we demand results now, then we will stop being human. Language cannot be reduced to a code; it plays itself out in a certain context. What is more, even if we immerse ourselves in the context we have to wait beyond the last sounds. When all is said, the meaning has not finally arrived yet. It is the anticipation of what it could yet mean that draws us forward. Einmal ist keinmal. Many may feel that the novel is an outdated art form, something to be replaced with the fast and immediate communication of the digital code. In his book The Moment of Complexity, Mark Taylor (2003) seems to lean in this direction. For him, something of a paradigm shift has occurred in the last few decades. We live in a new world with new forms of communication and new forms of learning and human interaction – something he seems quite willing to sell. He resonates with a fast world, something new and exciting. Taylor’s emphasis on the new goes hand in hand with a nearly interchangeable use of the notions “noise” and “complex”. This problematic conflation is the result of an understanding of complexity primarily informed by chaos theory and of information as entropy. Such an understanding, inspired by the theories of Shannon and Chaitin, will attribute the highest information content to a purely random sequence.9 Although these notions are important in the context of computation, they are less useful when talking about complex systems in general. Living systems, including the social systems that Taylor explores, are neither random nor chaotic. Despite the fact that they are constituted through nonlinear interaction and that they are capable of novel and surprising behaviour, they are well structured and robust. They persist through time and maintain themselves. When we encounter behaviour that we do not understand, or cannot decode, it often looks like noise, but once it is understood we can see the patterns. These patterns are not merely or only an order imposed by the observer, but also characteristics of the system itself. Complexity may look like noise, but all noise is not something complex waiting to be decoded. Sometimes noise is just noise.10 Taylor’s argument is seductive, but, to my mind, wrong if not harmful. In his fervour to embrace the posthuman, he looks at the history of being human with a

9 See Hayles (1999) for a discussion of these issues. Primary sources are Shannon (1949) and Chaitin (1987). 10 A similar, and more detailed argument is made in Cilliers (2005).

220 

 Paul Cilliers

certain disdain. It seems as if he thinks that complexity is a recent discovery and forgets that being human has always been complex. He embraces the present and wants to deal with it quickly and efficiently. We can be educated instantaneously by electronic means and thus we should make a radical break with old methods. In his excitement he forgets that complex systems, even those in our postmodern world, are constituted historically, that they develop and change, and that one of their primary functions is to distinguish between information and noise. This cannot be done at the press of a button. The ideas of the posthuman and the cyborg are of undeniable importance,11 but in our enthusiasm to embrace new modes of being we should be careful not to effect a transformation into something inhuman.12 Machines are fast, but they are machines. The present argument is not for an a priori rejection of the possibility of machines with human capabilities, or one that denies the intimate relationship that humans have always had with technology. Our cultural existence presupposes the use of tools. The difference between using a quill and a word processor may have huge implications on a practical level, but they also share some essential features.13 The notion “posthuman” is thus an ambiguous one. If it signifies a tight coupling between the body and technology, we have always been posthuman. If it signifies the obsolescence of the body, perhaps in the sense that a “person” could be downloaded instantaneously as software and run on a machine, it becomes a problematic notion at odds with the idea that human identity is also the result of a certain temporal embodiment. The general argument presented here maintains that any complex system, organic or not, would have to incorporate a certain slowness. The need for slowness, and a warning against the embracing of the fast, can perhaps be motivated best from the perspective of philosophy. Philosophy, in its most general form, is essentially the art of reflection. Wendy Parkins (2004) analyses contemporary culture as one moving away from reflection, and argues that what we need is an “ethics of time”. She does not elaborate much on what such an ethics should look like, but it is something that needs careful attention, not only from a moral perspective but also from the purely pragmatic perspective of

11 See Badmington (2000) for a collection of philosophical essays on the posthuman. 12 See Hayles (1999) for a detailed discussion of cybernetics, the development of the posthuman, and the importance of embodiment. See Braidotti (2005) for an affirmative discussion of the posthuman that is neither a euphoric, uncritical acceptance of advanced technology, nor a nostalgic lament for the decline of classical humanism. 13 There would definitely be a lot less drivel to wade through if it were not possible to write so quickly.



Complexity and philosophy 

 221

how to live and survive in a fast world. Such an “ethics” will be complex in itself. It will have to unfold in time and be conscious of its own temporal nature. For now, instead of rushing around like the Red Queen in a world where change is virtuous merely because it is change, we can start by taking some time out to reflect. At this point the argument for slowness becomes a political one: We should put up some resistance to a culture in which being fast is a virtue in itself. We should say “no” with a little more regularity.

References Badmington, N. (ed.). 2000. Posthumanism. London, England: Palgrave. Bauman, Z. 1992. Intimations of Postmodernity. London, England: Routledge. Braidotti, R. Cyberfeminsim with a difference. http://www.let.uu.nl/womens_studies/rosi/ cyberfem.htm Chaitin, G.J. 1987. Algorithmic Information Theory. Cambridge, England: Cambridge University press. Cilliers, P. 1998. Complexity and Postmodernism: Understanding Complex Systems. London, England: Routledge. Cilliers, P. 2001. Boundaries, hierarchies and networks in complex systems. In: International Journal of Innovation Management 5(2): 135–147. Cilliers, P. 2005. Complexity, deconstruction and relativism. In: Theory, Culture & Society 22(5): 255–267. Coetzee, J.M. 2005. Slow Man. London, England: Seeker and Warburg. Derrida, J. 1976. Of Grammatology. Baltimore, MD: John Hopkins University Press. Derrida, J. 1982. Différance. In: Derrida, J. Margins of Philosophy, Chicago, IL: The Harvester Press, 1–27. Eriksen, T.H. 2001. Tyranny of the Moment: Fast and Slow Time in the Information Age. London, England: Pluto Press. Hayles, N.K. 1999. How We Became Posthuman: Virtual Bodies in Cybernetics, Literature, and Informatics. Chicago, IL: The University of Chicago Press. Honoré, C. 2004. In Praise of Slowness: How a Worldwide Movement is Challenging the Cult of Speed. London, England: Orion. Kundera, M. 1996. Slowness. London, England: Faber and Faber. Nadolny, S. 2003. The Discovery of Slowness. Edinburgh, Scotland: Canongate. Nowotny, H. 1994. Time: The Modern and the Postmodern Experience. Oxford, England: Polity Press. Parkins, W. 2004. Out of time: Fast subjects and slow living. In: Time and Society 13(2): 363–382. Richardson, K.A. 2005. The hegemony of the physical sciences: An exploration in complexity thinking. In: Futures 37: 615–653. Shannon, C.E. 1949. Communication in the presence of noise. In: Proceedings of the IRE. 37: 10–21. Taylor, M.C. 2003. The Moment of Complexity: Emerging Network Culture. Chicago, IL: The University of Chicago Press.

Part 2: Posthumous after 2011 Theme 1: Critical Complexity

Rika Preiser, Paul Cilliers & Oliver Human

Deconstruction and complexity A critical economy

1 Introduction The notion of “complexity” is currently subject to a number of interpretations (Rasch 1991, Chu et al. 2003, Alhadeff-Jones 2008). We argue that a perspective on complexity informed by deconstruction is particularly fruitful. Not only does it help to develop a theory of complexity that resists positivist reductionism, it also contributes to the development of deconstruction in a way that resists faddish extravagances. After a brief introduction to what is meant by the notion of complexity and how it relates to deconstruction, the focus will be on the notion of “economy”. It will be illustrated how the economy of deconstruction can be applied to a critical theory of complexity. The general economy upon which deconstruction is based allows one to make provisional interventions in the world, yet forces one to acknowledge that these interventions are always contingently based upon a set of exclusions one necessarily has to make in order to be able to say anything about the world at all. This forces one into a position of modesty. At the same time it opens up the possibility for a truly critical project to arise. In the third part of this paper we therefore explore the notion of critical complexity as a basis for opening up a new form of critique made possible by the joining of complexity thinking and deconstruction. Throughout this paper we underline the inevitable normativity of our engagement with complex systems and in conclusion we expand upon the possibility of an ethics of complexity, underlining the provisional nature of such a programme.

2 Systems of meaning Complexity is the characteristic of a system, not of atomistic entities. The link between deconstruction and complexity can thus be made through the notion of “system of difference”. Such a system of differences can be used to describe how a complex system works in much the same way as Ferdinand de Saussure Originally published in the South African Journal of Philosophy, 2013, 32(3): 261–273. DOI: 10.1080/02580136.2013.837656 © South African Journal of Philosophy.

226 

 Rika Preiser, Paul Cilliers & Oliver Human

(1974) described how language works. These descriptions can then be meaningfully expanded by incorporating insights gained from Derrida’s deconstruction of Saussure (Cilliers 1998: 1–7, 37–47). Systems understood in this way consist of a number of components that interact nonlinearly. The properties of the system do not reside in the components, but is a result of these interactions. If these interactions were ordered, homogenous and symmetrical, no interesting behaviour would arise. There has to be asymmetry. This is another way of stating that the relationships between the components are relationships of difference. If one sticks to a purely structuralist (i.e. Saussurian) understanding of complex systems, one ends up with a model which argues that things may be relational and very complicated, but if you work hard enough, with clever enough techniques, you can figure the system out – essentially the general structuralist claim. This understanding would correspond to what Morin (2007) calls “restricted complexity”. A “general” understanding of complexity requires a more reflexive and transformative approach. It is exactly in this respect that deconstruction makes a vital contribution. It allows us to describe the dynamic nature of the play of differences. The interactions in a complex system can be described by Derrida’s concepts of trace and différance. The concept trace can be used to refer to the individual differences between the components in a system. Each trace has no meaning in itself, but through their interaction the meaning of a sign emerges. The notion of différance can be used to describe the dynamics of complex networks. The analogy works in the following way: the interaction between a number of components in the system generates a pattern of activity, traces of which reverberate through the whole network. Given there are loops in the network, these traces are reflected back after a certain propagation delay (deferral), and alter (make different) the activity that has produced them in the first place. Given that complex systems always contain loops and feedback, delayed self-altering will be one of the network’s characteristics; a characteristic described quite precisely by Derrida’s notion of différance – a concept that indicates difference and deference, that is suspended between the passive and active modes, and that has both spatial and temporal components (Derrida 1982: 1–27). Difference is therefore not simply the static differences between components in the system; they are constantly transformed. Nevertheless, if we merely insist on an abundance of difference that is irreducible, we are not saying enough about how complexity is constituted. A limitless play of difference does not lead to the generation of meaning, nor can a complex system function without being constrained in some way. On the epistemological level (our descriptions of complex systems) as well as the ontological level (the functioning of complex systems in the real world), boundaries are required. There



Deconstruction and complexity 

 227

is a certain “economy” involved. A better understanding of this requires a closer look at the roles of limits and constraints.

2.1 Knowledge and limits To fully understand a complex system, we need to understand it in all its complexity. Furthermore, because complex systems are open systems, we need to understand the system’s complete environment before we can understand the system, and, of course, the environment is complex in itself. There is no human way of doing this. The knowledge we have of complex systems is based on the models we make of these systems, but in order to function as models – and not merely as a repetition of the system – they have to reduce the complexity of the system. This means that some aspects of the system are always left out of consideration. The problem is compounded by the fact that that which is left out, interacts with the rest of the system in a nonlinear way and we cannot, therefore, predict what the effects of our reduction of the complexity will be, especially not as the system and its environment develop and transform in time.1 We cannot have complete knowledge of complex systems; we can only have knowledge in terms of a certain framework. There is no stepping outside of complexity (we are finite beings), thus there is no framework for frameworks. One should be careful not to interpret this state of affairs as somehow inadequate, as something to be improved upon. There is a necessary relationship between the imposition of a limiting framework and the generation of knowledge. One cannot have knowledge without a framework. Despite the fact that our knowledge is of necessity limited, these limits are enabling, they allow us to make claims that are neither relativistic nor vague (see Cilliers 2005). At the same time, however, such knowledge is not the result of free-floating truths; it is contextualised in time and space. Because it is not objective, and because we know that, we cannot use this knowledge as if it is objective. There is always a normative dimension to the claims we make, and we have to stand in for them. We cannot shift the responsibility for the effects of our claims onto some process we call “scientific”. It should be clear that this position shares the most important aspects of that which constitutes deconstruction.

1 These ideas are elaborated upon in Cilliers (2000, 2001).

228 

 Rika Preiser, Paul Cilliers & Oliver Human

2.2 Derrida and Morin It is clear that Derrida’s argument is based on the fact that meaning is constituted through complex interaction. Although he did not explicitly develop a theory of complexity, a sensitivity to complexity permeates his thinking. There are still many fruitful insights to be gained from a sustained interaction between deconstruction and complexity theory.2 This interaction can be commenced by comparing some aspects of deconstruction with the work of Edgar Morin.3 Morin distinguishes between a “general” and a “restricted” understanding of complexity (Morin 2007). The “restricted” understanding is clearly related to the Saussurian position. It acknowledges the basic structure of complexity, but balks before the more radical consequences. In Morin’s terms, it opens up the understanding towards relational thinking, but it cannot get rid of the reductive apparatus that should qualify this work as “science” (Morin 2007). As a result, this approach to complexity – and we would put much of the work historically done under the umbrella of the so-called Santa Fè School in this category – reverts to an instrumental strategy in the hope of making purely objective claims in the same way as Saussure’s claim that we can get at the correct meaning of the sign. It is precisely this denial of a normative element in our dealing with complexity which makes this position “restricted”. In developing a deeper understanding of what a “general” understanding of complexity could be, something for which Morin thinks we do not yet have a language, insights from deconstruction could play a vital role (Morin 2007). In the paradigm of “general” complexity the assumptions made by classical science are not taken for granted. Although one has to make some assumptions in order to get something done at all, these assumptions should be seen for what they are, namely reductions of the complexity. In opposition to reduction, complexity requires that one tries to comprehend the relations between the whole and the parts. The principle of disjunction, of separation (between objects, between disciplines, between notions, between subject and object of knowledge) should be substituted by a principle that maintains the distinction, but that tries to establish the relation (Morin 2007: 10–11).

2 See also Preiser and Cilliers (2010: 271–274). 3 Edgar Morin is a French intellectual. He is Emeritus Director of Research at the CNRS (the French National Research Centre). His latest book, On Complexity (2008), is a summary of his multivolume work on the notion of complex thought.



Deconstruction and complexity 

 229

In many ways, this dialogical thinking agrees with the double movement of deconstruction. Derrida (1998) argues that the strategy of deconstruction involves a “double” activity. In deconstructing a system, one has to make use of the resources provided by the system itself. One is thus simultaneously confirming and undermining central elements of the system. This simultaneous give and take is a much more complex process than simply replacing something with something else. It implies that one transforms something by using the thing itself in novel ways. Deconstruction is thus not a critique from the outside, a critique that knows where it stands and what it wants to do. It is a critique that acknowledges that it is in transformation itself because it cannot depart from a perfect understanding, neither of itself, nor of that which it is transforming. In many ways dealing with complex problems has to engage with the same dynamic. When we use models we make exclusions and simplifications, consciously or not, and these exclusions have effects. The benefit gained from a “general” notion of complexity is precisely to make us aware of the fact that the process of modelling the world, an unavoidable process, is never neutral or objective. The difficulty in constructing models of the world is partly explained by the notion of “play”. Play is inaugurated by the differential and as a result of the nonlinear interactions that take place within complex systems. This play is subject to the economy of différance and therefore resists any form of comprehensive reduction. Nevertheless, complex systems, such as systems of language and living systems, are also robust, they persist in time despite their complexity. This robustness can fruitfully be compared to Derrida’s notion of “iterability” (Derrida 1977: 7). Although systems constantly change, there is simultaneously structural “repeatability” that allows one to say something about the system that is not arbitrary or relativistic. This iterable economy of the system opens up the possibility of modelling the system, albeit it never in a final way. Thus, complex systems are simultaneously limited enough for us to be able to say something about them and “open” enough for them to constantly slip from the models we create. One can look at the relationship between the notions that constitute our models, as well as the relationships that constitute the system itself, as a type of economy similar to that upon which deconstruction works.

3 The economy of complex systems When we think of the term “economy” a certain set of meanings is brought to mind. Primarily these include usages of the term in relation to the study of economics. The dictionary definition of an economy is a twofold definition of, first,

230 

 Rika Preiser, Paul Cilliers & Oliver Human

the concern with the production and consumption of resources and, second, the orderly interplay between the parts of a system (Collins English Dictionary 2006). In effect then one can state that an economy is the concern with the production and consumption of resources made possible by the orderly interplay of the parts of a system. However, when one thinks through the general use of the term a defining feature appears to be that in order for an economy to exist there must be some form of scarcity, thus its usage in statements such as “to economise” or “economical”. A consequence of the limited use of resources is that a system is developed which needs to prioritise what should be produced and in what amounts (Fleming 1969). However, this system of prioritisation not only determines what should be produced but also what can be produced, it determines or establishes the system, the orderly interplay between parts. An economy then operates on the principle of a relationship of feedback between the use of limited resources (production and consumption) and what the system is able to do with these resources and vice versa. This system of production and the limited resources it exploits results in the fact that certain areas or facets of life are prioritised over others. As Derrida (1992: 6–7) argues in relation to the classical Hegelian notion of economy: Among its irreducible predicates or semantic values, economy no doubt includes the values of law (nomos) and of home (oikos). Nomos does not only signify the law in general but also the law of distribution (nemein), the law of sharing or partition (partage) [...] As soon as there is law, there is partition: as soon as there is nomy, there is economy. Besides the values of law and home, of distribution and partition, economy implies the idea of exchange, of circulation, of return. The figure of the circle is obviously at the centre, if that can still be said of a circle [...] This motif of circulation can lead one to think that the law of economy is the-circular-return to the point of departure, to the origin, also to the home.

In order for a system to produce something it must receive in return, indeed it would be difficult to conceive of an economy, as an economy, that only distributes or disseminates without return. This can also be read as the foundation or basis of a system, in a certain sense then an economy, conceived of as a circle, is inherently conservative, it aims to preserve its point of departure or its structure. Yet these limits are not necessarily bad things, not only are they enabling (Cilliers 2001) but furthermore they are necessary for the existence of the system in the first place, something that has no boundaries, which is claimed to be and do everything, is indeed nothing. In comparison to the distinction between restricted and general complexity, George Bataille (1989, 1991, 1993) makes a distinction between restricted and general economies. In his discussion of Bataille, Derrida (1978) argues that one cannot postulate two economies, the restricted, utilitarian economy Bataille critiques and the “other” of this economy, the general economy, which experiences



Deconstruction and complexity 

 231

only loss and waste. Rather, one must read the notion of a general economy as if it is a restricted economy. It is this understanding of economy that we pursue in complexity thinking. This is due to the fact that when we model a complex system our analysis will always be restricted, due not only to our limited perspectives but also due to the fact that our models are obliged to fulfil the demands of reason, coherence or logic and will thereby always be exceeded by an excess. This is due to the fact that excess, by definition, exceeds reason (Derrida 1978: 255, quoting Bataille). It is as such that despite being limited in a certain sense, an economy at the same time will always be exceeded by an excess that it cannot capture and reincorporate into its circular imagery. What the notion of general economy does is to establish a relationship to this excess (Derrida 1978), it establishes a relationship to the “loss” that an economy experiences, to that which does not return to its origin and to the possibility of using limited resources without any utilitarian gain. Therefore, in Derrida’s exploration of Bataille we can note the double-handed movement of the deconstructive process that aims to maintain the radical nature of Bataille’s critique whilst at the same time illustrating the impossibility of a “pure excess” without an economy to which it corresponds. Derrida argues that we need to remember that we can only speak of one economy (of one discourse); it is senseless in this regard to postulate two different kinds of economy, one restricted and the other excessive or general. That is, we cannot postulate an economy of excess that runs parallel to the restricted economy or a general economy in which there is only excess. When we speak of a general economy, it is not an economy separate from a restricted economy, rather, it is a single economy that is not closed but is both open to random chance events as well as predictability, open to the possibility of destruction and yet robust, whether it comes from the play of forces inside the system or from its relationship to its environment (Derrida 1978). The models we construct of such systems are therefore often adequate for the systems we aim to model. However, we must keep in mind the fact that a system does not run on an entirely rational, utilitarian basis but is open to the possibility of paradox or inconsistencies, as well as the effects of play, which will always escape our ability to model. In this light our understanding of complex systems as operating under a general economy must walk a narrow strait. In one sense, our models of complex systems are often adequate for the task at hand. However, we must be careful to build on top of this adequacy an arrogance that assumes mastery over the world. In the same light, stepping back and stating that we cannot act in the world due to the excess that exceeds us, leaving this world to the hands of Gods or scientism, is equally problematic. The acknowledgement of complexity therefore demands a critical stance that includes deconstructing (opening up, transforming, not dismissing) the foundations and limits upon which knowledge is built.

232 

 Rika Preiser, Paul Cilliers & Oliver Human

3.1 Complexity and critique In her book called Philosophy in Turbulent Times, Elisabeth Roudinesco (2008) argues that the lack of critical thought is marked by intellectual activity that only knows how to “classify, rank, calculate, measure, put a price on [and] normalize [...] in the name of a bogus modernity that undermines every form of critical intelligence grounded in the analysis of the complexity of things and persons” (p. xi). Roudinesco reminds us that the last word on the critical project has not yet been spoken. An “and” remains. Reflecting on the meaning of the concept “and”, Derrida (2000) responds to an edited collection of essays (Royle 2000) by offering a parody on the logic of the “and”, and notes that the conjunction “is resistant not only to association but also to serialization, and it protests against a reduction that is at bottom absurd and even ridiculous [...]” (Derrida 2000: 283). The opening lines of his essay read as follows: And in the beginning, there is the and. What is there in an ‘and’? And I wonder what a deconstruction can do with such a little, almost insignificant word (Derrida 2000: 282).

And here we are again, suggesting an “and” of some sort. This time, the workings of deconstruction are supplemented with the notion of “complexity and critique”. This “and” not only joins seemingly unrelated paradigms of thought but, as Derrida reminds us, it also represents a gap between the two traditions. Hence, the double bind of the gap and the bringing together of these two traditions cause a rupture in the unsurpassable dilemma of the somewhat stale critical project. From this rupture the possibility of critique is born. By thinking “deconstruction” and “complexity” together, a liminal space is opened in which both traditions are challenged to negotiate new ways of thinking and knowing that may infuse “critique’s grammar” (Pavlich 2005) with new vigour.

3.2 The crisis of critique Recounting the possible difficulties one is faced with by subscribing to postmetaphysical forms of critique, recent literature on the notion of critique suggests that “the search for new critical directions” is driven from “a sense that critique has become stagnant – has fallen into crisis or malaise” (Chryssostalis & Tuitt 2005: 1). To explain in more detail, it seems that the “crisis” critique faces today stems from the historical changes it underwent since the inception of the



Deconstruction and complexity 

 233

critical project with Kant. The problem of the possibility of a postmetaphysical form of critique more specifically, is directly related to the problem of the “loss of the outside”, which in turn can be identified as a problem of legitimisation as famously argued by Habermas (1975). The insistence of poststructuralist positions to reject any foundationalist position from where to ascertain the validity of criteria for legitimising or justifying the acceptance of certain norms remains a challenge to the critical tradition. If there is no outside or metaposition from where critique can assume a basis in which to situate its grounding norms, then the whole critical project falls prey to relativism and indifference or to what Sloterdijk (1987: xxxii) calls the “self-abdication of critique”. By asking the questions “What can critique achieve today? What can it still hope for in a time that is so sick of theory?”, Sloterdijk (1987: xxxii–xxxiii) suggests that in the process of Enlightenment critique, the critical project has undermined its own impetus. By having exposed the foundations from where the norms of critique could be legitimised, the loss of the outside also implies that there is no standpoint for a description, no central perspective for a convincing form of critique. Moreover, along the way critique has lost its ability to perform exactly what it set out to do originally, namely to be a cutting force that lays bare the conditions of possibility under which reason can know and judge the world in which we live. With the loss of a fixed foundation on which critique can be grounded, forms of postmetaphysical critique has not only run into the difficulty of finding alternative concepts or groundings from where to justify a non-foundational, non-essentialist critique but, moreover, the notion of “critique” itself has been compromised. In the process of wanting to avoid transcendental categories and metaphysical positions from which it launches its judgements, critique has lost its conceptional ground from which it could offer any substantive norm or name in which to incur judgement. It seems that the end of the age of grand narratives and the death of God also silenced critique’s capability to start a revolution in philosophical thinking. Fear that the modernist project might be revoked by the ghost of utopia and the accusation that one might actually know what a perfect society should look like are also affecting the possibility to come up with vibrant notions of postmetaphysical forms of critique. The loss of metaphysical claims robbed poststructural forms of critique from the possibility to bring forth a “philosophy of heroism” (Roudinesco 2008), seeing that postmetaphysical forms of critique are often criticised to subscribe to a form of pluralism that is “as anarchistic as anything goes” and as “nihilistic as ‘nothing matters’” (Hoy 2005: 231).

234 

 Rika Preiser, Paul Cilliers & Oliver Human

3.3 C  ritique as stricture: simultaneous rupture and reconciliation In what follows, a deconstructive reading of the concept critique will be undertaken in order to find a solution to the problem of legitimation. An analysis into the nature and meaning of the word “critique” reveals that the concept itself deconstructs4 when its roots are traced back to the meaning Kant (1998) ascribed to it in his critical project. The implementation of the word Kritik in the titles of Kant’s three Critiques, refers to an understanding of the word as found in the tradition dating back to Cicero and the appropriation thereof in relation to the interpretation of “ars critica” (as read in French) as “critique” during the seventeenth century (Höffe 2004: 35). As argued by Höffe (2004) this notion of critique (as it was used in theatre and art criticism) was appropriated by Kant and employed in a manner contrary to the common use of the word, which was understood to mean “disapproval”, “objection” or “disagreement”. In addition, Höffe (2004: 35) explains that the term Kritik in Kant’s critical project should be interpreted to coincide with the process of judgement or thorough examination as found in a court of law. Kant’s interpretation of critique as a form of judgment unmasks the assumptions of traditional metaphysics in the same way criminal proceedings examine the false intentions and statements of the accused. Hence, the examination of the possibilities and limits of reason resembles a judiciary trial (Van Niekerk 1980: 164, Höffe 2004: 35). As Bernstein (1991: 6) argues, a judicious critique is “the type of critique where one seeks to do justice to what is being said and also ‘steps back’ in order to evaluate critically strengths and weaknesses, insights and blindnesses, ‘truth’ and ‘falsity’” (italics in original text). Related to this interpretation of the notion of critique as judgement, the task of the critic can be explained as that of being a “good judge” who can evaluate the “authenticity, truth, validity or beauty of a given subject matter” (Benhabib 1986: 19). For Kant, the use of the judiciary metaphor gave him the opportunity to put a case forward that philosophical method should not exclusively use the mathematical method as standard model for verifying truth claims. However, a closer look at the concept “critique” reveals that in addition to signalling the act of passing judgement, it also means “incisive cutting” (Hanssen

4 Gasché (1994: 26) describes the operation of deconstruction as to “represent the moment where, in a text, the argument begins to undermine itself; [...] or the relation of a message of communication to itself that, thus, becomes its own object; or finally, the self-revelation and indication by the text of its own principles of organisation and operation”.



Deconstruction and complexity 

 235

2000: 4, Alhadeff-Jones 2010: 48). Etymologically, the word kritik is derived from its Greek roots, Krinein and kritikos, which can be translated to mean “to cut, rift, separate, discriminate, but also to decide” or “passing judgment” (Hanssen 2000: 4). Following the economy of the double movement, the interpretation of critique that coincides with the process of judgement ruptures itself when the German form of the word “judgement”, namely Entscheidung (meaning “decision” or “judgement”), is considered in deconstructive fashion. A certain self-division (or deconstruction)5 occurs that can be explained as follows: When the German term Entscheidung is orthographically divided into “Ent-” plus “-Scheidung”, it literally becomes to mean “de-” plus “separate”, thus: to de-separate or to put an end to Scheidung, therefore, to bring together or reconcile. Hence, the self-cutting overturns the hierarchy (to judge/to cut) and a new or supplemented understanding of critique arises. A displacement takes place. Not only does it become apparent that what seemed to be the primary term “to judge” is dependent on “to cut”, but when we overturn the hierarchy and use the secondary meaning “to separate” (or “to rupture”), the displacement reveals the double bind that is present in the concept of critique: at once it means to judge and cut apart, but also to de-separate, thus to bring together and reconcile. Here the logic of the “and” becomes apparent as Derrida (2000: 282) proposes that “deconstruction introduces an ‘and’ of association and dissociation at the very heart of each thing, [...] it recognises this self-division within each concept”. The key for this new impetus lies in the “and” that is also located in this deconstructed understanding of critique. Moreover, it is exactly in this double bind that the inner logic of the notion of critique is exposed to be what Derrida calls a “stricture” that characterises the new displaced form of the concept critique. Derrida (1998: 36) proposes that it is [...] this double bind [that] is the question of analysis itself. Not that one must assume the double bind. By definition a double bind cannot be assumed; one can only endure it in passion. Likewise, a double bind cannot be fully analysed; one can only unbind one of its knots by pulling on the other to make it tighter, in the movement called stricture.

Falling into the same category as Derrida’s other terms such as différance, trace, supplement, the double bind and stricture are not “substitutions for a truth”or “one-place function” (Hobson 1998: 162). It is this sense of stricture that inhabits

5 “Through the play of this separation between the two markings, it will be possible to carry out at the same time a deconstruction by inversion and a deconstruction by positive displacement, transgression” (Derrida and Houdebine 1973: 35).

236 

 Rika Preiser, Paul Cilliers & Oliver Human

the notion of critique: the constant movement of bringing together and pulling apart, that allows for some legitimate (albeit temporal) position from where critical practices can emerge.

3.4 The economy of critique as stricture Drawing on the earlier discussion concerning the relation between the restricted and general economy, it is obvious that “critique as stricture” also operates within the currency of the general economy. From this deconstructed interpretation of the concept of critique, “critique as stricture” changes our critical engagement with complex phenomena to such an extent that it can be described as a mode of critical practice that can be employed to negotiate the labyrinth of different and often opposing epistemologies. Critique as stricture offers us the possibility to find some legitimisation in the fact that it surpasses the reified positions as found in the binary logic of seemingly opposing positions. Critique as stricture poses a restorative critical practice, which allows for new and alternative ways of negotiating complex realities. The role of critique is thus changed from providing a measure or standard against which to judge norms and notions of truth, to that of becoming a resisting force that undermines and frustrates totalising projects and calls in question the binary oppositions through which knowledge about reality is mastered. Moreover, critique as stricture cannot be an absolute negative critique in the sense that it just resists any kind of affirmation blindly. It aligns itself to some positive affirmations whilst simultaneously knowing that the moment of affirmation also implies the transgression thereof. Its scepticism towards a blind faith in the power of reason and totalising discourses that suggest a unity of science or reification of knowledge on the one hand is mirrored by the same measure of scepticism towards total negativity that amounts to an irresponsible kind of nihilism or relativism. By being mindful of the fact that the in-betwix is a constellation of clustered and juxtaposed differences and irreducible propositions and principles, our way of thinking is challenged to change in such a way that it resists total reduction and total reconciliation simultaneously. In the name of this simulated kind of thinking, the affirmation of a critique that has the capability to rupture and reconcile at once is found. Moreover, as a theory that exposes the limits of scientific methods and the implications of reductionist methodologies, the grammar of critique as stricture not only coincides with the grammar of the critical project but also transcends it in the same way that deconstruction transcends classification as being a critical theory. Hence, the act of criticising not only results in the bridging of opposing



Deconstruction and complexity 

 237

paradigms, but simultaneously disrupts and ruptures unified ideas and totalitarian dogmatism. As an intervention, critique as stricture becomes a relentless undercurrent of resistance that never rests. In a more existential sense, the general economy of critique as stricture can be understood to simultaneously mean a “critique–of” (being a strategy or mode of questioning) as well as to mean “critique–as” (becoming an attitude or strategy of thinking). It is this double understanding of critique that allows us a kind of thinking that is more thoughtful to the questions of what it means to be human in a complex world. As will be discussed here below, this kind of thinking can be found in the notion of critical complexity thinking – a mode of thinking that is informed by the logic of the general economy.

4 The economy of critical complexity thinking By inscribing the generative movement of the Derridean double bind into both the concepts “critique” and “complexity”, a synergy is formed that amounts to the notion of “critical complexity”. Influenced by this double reading of Derrida and Morin, it is possible to interpret the idea of “critical complex thinking” as the embodiment of both différance and the double bind. It alludes to a kind of thinking that takes place in the force field where the tension between differences are upheld, brought together and kept apart at the same time. Considering that the first meaning of the word complexity stems from the Latin noun complexus, which means “what is woven together” (Morin 2007: 6), the notion of critical complexity thinking can supplement the new deconstructed meaning of the concept of critique as stricture. The way in which the weaving together should be understood is demonstrated by Morin (2008) and Cilliers (1998, 2007, 2010) when they speak of “complex thinking”. Derived from a complex systems understanding of how meaning and our interpretation of reality arises, the logic of critical complex thinking proposes a type of thinking that necessitates a double movement similar to what Derrida calls the double bind. It suggests that the concept and its counterpart (the yes and the no) are thought simultaneously. Morin (2007) calls this the “logical core of complexity”, which is dialogical and economical in nature. However, the art lies not in thinking one in terms of the other in binary motion, but in terms of how the one is dependent and determined by the other. The knack lies not in describing opposites when making knowledge claims, but in thinking both at the same time. It is described as a “dialogic [that] is not the response to these paradoxes, but the means of

238 

 Rika Preiser, Paul Cilliers & Oliver Human

facing them, by considering the productive play of complementary antagonisms” (Morin 2007: 20). Not only does the economy of critical complexity thinking join seemingly unrelated paradigms of thought, but it simultaneously also represents a gap between different traditions. From this double bind of the gap and the bringing together of seemingly contradictory traditions, a rupture emerges. From this rupture, in turn, the possibility of interruption is born where a liminal space is opened in which all orientations are challenged to critically negotiate new ways of thinking and knowing. The economy of the critical complexity approach takes on a dynamic character that has the ability to remain alive whilst operating in the tension of the paradoxical stricture. As a mode of thinking it becomes a relentless undercurrent of interruption that never rests. The interruption is characterised by the recognition of the limitations that each different orientation of thought has to offer. In the conceptual opening that is generated by the performative tension in which the logic of critical complexity negotiates between crude reductionism and totalising holism or stringent forms of rule-making and interpretation, a crack emerges from where dichotomies are destabilised and exposed. From this position of interruption a gap opens up that offers us a way through discourses that pronounce “the end of science” (Horgan 1996) and philosophy (Badiou 2006). We are not left at some dead end where all scientific and philosophic endeavours or the nature of reality is just a construction of the human mind with no relation to an outside. An understanding of the relatedness of the different orientations affords us the possibility to engage in philosophic practice that amounts to an endeavour that is neither absolute, nor eternal, but open to excess, innovation and creativity. Complex thinking and its critical imperative allow us to be more thoughtful and open to the call of questions relating to injustice, exploitation and different forms of oppression. Hidden in the heart of critical complex thinking is the possibility of thinking differently about what it means to be human in a world that is changing more rapidly than we can control.

4.1 Towards an ethics of critical complexity In this paper we have argued that Derrida’s work can contribute towards a deeper understanding of complex systems. Part of the argument was to show that when dealing with complexity, we never escape the realm of choice and that consequently, there are always – implicitly or explicitly – normative issues involved. In the final part of the paper we hope to make some first steps in developing an ethics of complexity that can give some content to this normative dimension.



Deconstruction and complexity 

 239

The argument for the inevitability of an ethical position when dealing with complexity can be made in the following way: since we cannot have complete knowledge of complex things we cannot “calculate” their behaviour in any deterministic way. We have to interpret and evaluate. Our decisions always involve an element of choice that cannot be justified objectively. What is more, no matter how careful our actions are considered, they may turn out to have been a mistake. Thus, acknowledging that values and choice are involved does not provide any guarantee that good will come of what we do. Critical complexity tells us that ethics will be involved, but does not tell us what that ethics actually entails. The ethics of complexity is thus radically or perpetually ethical. There is no a priori principle we can follow, nor utility we can compute. We do not escape the realm of choice. The problem with developing an “ethics of complexity” is, of course, nothing new that one could ascribe to be unique to the study of complexity.6 There are many different forms of postmetaphysical theories that deal with the problem of the loss of the outside or the so-called “death of the transcendental subject” that makes it impossible to find a legitimate point from outside our systems of knowing from where to justify what norms are acceptable and which are objectionable. Thus, the dilemma that all poststructural ethical theories face is the problem that there is no a priori principle that can be followed that could guarantee that the norms and criteria we choose will produce ethical behaviour or that justice might flow from it. This problematic is expressed clearly in Fraser’s (2008: 406) critique of poststructural theories of justice when she states that the so-called “loss of the outside” amounts to the fact that “the basic parameters of justice are contested” and that it results in a lack of “authoritative standards for assessing the merits of justice claims”. Stated in different terms, allegiance to a poststructural position implies that there is no transcendental Archimedic point that can provide us with something to hold unto. The implications of trying to find legitimate justifications for adopting ethical standards after having denounced all objective or transcendental positions is summed up poignantly by Praeg (2010: 259) in his discussion of the hegemonic powers that are at work in developing countries and how they position themselves in a globally changing world of power relations: There is no ‘outside’ of/to a global complex network. This means that the signifiers hitherto used in order to invent and legitimize such systems of differentiation from the outside –

6 See Woermann and Cilliers (2012) for a more detailed discussion on the idea of an “ethics of complexity”.

240 

 Rika Preiser, Paul Cilliers & Oliver Human

God, time, autonomy, evolution, rationality, Freedom, Democracy and so forth – have and will increasingly become drawn into the system [...] Contorted remnants of previous legitimations, they will increasingly give rise to a symptomatic relativism: whose rationality? Whose concept of freedom? Whose notion of democracy? Whose freedom fighter and whose terrorist? When this occurs we will have moved, in the terms of analysis offered here, from a politics to an ethics. More precisely, we will have witnessed the end of (modernist) politics and a return of/to the ethical.

What is highlighted in Praeg’s quote is the fact that the ethical is not to be found in the moment of organised politics or in any moralising strategies. Moreover, there is a change in focus in terms of when ethics enters into the picture. This view aptly describes the way in which the ethical turn takes place in the formulation of the notion of critical complexity: when meaning emerges relationally in a system of differences, the ethical moment is born when we have reached the limits of our computing or equalling out strategies. When we know which decisions or strategies lead to what results or outcomes, we don’t need “ethics”, but moral codes of conduct protocols or best-practice manuals. The ethical moment can thus be redefined as being situated in the moment in which we take the leap from that which is known to that which is uncertain or unknown. Derrida (1999: 280) calls this leap into the unknown the moment of “undecidability” and he describes it as follows: I have to prepare a decision to know where I can go as well, as consciously as possible, but I should acknowledge that between the accumulation of knowledge and the moment I make a choice, I take a responsibility, I make a decision, there is an infinite abyss because of the heterogeneity of these moments. That is why I so constantly insist on the undecidability, which does not mean you are simply paralysed and neutralised because you don’t know what to do. Simply, in order for a decision to be a decision it has to go through a moment when irrespective of what you know, you make a leap into the decision. This leap into the responsibility is an infinite one and you take a decision only in a situation when there is something undecidable, when you don’t know what to do. You don’t know. That is, if you knew what to do, there would be no decision, you would have already done [...]

What is very clear from the above is the fact that the ethical moment is born once we enter into the gap of the infinite abyss that is created by the limits of our models. Being face to face with the limit forces a change in terms of how we define the notion of ethics. Subsequently, it is suggested that critical complexity thinking is informed by a kind of normative turn that emerges from an understanding of ethics that is defined in terms of its situatedness of being “in the gap” where it is faced with the unknown – or stated differently – the limit. Moreover, such an understanding of ethics calls the moral agent not to choose between either a transcendental position or a pragmatic one. Rather, it requires from the agent to enter into the ambiguity of the both/and logic of the general economy. This argu-



Deconstruction and complexity 

 241

ment is echoed by Baudrillard (1994) when he argues that poststructural notions of ethics can only stand a chance of achieving any form of credibility if it becomes a strategy of radical antagonism, a play upon reality, the issuing of a challenge to the real, an attempt to put the real on the spot. Thus, the movement of stricture and the economy of the double bind remain inescapable also for rethinking the notion of ethics when faced with the limits of what we can know and do.

5 Conclusion The project of critical complexity makes clear that we are always acting from a particular contingent position with a limited amount of resources, in other words, from a particular economy. This economy would be sufficient were complex systems closed and did not contain play. However, things are not that simple and our economies of thought are always exceeded by an excess. It is for this reason that the project of critical complexity can never rest on solid ground; we are always forced to reassess our position. Yet we must act in the world and therefore we are forced to act from the restricted economies we function with. Yet this is the possibility of action, otherwise our actions would not be distinguishable. This is why often the strongest actions, be they good or bad, arise from a very restricted worldview. But what is important to note is that even though we may be successful for the time being, the ever-changing conditions of life mean that this success is always limited in time and space, we will eventually be forced to find other means of reaching our goals. The critical project grants us this insight that possibilities can be found between the limits and in the stricture of the restricted economies we operate from and the general milieu in which these actions rest. These possibilities are brought forth by the cutting up and weaving together of possibilities, like the possibilities found in the weaving together of deconstruction and complexity. However, when we cut and weave we not only create new possibilities but we exclude others. The ethics of complexity rests in this tension between the different economies, between what remains in our economy and what we exclude. The ethics of complexity attempts to guide us in acting in meaningful ways whilst taking into consideration that these actions are meaningful precisely because they close down other possibilities. Our actions should always therefore be seen as provisional (note this does not mean that we cannot be held accountable for our actions, quite to the contrary). This ethics thereby reminds us that there is always the possibility for another process of cutting and weaving, for another engagement with the general economy, in short this ethic reminds us of the

242 

 Rika Preiser, Paul Cilliers & Oliver Human

necessity of alternatives. The acknowledgement of the fact that our economies are always restricted, that play exists within our economies, that there are always more possibilities than we grant (both good and bad) is becoming increasingly important in a world more and more committed to a single restricted economy. In order to make other worlds possible, even to simply imagine an alternative means of existence, it is imperative that we engage with the restless resistance that can be found in the cutting and weaving together of complexity and deconstruction, in the possibilities that this dialogic reveals.

References Alhadeff-Jones, M. 2008. Three generations of complexity theories: nuances and ambiguities. In: Educational Philosophy and Theory 40(1): 66–82. Alhadeff-Jones, M. 2010. Challenging the limits of critique in education through Morin’s paradigm of complexity. In: Studies in Philosophy and Education 29(5): 477–490. Badiou, A. 2006. Infinite Thought. London: Continuum. Bataille, G. 1989. The notion of expenditure. In: Stoekl, A. (ed. and transl.). Visions of Excess: Selected Writings, 1927–1939. Theory and History of Literature, Volume 14. Minneapolis: University of Minnesota Press. Bataille, G. 1991. The Accursed Share, Volume 1. transl. Hurley, R. New York: Zone Books. Bataille, G. 1993. The Accursed Share, Volumes 2 & 3. transl. Hurley, R. New York: Zone Books. Baudrillard, J. 1994. Simulacra and Simulation. Ann Arbor: University of Michigan Press. Benhabib, S. 1986. Critique, Norm, and Utopia: a Study of the Foundations of Critical Theory. New York: Columbia University Press. Bernstein, R. J. 1991. The New Constellation: the Ethical-political Horizons of Modernity/Postmodernity. Cambridge: Polity Press. Chryssostalis, J. & Tuitt, P. 2005. Introduction. In: Law and Critique 16(1): 1. Chu, D., Strand, R. & Fjelland, R. 2003. Theories of complexity: common denominators of complex systems. In: Complexity 8(3): 19–30. Cilliers, P. 1998. Complexity and Postmodernism: Understanding Complex Systems. London: Routledge. Cilliers, P. 2000. Knowledge, complexity and understanding. In: Emergence 2(4): 7–13. Cilliers, P. 2001. Boundaries, hierarchies and networks in complex systems. In: International Journal of Innovation Management 5(2): 135–147. Cilliers, P. 2005. Complexity, deconstruction and relativism. In: Theory Culture and Society 22(5): 255–267. Cilliers, P. (ed.). 2007. Thinking Complexity: Complexity and Philosophy, Volume 1. Mansfield, MA: ISCE Publishing. Cilliers, P. 2010. Difference, identity and complexity. In: Philosophy Today 54(1): 55–65. Collins English Dictionary. 2006. Glasgow: Harper Collins. Derrida, J. 1977. Limited Inc. Evanston, IL: Northwestern University Press. Derrida, J. 1978. Writing and Difference. Chicago: Chicago University Press. Derrida, J. 1982. Margins of Philosophy. Brighton: Harvester Press.



Deconstruction and complexity 

 243

Derrida, J. 1992. Given Time: 1. Counterfeit Money. transl. Kamuf, P. Chicago: University of Chicago Press. Derrida, J. 1998. Resistances of Psychoanalysis. Stanford: Stanford University Press. Derrida, J. 1999. Justice, law and philosophy – an interview with Jacques Derrida by Cilliers, P., van der Merwe, W., Degenaar, J. In: South African Journal of Philosophy 18(3): 279–286. Derrida, J. 2000. Et cetera. In: Royle, N. (ed.). Deconstructions: a User’s Guide. Basingstoke: Palgrave, 282–305. Derrida, J. & Houdebine, J.L. 1973. Positions: interview with Jacques Derrida. In: Diacritics 3(1): 33–46. Fleming, M. 1969. Introduction to Economic Analysis. London: George Allen and Unwin. Fraser, N. 2008. Abnormal justice. In: Critical Inquiry 34 (Spring): 393–422. Gasché, R. 1994. Inventions of Difference: on Jacques Derrida. Cambridge: Harvard University Press. Habermas, J. 1975. Legitimation Crisis. Boston: Beacon Press. Hanssen, B. 2000. Critique of Violence: Between Poststructuralism and Critical Theory. London: Routledge. Hobson, M. 1998. Jacques Derrida: Opening Lines. London: Routledge. Höffe, O. 2004. Kants Kritik der reinen Vernunft: die Grundlegung der modernen Philosophie. München: Verlag C.H. Beck. Horgan, J. 1996. The End of Science: Facing the Limits of Knowledge in the Twilight of the Scientific Age. Reading, MA: Helix Books. Hoy, D.C. 2005. Critical Resistance: from Poststructuralism to Post-Critique. Cambridge, MA: The MIT Press. Kant, I. 1998 [1781]. Critique of Pure Reason. Guyer, P. & Wood, A. (eds. and transl.). Cambridge: Cambridge University Press. Morin, E. 2007. Restricted complexity, general complexity, in: Gershenson, C., Aerts, D. & Edmonds, B. (eds.). Worldviews, Sciences and Us – Philosophy and Complexity. Singapore: World Scientific Publishing, 5–29. Morin, E. 2008. On Complexity. Cresskill, NJ: Hampton Press. Pavlich, G. 2005. Experiencing critique, Law and Critique 16(1): 95–112. Praeg, L. 2010. Africa: globalisation and the ethical. In: Cilliers, P. & Preiser, R. (eds.). Complexity, Difference and Identity: an Ethical Perspective. Dordrecht: Springer, 241–264. Preiser, R. & Cilliers, P. 2010. Unpacking the ethics of complexity: concluding reflections. In: Cilliers, P. & Preiser, R. (eds.). Complexity, Difference and Identity. Dordrecht: Springer, 265–287. Rasch, W. 1991. Theories of complexity, complexities of theory: Habermas, Luhmann, and the study of social systems. German Studies Review 14(1): 65–83. Roudinesco, E. 2008. Philosophy in Turbulent Times. New York: Columbia University Press. Royle, N. (ed.). 2000. Deconstructions: a User’s Guide. Basingstoke: Palgrave. Saussure, F. de. 1974. Course in General Linguistics. London: Fontana. Sloterdijk, P. 1987. Critique of Cynical Reason. transl. Eldred, M. Minneapolis: University of Minnesota Press. Van Niekerk, A.A. 1980. Die Grense van die Kritiese Rede. MA(Phil) thesis. University of Stellenbosch, South Africa. Woermann, M. & Cilliers, P. 2012. The ethics of complexity and the complexity of ethics. In: South African Journal of Philosophy 31(1): 403–419.

Oliver Human & Paul Cilliers

Towards an economy of complexity Derrida, Morin and Bataille

1 Introduction There has been a sustained engagement with complexity and complex systems at least since the second half of the 1980s. In the last decade or so, this engagement has shifted from more traditional and reductionist approaches to approaches which emphasise the contingent nature of complex systems. It is argued that the characteristics of such systems are context dependent and irreducible to their constituent parts. Byrne (2005), Cilliers (1998), Smith and Jenks (2006) and Urry (2005) provide some overview of these developments. In summary, the following general description of complexity is normally given: A complex system is defined by a network of rich interactions which change over time. It is not the number of parts interacting which define complexity but rather the nature of their interactions: these interactions are nonlinear. The fact of nonlinearity has at least two implications. In the first place, since the law of superposition does not hold, a complex set of nonlinear relationships cannot be reduced to a simpler set which is equivalent. This is referred to as the “incompressibility’ of complexity (Cilliers 1998: 9–10). Secondly, there is not a linear relationship between cause and effect. Small causes can have large effects and vice versa (for a discussion on nonlinearity see Borgo & Goguen 2005; Knyazeva, 2004). Furthermore, complex systems are usually open systems, i.e. they interact with their environment. Instead of simply being a characteristic of the system itself, the extent of the system is also determined by the purpose of the description of the system, and is thus influenced by the position of the observer (Cilliers, 1998: 4). This is an important aspect of complex systems. The boundaries which we draw between the system and its environment are often a product of the description we use of such systems. What we exclude from analysis has just as strong an impact on our understanding of the system as that which is included (for a discussion of the radical openness of complex systems see below and also Chu et al. 2003; for a discussion of issues involved in observing complex systems see Rasch 1991). Originally published in Theory, Culture and Society, 2013, 30(5): 24–44. DOI: 10.1177/0263276413484070, © 2013 SAGE Publications.

246 

 Oliver Human & Paul Cilliers

An important consequence of these characteristics is that our knowledge of complex systems is always context dependent, determined by the models we construct of such systems (see Byrne 2005). Strong reductionism, where the aim is to reduce the complexity faced to a neat, comprehensive model or algorithm, becomes impossible. We cannot know complex systems completely (Cilliers 2002, 2005b). This implies that the possibilities which a model holds will always be the product of the frames we apply because we cannot comprehensively determine the limits of the system and hence accurately define the field of possibilities. The use of words such as “models”, “exclusion”, and “partial description” by no means implies that we can contrast this with the “complete” or “comprehensive” view of the “system itself” which we may have access to outside of the problems of modelling.1 Modelling is not only a limitation of epistemology; it is also an ontological problem. A “pure” description or “full” presence will always elude us, mainly due to the ontological characteristics of complex systems described above. Indeed, the problems of complexity arise precisely because of the inevitable contingency of our existence. In this paper we are trying to develop a means through which we can explain the simultaneously constrained and excessive nature of complex systems. We try to develop the notion of economy as a means to understand Morin’s (2007: 11) contention that a system is both more and less than the sum of its parts. In doing this we aim to explore and forge some links between recent work in complexity theory as exemplified by Edgar Morin (1992, 2007), and the work of Jacques Derrida (1977, 1978, 1982, 1992, 2004 [1972]) and Georges Bataille (1989, 1991, 1993). It will be shown that the complex thought of Edgar Morin shows striking similarities to elements of deconstruction, and that insights from Derrida can be used to augment our understanding of complexity.2 The key notion which will be examined is that of an “economy”. It is important to note here that when we use the term “economy” we are not referring to the discipline of “economics” or “economic systems” per se. Instead we use the term as it is used by Bataille (adopted

1 In short, when we use the term “model” or “modelling” we use it as a shorthand to describe any understanding of the world we express. In this regard, a theory, explanation and indeed art or music are all “models” which we use to understand complex phenomena. 2 For further discussion of the links that can be made between complexity and deconstruction see the work of Cilliers in general (especially Cilliers 1998, 2005b). In contrast to others who have adopted post-structuralist philosophy as an approach to dealing with complexity, in this paper we have adopted the approach of Derrida. Despite the “style” and “language” of philosophers such as Deleuze which lends itself to complexity, in this paper we find Derrida’s thoughts on the relationship between restriction and excess, presence and absence, especially helpful in thinking about complexity.



Towards an economy of complexity 

 247

from Hegel) and Derrida, in its philosophical sense, as developed by this tradition of critical thinkers (see below). We will attempt to answer the following questions: Why is the notion necessary when we talk of systems? How does one understand the difference between “open” and “closed” systems and between “general” and “restricted” economies? Are these notions related and what do they imply for the way in which we engage with complex phenomena? In thinking about complex phenomena, the notion of a “frame” has often been used as a tool to describe how we exclude certain phenomena from our view in order for us to develop a model of the phenomenon we are exploring. However, as will become clear below, we consider the notion of a frame too limiting as it fails to take into consideration the various sets of relationships which are established in the process of modelling. These relationships include both the sets of relationships inside the model as well as the sets of relationships established with that which is excluded from the model. The metaphor of a general economy, as read by Derrida, helps us out of this conceptual blind spot. The argument in this paper moves through five stages. First, we give a general introduction to the notion of an economy. We try to give a brief outline of how we use the term and which aspects of the term pertain to modelling complex systems. Following this we provide an overview as to the difficulties of modelling complex systems. We try to describe the sets of relationships which are established in the process of modelling, both those which occur inside the model and those established with that which is excluded from the model. Thirdly, we look at the distinction Edgar Morin establishes between “restricted” and “general” complexity. In this discussion we illustrate how it is impossible to reduce complex phenomena to simply the sum of their parts but rather argue for the fact that complex systems are always excessive and thereby incomplete. Finally, we illustrate how the view of complexity we have established here has affinities with the notion of a general economy as described by Bataille and Derrida. In this regard we propose that the notion of a “general economy” as understood by Derrida is a useful aid to assist us in thinking about complex systems. In the conclusion, we aim to illustrate some of the benefits of adopting the term economy in relation to thinking about complexity.

2 What is an Economy? The notion of “economy” evokes a certain set of meanings. Primarily these include usages of the term in relation to the study of economics. The dictionary definition of an “economy” is twofold. In the first place, it refers to a concern with

248 

 Oliver Human & Paul Cilliers

the production and consumption of resources, but it also indicates the orderly interplay between the parts of a system (Collins English Dictionary 2006). In effect, an economy is the concern with the production and consumption of resources made possible by the orderly interplay of the parts of a system. Furthermore, the notion of economy implies limited or scarce resources which constrain not only the amount of things produced within a system but also what is produced by that system (things which are not fixed and which can change over time). A consequence of these constraints is that processes have to be developed within a system which prioritise what should be produced and in what amounts (Flemming 1969). However, these processes, shaped by the orderly interplay between parts, not only determine what should be produced but also what can be produced. An economy then operates on the principle of a relationship of feedback (see below) between the use of limited resources (production and consumption) and what the system is able to do with these resources. This system of production, and the limited resources it exploits, produces a contingent set of priorities which favours some aspects but has to suppress or exclude others. There is a further meaning of the word “economy” which does not necessarily reflect the concern for production and consumption. In this use of the word, as the second part of the dictionary definition illustrates, an economy is more concerned with the limited sets of relationships between parts or the play between parts of a system. An economy in this regard is something internal to a system, not as concerned with the production or consumption of resources as with the constraints placed on the ability of parts to act in certain ways. This sense of the term economy affects our understanding of the general notion. An economy is dependent upon limits or constraints determined by the relationships between the components in the system. It is this, more philosophical, definition of the notion of an economy which we will be using in this paper. The limits of an economy imply that in order for a system to produce something, it must receive in return. Indeed, it would be difficult to conceive of an economy, as an economy, that only distributes or disseminates without return. There must be constraints on how much a system can do. In a certain sense, then, an economy can be conceived of as being conservative. It aims to preserve its structure through the priorities it establishes. The means by which an economy establishes its priorities, and hence preserves its foundations, can be described as the reason of the economy (see below). These limits are not to be understood only in a negative sense. Not only are they enabling (see Cilliers 2001: 139; Juarrero 1999: 132–3), they are necessary for the existence of the system in the first place. Something which has no boundaries, which purports to encompass everything, is indeed nothing.



Towards an economy of complexity 

 249

If one looks at different strategies or systems of thinking as economies, this point becomes clearer. A paradigm of thought requires that the theories or propositions which constitute it retain a certain coherence, a certain allegiance to limits which allows some components but excludes others. As a matter of fact, paradigm shifts are often initiated by the upsetting or disruption of such limits. These economies of thinking exclude that which does not suit the reason guiding the model and hence resist that which may disrupt the foundations and priorities established. For example, the shift which occurred from modernism to postmodernism was a result of the disruption of the taken-for-granted foundation of modernity, namely that of an accurately reducible world, by an economy of analysis which argued for an excess to such a foundation.

3 The “Open” Boundaries of an Economy For certain theorists, complex systems are conceptualised as being open to their environments, but at the same time they are operationally closed (see, for instance, Chu et al. 2003: 28; Cilliers 2005a: 608; Morin 2007: 10; Luhmann 1989). We thus need to explore the idea of limits, that which makes an economy possible, in more detail. Due to the fact that we are always partial and situated observers of the phenomenon we are studying, we inevitably exclude certain aspects which may have a bearing on our analysis from that analysis. What is excluded, as mentioned above, is a result of the process of prioritisation granted to certain facets of the phenomenon we are exploring. Early attempts to deal with complex problems took for granted the fact that this prioritisation was natural and sufficient to explain the phenomenon under consideration. These approaches assumed that what was irrelevant to analysis did not have an essential bearing on the functioning of the system as a whole. This approach can loosely be described as “modernist”. They considered their models of complex systems to be comprehensive since they were based on the essential, underlying properties or structure of such systems usually, but not solely, described in mathematics. This approach could be justified by the atomistic and rationalistic tendency inherent within Western metaphysics (Dreyfus & Dreyfus 1986; Dreyfus 1999). The reductionist rationality assumes that some kind of coherence, some kind of “reason”, is necessary in order for the model to be sensible at all.3 Therefore,

3 It is important to note here that we are not arguing that we must introduce or allow irrationality to dominate the models we use in science. Rather, this argument is simply to illustrate how

250 

 Oliver Human & Paul Cilliers

that which is excessive, that which is excluded, stands outside the particular logic or reason of that epistemology. In order to ensure the rational representation of what is investigated, that which is excluded is often depicted as being “noise” or “inconsequential”, or even to be rotten or defiled in some way. The “other” in this instance is marginalised in order to provide a sense of coherence and order to the model (for a discussion of this strategy in social systems see Laclau 2005: 139–144). Thinking in this instance demands that we exclude the heterogeneous, that we exclude that which challenges the taxonomy of our thought. Rationality will structure and determine the process of exclusion. However, that which is found outside the strict borders of our models can never be seen as a blank space or simply as noise. The process whereby a model is developed does not occur within a neutral, context-free environment. In order to function as models, our models can never contain everything of relevance. There is no a priori way of determining what is relevant and what is not (see Cilliers 2005b: 259). What occurs inside our models cannot be easily separated from what is excluded because what we exclude from our models constitutes them as much as that which is included. Nevertheless, this should not be seen as a weak, postmodern cry for heterogeneity, as a plea for the inclusion of everybody and their best friend into the economy. Such a position is as futile as the complete exclusion of everything heterogeneous. It is by exclusion, the setting of limits, that the economies become useful to us. These limits are productive not only because they are constitutive, but also because they allow the very antagonisms within the models to function. That which is excluded makes possible the debates or differences found inside the system. In science, for example, for a discipline such as botany to exist, for there to be debates within the discipline about the subject matter of that discipline, botanists need to behave and engage in a manner which by necessity excludes other fields from their discussion. In order to be botanists, they have to differentiate themselves from physicists, chemists and social anthropologists. This differentiation nevertheless remains problematic since it is constantly challenged. The nature of the boundary, of what is considered internal or external, is perpetually transformed by the threat of “the outside” since the “threat” simultaneously

rationality thins out, or excludes, the excess of data we are faced with during the process of generating understanding.



Towards an economy of complexity 

 251

structures the “inside”.4 In order to give more content to this “double movement”, a distinction can be made between heterogeneity and difference.5 We will label heterogeneous that which is perceived as noise from the perspective of the system or the model itself. In other words, heterogeneity is that which, again from the perspective of the model, makes no sense or appears to bear no influence on the outcomes we aim to achieve, despite the fact that it has an influence nonetheless. Heterogeneity can be shown to carry weight within the model through a process of analysis which makes use of other frames of analysis. Heterogeneity is thus not noise or a mystical force but simply that which does not make sense from the limited perspective of this model. Following Ernesto Laclau (2005: 141), one can use the example of chess players to describe heterogeneity and difference. Difference makes the game of chess possible, the fact that two players compete with different pieces which have different capabilities in terms of how they move around the board. Difference is both the set of capabilities of the chess pieces as well as the two competing players. Heterogeneity intervenes when another person knocks over the chess board, for instance, or some event disturbs the game. The field on which the competing players were playing is now destroyed or changed. Difference refers to the discriminations which can be made from the perspective of the model under consideration. Differences can be recognised only in terms of a common frame. For example, the difference between a dictionary and a novel occurs within the framework of what we understand as books. The difference between a dictionary and a tree requires that one shift frames of reference considerably in order to make these differences understandable (see Cilliers 2010). To recognise heterogeneous objects one has to establish a chain of different frameworks, on different scales, which will eventually allow some comparison to emerge. Consider the following very simplified example. The notion of tree is heterogeneous to the frame “books”, but the frame “paper” allows us to see the importance of the notion to the frame by which this heterogeneity is defined. Yet, as Bataille (1989: 98) pointed out, the inclusion of the heterogeneous as a difference within an economy necessarily destroys its radical status. As we include the heterogeneous into our models we necessarily homogenise these possibilities within current frameworks at the same time as producing other forms of heterogeneity outside the newly accepted range of differences.

4 For a discussion of the relationship between the notion of “inside” and “outside” see Derrida (2004 [1972]). 5 We borrow these terms from Laclau (2005: 140).

252 

 Oliver Human & Paul Cilliers

There is a contextual dimension to difference and heterogeneity. Due to the adaptability of the system we cannot categorically state that something will remain heterogeneous to a system because, in a different context, it may make sense to that economy. For example, if we were building a lightweight car out of composite materials, the differences within our economies would be concerned with the different pliabilities and strengths of the materials. The colour of these materials would be heterogeneous to our discussion. However, if we were told we were now building this car in a country with a warm climate, the relative colours of the composites, such as black carbon, would begin to function as differences within our economy due to concerns with heat. What was heterogeneous to this economy now becomes a difference within that economy as it fits within the logic of the economy.6 An important consequence of these dynamics is that it is difficult to illustrate the incoherence and excess of any economy clearly since the language we use within models relies only on difference. In other words, the language we use within models recognises difference, not heterogeneity, as a result of the fact that they are built around a concern for coherence. This is the reason why critiques from outside a “system” are often not considered valid critiques at all by proponents working within the system. An excellent example of this is the difficulty encountered in finding a meaningful interaction between “analytic” and “continental” philosophy.

4 From Restricted to General Complexity As argued above, classical science attempted to reduce the problems of complexity to a level at which the unique problems faced in complex systems are not visible. Edgar Morin (1992, 2007) argues that this restriction of the problem of complexity was achieved by relying on three explanatory principles. The first of these principles is that of epistemological determinism, which implies that all future and past events must be known within the present state of a system. The principle of determinism argues that a complex system rests on a neat historical trajectory and, based upon its current state as well as the next state description, we can trace, as well as predict, the shape of the system as it has been and will

6 This does not imply that the economy itself remains exactly the same. Like the meaning of words, with the possibility of including something previously heterogeneous to the economy comes the loss of previous difference(s).



Towards an economy of complexity 

 253

be (see also Dekker 2010). The second explanatory principle adopted is that of reduction. Reduction is the assumption that “consists in knowing any composite from only the knowledge of its basic constituting elements” (Morin 2007: 5). The reductionist argument states that a system consists simply of a sum of its parts, that the higher or emergent properties of a system can be reduced to the characteristics of the parts which determine them. Finally, classical science argues for the explanatory principle of disjunction, which “consists in isolating and separating cognitive difficulties from one another, leading to the separation between disciplines, which have become hermetic from each other” (Morin 2007: 5). One can see that all the explanatory principles of classical science are predicated upon a very similar principle, that a scientist can objectively and comprehensively know what is essential to the functioning and survival of a system. Classical science, therefore, works on the principle of simplification. The complexity of a system is simplified to sets of laws and universal truths which operate as the foundation of science. Morin refers to this strategy as “restricted complexity”. Against the approach of “restricted complexity”, Morin (2007) proposes the concept of “general complexity”. In the paradigm of general complexity the assumptions made by classical science are not taken for granted as simple truths. Although such assumptions are necessary for the process of science, it should be kept in mind that they are assumptions made in order to reduce complexity to a point where practical research becomes possible. The strategy of general complexity is to recognise this dilemma. In opposition to reduction, complexity requires that one tries to comprehend the relations between the whole and the parts. The principle of disjunction, of separation (between objects, between disciplines, between notions, between subject and object of knowledge) should be substituted by a principle that maintains the distinction, but that tries to establish the relation (Morin 2007: 10–11).

General complexity points towards an epistemology of complex systems which examines the relationships between the parts as well as the parts themselves. A strong reductionism thus ceases to be possible as the focus of analysis shifts away from the parts to a consideration of the contingent sets of relationships between the parts. It is this contingency which denies simple and universal models. The best we have are models which are partial and provisional. The simplified models of a modernist science, we argue, only recognise difference, not heterogeneity. A “general” approach to complexity is one which recognises that we have to reduce and constrain, but that the heterogeneous will remain a force which disrupts our provisional reductions. Thus, consciously or not, we make exclusions when we build models, but these exclusions always have an impact on both the system and the model. Some of these exclusions hold

254 

 Oliver Human & Paul Cilliers

the potential to destroy the systematic representation of what we are modelling, but the notion of general complexity is precisely to make us aware of this. The strict disjunctions, separations and exclusions are imposed by us in order to make these systems intelligible to us. In what follows we will propose that the economies of complex systems, along with the models we create of them, should be conceived of as “general” economies. We will argue that, despite the necessity of restricted models, when we are faced with a complex system we are forced to concede that our models are context dependent and therefore always open to chance and error as well as certain types of paradox, even to what may be seen as “irrational”. A restricted economy of analysis, as we will show, is one which does not take heterogeneity into consideration. In contrast, like Morin’s general complexity, the notion of a general economy aims to keep heterogeneity in mind whilst granting us the use value of a restricted economy.

5 Towards an Economy of General Complexity Thus far we have presented the notion of an economy and then illustrated how complex systems can be conceived of as operating within a particular economy. Since we can never have a complete view of a complex system, we are forced to acknowledge that the economy we postulate of a complex system is always the product of the particular viewpoint we adopt. Following Morin, we argued that classical science adopted a restricted view of complexity. However, despite our critique of the reductionism of classical science, we argued that, to a certain extent, this reductionism is necessary in order for us to be able to say anything about complex systems in the first place. The question before us now is the following: How then does this “new” approach to complexity differ from that of classical science? In other words, if we criticise the economy of thought established by classical science yet, at the same time, argue for the necessity of that which we are criticising, what is different about this economy? How can we speak of an economy, as an economy, if we are forced to speak of it as open and closed at the same time? These questions can be tackled by first investigating another perspective on the notion of an economy, one offered by the philosopher George Bataille (1989, 1991, 1993).



Towards an economy of complexity 

 255

5.1 Bataille’s general economy Bataille (1989) argues that traditional political economy restricts its analysis of a system to the production and consumption of resources, limited to the immediate ends they serve. According to restricted economics, “on the whole, any general judgement of social activity implies the principle that all individual effort, in order to be valid, must be reducible to the fundamental necessities of production and conservation” (1989: 117). Restricted economics is therefore utilitarian, only interested in the use value any object or activity may have. The problem with this model, for Bataille, is that it does not take into consideration the “excesses” and so-called “waste” produced by a system. In contrast to the notion of restricted economy, Bataille argued for the notion of “general economy” which aimed to include within its analysis the excesses and waste not considered by a restricted worldview. The “law of general economy” states: On the whole a society always produces more than is necessary for its survival; it has a surplus at its disposal. It is precisely the use it makes of this surplus that determines it: the surplus is the cause of the agitation, of the structural changes and of the entire history of society. But the surplus has more than one outlet, the most common of which is growth (Bataille 1991: 106).

For Bataille, the impact which a surplus has on the nature of a system is not reflected upon by restricted economies of analysis which limit their analysis to notions such as utility and thereby exclude, and are unable to explain, other forces which act upon the system. It is for this reason that Bataille (1989) argues that restricted economic worldviews struggle to explain the occurrence of such phenomena as war, sacrifice or eroticism. These aspects of human life and history remain side-lined and are seen as marginal to the “more important” aspects of survival. Yet it is precisely these marginalised forms which give shape to the societies we live in (Bataille 1991). General economics tries to incorporate these aspects of life which are considered pure expenditure, or “excess”, into its frame of analysis. However, as we have already argued, we always operate from a limited, context dependent position. Bataille in this sense tries to “re-economise” our thinking by attempting to take into consideration the excess of the frames we use when describing economies purely in utilitarian terms. In other words, Bataille tries to include all social activity in his analysis; he aims to overcome the limits of the economies we create due to our situated perspectives. However, as Bennington (1995) illustrates, by focusing on the waste or excess produced by a system, Bataille is structuring his analysis around a single concept (that of excess) in the same vein as the restricted economies he is critiquing:

256 

 Oliver Human & Paul Cilliers

In its most abstract form, this suggestion would say that ‘general economy’ is not the other of ‘restricted economy,’ but is no other than restricted economy; that there is no general economy except as the economy of restricted economy; that general economy is the economy of its own restriction (Bataille, 1991: 47–8, emphasis in original).

The argument here is simple enough: in order to be an economy, it must by definition operate as if it is restricted; an economy cannot contain everything. We argued above that when we model a complex system our analysis will always be restricted, due not only to our limited perspectives but also due to the fact that our models need to deal with the inevitable uncertainty of our existence and still be coherent and logical. We deal with this uncertainty through the use of reason, which Derrida (2005: 151) defined as a wager between the calculable and the incalculable. As such there will always be an excess. Excess, by definition, exceeds reason (Derrida 1978: 255, quoting Bataille). This excess we have labelled heterogeneity.7 What the notion of general economy does is to establish a relationship to this excess (Derrida 1978: 270). In Derrida’s exploration of Bataille we can note the double handed movement of the deconstructive process which aims to maintain the radical nature of Bataille’s critique whilst at the same time illustrating the impossibility of a “pure excess” without an economy to which it corresponds (or “sovereignty” as Bataille labels it). Derrida argues that we need to remember that we can only speak of one economy (of one discourse); it is senseless in this regard to postulate two different kinds of economy, one restricted and the other excessive or general. That is, we cannot postulate an economy of excess which runs parallel to the restricted economy or a general economy in which there is only excess. When we speak of a general economy it is not an economy separate from a restricted economy; rather, it is a single economy which is not closed but is both open to random chance events as well as predictability, open to the possibility of destruction and yet robust, whether it comes from the play of forces inside the system or from its relationship to its environment (Derrida 1978: 272). The models we construct of such systems must keep in mind that a system does not run on an entirely rational, utilitarian basis, but is open to the possibility of

7 It is important to note that Bataille uses the term heterogeneity in a far more radical way than the way we have defined it in this paper. For Bataille (1989: 140), heterogeneity can never be reincorporated into any economy; it is that which stands radically outside all economies. His “science” of heterologie therefore only hopes to examine the effects of heterogeneity rather than their reincorporation into a restricted economy (1989: 97–102). This is because the heterogeneous is a product of an active force for Bataille. In contrast, we argue that although the heterogeneous carries weight for that which excludes it, it doesn’t have the same active energy as Bataille hypothesises.



Towards an economy of complexity 

 257

paradox and inconsistency yet still displays enough stability in order to be comprehended. The notion of general economy describes an economy with open boundaries and also a play of forces inside the system. Such an economy is neither the strictly restricted economy of traditional political economy or of classical science which denies the partiality of any perspective, nor is it an economy of excess (whatever that may look like). The double handed logic of Derrida allows us to conceive of this economy as being limited, constrained and restricted and, at the same time, as being open and excessive. What makes this possible? The notion of “play’ is crucial to this understanding.

5.2 The play of the general economy Under a restricted economy, a single term or centre, a single logic or reason, defines and limits the structure of a system. As Derrida (1978: 278) argues, “the function of this centre was not only to orient, balance and organize the structure – one cannot in fact conceive of an unorganized structure – but above all to make sure that the organizing principle of the structure would limit what we might call the play of the structure’” However, in a complex system, a sort of totalisation by means of proposing a centre is not possible. Due to the impossibility of reducing the system to some essential truth, or algorithm, in the style of restricted complexity, complex systems are “centreless”. Totalization, therefore, is sometimes defined as useless, and sometimes as impossible. This is no doubt due to the fact that there are two ways of conceiving the limit of totalization [...] Totalization can be judged impossible in the classical style: one then refers to the empirical endeavour of either a subject or a finite richness which it can never master. There is too much, more than one can say. But nontotalization can also be determined in another way: no longer from the standpoint of a concept of finitude as relegation to the empirical, but from the standpoint of the concept of play. If totalization no longer has any meaning, it is not because the infiniteness of a field cannot be covered by a finite glance or a finite discourse, but because the nature of the field [...] excludes totalization. This field is in effect that of play, that is to say, a field of infinite substitutions only because it is finite, that is to say, because instead of being an inexhaustible field, as in the classical hypothesis, instead of being too large, there is something missing from it: a centre which arrests and grounds the play of substitutions (Derrida as quoted by Johnson 1993: 51, emphasis in original).

Play thus constitutes another, different form of excess. This is no longer the excess “outside” the system but rather a certain indeterminacy “inside” the system. In the language of complexity, this is due to the nature of the multiple nonlinear interactions and feedback paths within complex systems, as well as to the

258 

 Oliver Human & Paul Cilliers

fact that the boundaries of complex systems are always open. This makes them non-totalisable. Play is a product of the relationships which a system establishes internally as well as with its environment. The notion of excess is not external to a system. Not only is the outside folded into the inside, the internal processes themselves break open gaps as the result of the process of différance (see below). As a result, we cannot reduce this system to a set of preconceived terms or truths, nor imagine that everything which happens within the system is reincorporated for utility within the system (Hegelian Aufhebung). Rather, we must grant, in the same logic as that of Derrida’s notion of iterability, that systems, even though they are robust, exist at a point which is open to change and even destruction, open to chance and a play of forces which we can never comprehensively predict or calculate. It is for this reason that the notion of general economy, even though it is read or understood in restricted economic terms, places our understanding of a system in a space in which we are forced to grant the vulnerability of both the system and our understanding of it. There are forces at play which we can never completely predict or control (Derrida 1978: 270). The excess which escapes our models of systems must pass through two straits. Firstly, we cannot exclude the excesses of the system as if they exist in some mysterious, unknowable form, “outside” of our understanding of anything, as if they exist in their own context, as if we could simply write off that which we cannot model whilst claiming that they have an influence on the system we are studying, even if only as inconsequential “noise” or “chaos”. This is what is implied by the use of the term “heterogeneity”. Heterogeneity can also be found within the play of the system and is not something simply external to the system. We must remember that these forces are defined by and thus exist in a relationship with the model we have constructed. The inaccessible, the unknowable in this regard, is not some mystical force outside of restricted economies which guides their interactions. The unnameable or unknowable is an inevitable product of the limits to our model. Secondly, we cannot assume at the same time that we have mastered these forces, that we can comprehensively model a system and the play of its processes, reduce its contingent existence to a single framework. The notion of a general economy reminds us that our understanding of complex systems in the world must walk this narrow edge. Keeping this in mind, the notion of différance (Derrida, 1982) becomes central to our understanding of complex systems and the economies under which these systems operate. Différance in fact establishes the relationship between a restricted and a general economy. Difference, heterogeneity and noise collapse into the notion of différance. Thus we can never comprehensively define a clear distinction between the general and the restricted. The temporal nature of différance (to defer) implies that what is noise today may be central to our under-



Towards an economy of complexity 

 259

standing of the system tomorrow, as we gain new means of interpretation or new understandings of the system. In this light, there can only be provisional discriminators between noise and structure. However, as the inside and the outside of the system are a product of the boundaries we draw rather than something natural or essential to the system itself (Allen 2000: 80; Cilliers 2001: 141; see also Derrida’s chapter “Plato’s Pharmacy” in Derrida 2004 [1972]), the notions of difference and heterogeneity cannot be neatly distinguished, as it is often the heterogeneity of the context which defines which differences will be important, or seen as differences, inside the model. It is also différance which represents the precarity of the system, the fact that any system is always open to (self-)destruction as it makes place for heterogeneous forces which may challenge the coherency of the system. By drawing boundaries, we create the “space” which allows us to say something about the system. This space is not static but a site of action (Cilliers 1998: 45). It is in this space that we create differences, including the difference between inside and outside, which allows us to create models and indeed to act in the world. Therefore, in the relationship which the general economy establishes between the restricted economy and its excess, one can find the play of différance. What makes it possible to model systems is the fact that différance remains undecided between activity and passivity (Cilliers 1998: 45; Derrida 1982: 8–9). “Pockets of stability” make it possible to contingently model a system, as long as we remember this contingency and the conditions under which it was established (Cilliers 1998: 43). Models have to be reinterpreted constantly and critically re-evaluated when used in different contexts. The process of différance thus creates the possibility for the deployment of a restricted economy in our creation of models whilst at the same time precluding the finality of such a restriction.

6 Conclusion In this paper we have argued that complex systems, and the models we make of them, operate under a particular economy. We have illustrated that classical science operates under a restricted economy of analysis which does not acknowledge a model’s relationship to that which it needs to exclude in order to function. In contrast, we argued for the notion of a general economy in which this excess is acknowledged. However, this does not imply that one can operate from a general economy. As Derrida illustrated of Bataille, we can only operate from a restricted economy. This does not mean that we are arguing for a positivistic reduction of a system to some central economy. The fact that we have to reduce does not imply that these reductions are comprehensive. At the same time, we are not arguing for

260 

 Oliver Human & Paul Cilliers

a relativism in which anything can be constituted as the economy of the system. One cannot privilege either form of economy. We are always dealing with economies which are simultaneously restricted and general. We cannot privilege either pole of this dichotomy, nor can we find a compromise.8 This is not a debilitating position. The robustness of complex systems does cater for the restricted economies upon which models are built. Yet the excess of these models leaves novel possibilities open for the future. In this regard, the term “general economy” is perhaps not the best one to use. An “open”, “folded” or “excessive” economy may be terms which could be used to replace the idea of a general economy. However, these still fall prey to the inside/ outside dichotomy and the assumption that a single, essential economy exists. However, the term “general economy” holds significance for the critical tradition in French philosophy as it marks the point at which Bataille illustrated for us the impossibility of a closed system. We have therefore continued using this term as its history speaks to our current concerns. In this paper we have attempted to illustrate how both Derrida and Morin drew on Bataille’s development of the term, and thus the existence of excess was acknowledged by two thinkers in apparently different domains. In fact, we are led to wonder whether Morin’s (2007) adoption of the terms “general” and “restricted” was not a product of reading Bataille. What the above argument has aimed to illustrate is that, despite the different domains in which Derrida and Morin worked, their work shares many common critiques of rationalist and reductive approaches to a complex world. In this article we have illustrated the connection between Morin and Derrida through the work of George Bataille, and the notion of an economy. This link, to the best of our knowledge, has not been made before. The idea of an economy moves away from the limits of speaking about our knowledge in terms of frames, which rely upon a metaphorical distinction between an inside and outside. “Economy” allows us to speak about the boundaries of a system without conceiving of these boundaries as being the limits to the system, as being the borders of influence to the system. To be an economy means to be open. An economy can be seen as an interface between resources and the use of those resources. In other words, an economy is constituted by sets of relationships rather than individuated components. The structure of the economy is therefore a relational one, yet it maintains enough form in order to be spoken of as “a system”. The term therefore overcomes the conceptual constraints of speaking about systems as both being, and being constituted by, isolated entities. Further-

8 This does not mean that we are arguing for some vague holism either, which still commits the same error as restricted economies.



Towards an economy of complexity 

 261

more, the notion of frame limits what we can say about the play inside a complex system. The frame is only concerned with what we can say about the system and its environment. It conjures up an image of that which happens inside the system not being affected by what happens outside the system. The idea of economy allows us to begin to say something about the nature of the relationships “inside” a complex system whilst acknowledging the openness of a system’s boundaries. The notion maintains that in a complex system everything is simultaneously close to the boundary and embedded (Cilliers 2001: 142). The inside/outside distinction thereby collapses under the aegis of the general economy. In short, the term “economy” is a useful placeholder to describe the “slippery” nature of dealing with complexity. The term “economy” holds five implications for the way we approach complex problems. Firstly, the concept of an economy allows us to think about the “economy” of possibilities which any system, and indeed any model of a system, holds. Under a restricted view of complexity we can see that any model of a complex system will only acknowledge those possibilities which can be reduced to the model itself. The range of these possibilities is defined by the field of differences allowed inside the economy. This closed economy of possibilities gives little leeway for dealing with unforeseen and unexpected occurrences. If we assume that our models of the world capture the essence of the world, prediction should be a perfect science. However, under a general economy of complexity we are forced to grant that the unforeseeable is a necessary or constitutive aspect of modelling complexity. This implies that we are forced to develop measures which can deal with this contingency. This “open” horizon of possibilities implies that the world, and our models of the world, always contain more potentialities for realising novelties than can ever be measured at any time. This does not mean that prediction is impossible, only that prediction is a context specific endeavour. This is both a positive feature of the notion of a general economy, that there are always divergent ways of living possible, and a negative consequence, that the future is never guaranteed, the possibility of war or atrocity always remains. Secondly, the acknowledgement that the economies through which we view the world always contain a certain set of possibilities allows us to develop a definition of conservatism. Under this definition we can begin to see how conservative positions are able to be progressive in appearance, as they maintain a faith in science and the possibilities of the future, yet remain tethered to the possibilities of the present economy. This is because we can argue that a conservative position only acknowledges and aims to forward those possibilities their economy may currently cater for or realise (see Fukuyama 1992, for an example of this). So although these positions may be future orientated they are not necessarily progressive. In contrast then to understandings of conservatism as always being

262 

 Oliver Human & Paul Cilliers

backward looking, we can begin to understand current modes of conservatism which argue for progress (see Bourdieu 1998). The term “economy” then complicates what we see as progressive or conservative positions by disallowing reductionist understandings of these terms. Thirdly, the notions of “difference” and “heterogeneity”, as worked out here, illustrate the difficulties we face in coming to some “common ground” or Habermasian ideal speech situation when faced with other worldviews. We are always limited by the economies we deploy and the discriminations (differences) they harbour. We cannot operate under any other set of circumstances. However, this does not mean that our worldviews are limited. The fourth consequence of the term “economy” is that it allows us to realise the wealth of potentialities any model may contain. The folded or general nature of general economies of complexity forces us to grant that possibilities for novelty do not only rest outside of the economies we deploy. We needn’t wait for an event to reveal some other possibility for existence (see Badiou 2007) nor for some interaction with a radical “other”. The play of general economies means that there is always a wealth of possibilities unrealised in the present, inside the very economies we deploy on a daily basis. The challenge is to begin conceptualising ways of realising these novelties in the present without them being consumed, or incorporated as differences, within the economies of the present. Finally, the excess of the general economy of complexity implies a normative dimension to our engagement with the world. We can only engage with a complex system by means of a restricted economy, yet both our economy and the economy we are dealing with are always open. We cannot thereby comprehensively calculate both the effects of our actions as well as the reasons for acting in this way. The ethics of complexity rests in the tension between the different economies, between what we include in the restricted economy which constitutes our models and what we exclude. The normative dimension rests in the fact that we have to choose what we include in the economy, but there is no objective means of doing so. We cannot escape this normative dimension when dealing with complexity as we are forced to make exclusions in order to maintain the coherence of our economies. Our actions are meaningful precisely because they close down other possibilities. This position demands that we take a radically critical stance when engaging with complexity, the main component of which is self-critique (Preiser et al. 2013). Because the economies we are using to analyse the world are always open to forces we cannot account for, we are forced to reconsider our position when facing complexity. Self-critique does not imply that we undermine our stance in the world; it implies that we open this stance up to other alternatives. The idea of a general economy reminds us of both the ubiquity and the necessity of alternatives. This reminder is as much an acknowledgement of the inherent



Towards an economy of complexity 

 263

normativity of complexity as it is a demand. We have to acknowledge this excess in order to begin imagining alternative means of engaging with the problems the world faces. There is a wealth of possibilities which goes unacknowledged by restricted or utilitarian economies which we have to begin to realise in order to improve living conditions on this planet. Acknowledgement: We would like to thank Rika Preiser for her invaluable insight and discussions of the various earlier drafts of this paper. We would also like to thank the reviewers of TCS for their insightful and patient comments on earlier drafts.

References Allen, P. 2000. Knowledge, ignorance and learning. In: Emergence 2(4): 78–103. Badiou, A. 2007. Being and Event. London: Continuum. Bataille, G. 1989. The notion of expenditure. In: Stoekl, A. (ed.). Visions of Excess: Selected Writings, 1927–1939 (Theory and History of literature, Vol. 14). Minneapolis: University of Minnesota Press. Bataille, G. 1991. The Accursed Share, Vol. 1. transl. Hurley, R. New York: Zone Books. Bataille, G. 1993. The Accursed Share, Vols 2 and 3. transl. Hurley, R. New York: Zone Books. Bennington, G. 1995. Introduction: Economics 1: Because the world is round. In: Bailey Gill C (ed.). Bataille: Writing the Sacred. London: Routledge. Borgo, D. & Goguen, J. 2005. Rivers of consciousness: The nonlinear dynamics of free jazz. In: Fisher, L. (ed.). Jazz Research Year Book (Proceedings of IAJE Conference, Long Beach, CA, 5–8 January 2004). Bourdieu, P. 1998. A reasoned utopia and economic fatalism. In: New Left Review 227(January/ February): 125–130. Byrne, D. 2005. Complexity, configurations and cases. Theory, Culture & Society 22(5): 95–111. Chu, D., Strand, R. & Fjellan, R. 2003. Theories of complexity: Common denominators of complex systems. In: Complexity 8(3): 19–30. Cilliers, P. 1998. Complexity and Postmodernism: Understanding Complex Systems. London: Routledge. Cilliers, P. 2001. Boundaries, hierarchies and networks in complex systems. In: International Journal of Innovation Management 5(2): 135–147. Cilliers, P. 2002. Why we cannot know complex things completely. In: Emergence 4(1/2): 77–84. Cilliers, P. 2005a. Knowledge, limits and boundaries. In: Futures 37: 605–613. Cilliers, P. 2005b. Complexity, deconstruction and relativism. In: Theory, Culture & Society 22(5): 255–267. Cilliers, P. 2010. Difference, identity and complexity. IN: Philosophy Today (Spring): 55–65. Collins English Dictionary. 2006. Glasgow: Harper Collins. Dekker, S. 2010. Newton’s bastards. Unpublished paper presented at CSC reading group in Stellenbosch, South Africa. Derrida, J. 1977. Limited Inc. Evanston, IL: Northwestern University Press. Derrida, J. 1978. Writing and Difference. transl. Bass, A. London: Routledge. Derrida, J. 1982. Margins of Philosophy. transl. Bass, A. Chicago: University of Chicago Press.

264 

 Oliver Human & Paul Cilliers

Derrida, J. 1992. Given Time: 1. Counterfeit Money. transl. Kamuf, P. Chicago: University of Chicago Press. Derrida, J. 2004 [1972]. Dissemination. London: Continuum. Derrida, J. 2005. Rogues: Two Essays on Reason. transl. Brault, P.A. & Naas, M. Stanford: Stanford University Press. Dreyfus, H.L. 1999. What Computers Still Can’t Do: A Critique of Artificial Reason. Cambridge, MA: MIT Press. Dreyfus, H.L. & Dreyfus, S.E. 1986. Mind over Machine: The Power of Human Intuition and Expertise in the Era of the Computer. New York: The Free Press. Flemming, M. 1969. Introduction to Economic Analysis. London: George Allen & Unwin. Fukuyama, F. 1992. The End of History and the Last Man. New York: Penguin. Johnson, C. 1993. System and Writing in the Philosophy of Jacques Derrida. Cambridge: Cambridge University Press. Juarrero, A. 1999. Dynamics in Action: Intentional Behaviour as a Complex System. Cambridge, MA: MIT Press. Knyazeva, H. 2004. The complex nonlinear thinking: Edgar Morin’s demand of reform of thinking and the contribution of synergetics. In: World Futures 60: 389–405. Laclau, E. 2005. On Populist Reason. London: Verso. Luhmann, N. 1989. Ecological Communication. Chicago: University of Chicago Press. Morin, E. 1992. Method: Towards a Study of Humankind. Volume 1: The Nature of Nature. trans. Bélanger, J.L.R. New York: Peter Lang. Morin, E. 2007. Restricted complexity, general complexity. In: Gershenson, C., Aerts, D. & Edmonds, B. eds. Worldviews, Science and Us: Philosophy and Complexity. London: World Scientific Publishing, 5–29. Preiser, R., Cilliers, P. & Human, O. 2013. Deconstruction and complexity: A critical economy. South African Journal of Philosophy (forthcoming). Rasch, W. 1991. Theories of complexity, complexities of theory: Habermas, Luhmann and the study of social systems. In: German Studies Review 14(1): 65–83. Smith, J. & Jenks, C. 2006. Qualitative Complexity: Ecology, Cognitive Processes and the Ee-emergence of Structure in Post-humanist Social Theory. London: Routledge. Urry, J. 2005. The complexity turn. In: Theory, Culture & Society 22(1): 1–14.

Minka Woermann & Paul Cilliers

The ethics of complexity and the complexity of ethics 1 Introduction The term “complexity” is often loosely appropriated to describe things that lack simple explanations. More specifically, the paradigm of complexity offers a challenge to traditional reductive explanations, which are premised on the assumption that complex systems can be completely understood in terms of their component parts. If we cannot know phenomena in their full complexity, then an engagement with complexity thinking implies a critical engagement with the status and limits of our knowledge claims. However, the challenge posed by complexity thinking moves beyond this general epistemological level, to influence the whole system of thought that defines our specific thinking on matters related to our practices, politics and ethics (Morin 2008). It is particularly this latter issue – i.e. ethics – that is of interest in the context of this paper. When we utilise a complexity perspective in our thinking about the world, we are busy with a task that is both descriptive and normative in nature. As soon as we engage with complexity, we have to make certain modelling choices when describing phenomena. In other words, since we cannot have complete knowledge of complex things, we cannot “calculate” their behaviour in any deterministic fashion. We have to interpret and evaluate. Our decisions always involve an element of choice that cannot be justified objectively, but are, in part, based on normative judgements. Otherwise stated, our modelling choices are based on subjective judgements about what matters  – both in terms of our work and in terms of our personal lives. This introduces an unavoidable ethical component into our thinking about complex phenomena (Preiser & Cilliers 2010; Cilliers, de Villiers & Roodt 2002; Derrida 1988). In this regard, ethics should be understood as something that constitutes both our knowledge and us, rather than as a normative system that dictates right action. Hence, the ethics of complexity is not an add-on, but inherent to any real engagement with complex phenomena. Otherwise stated, the ethics of complexity is a structural element of complexity thinking. In practice, this means that we should assume a critical attitude when modelling phenomena, where the critical attitude amounts to both the recognition of, and engagement with, the limits of knowledge (Preiser & Cilliers 2010). Originally published in the South African Journal of Philosophy, 2012, 31(2): 448–463. DOI: 10.1080/02580136.2012.10751787 © South African Journal of Philosophy.

266 

 Minka Woermann & Paul Cilliers

In this paper, we investigate, and offer ways in which to deal with the challenges that arise from an engagement with the ethics of complexity. The most important implication in this regard concerns the possibility of a substantive ethics. In simple terms, this amounts to the fact that, although complexity thinking necessarily involves ethics, it cannot provide any information regarding the content of such an ethics since our sense of right and wrong, good and bad, and what deserves respect and what does not, cannot be justified on a priori grounds. Moreover, such a position implies that any substantive notion of ethics must itself be subjected to a deconstruction of sorts, since our ethical models are limited and, hence, exclusionary (Cilliers 2005; Derrida 2005, 2002a, 2002b, 1999). As such the logic which informs the ethics of complexity commits us to accepting the complexity of ethics. Although our position prevents us from giving a substantive account of ethics, we argue that the critical position that we develop nevertheless constitutes a type of ethical strategy, similar to Immanuel Kant’s (1993) categorical imperative, which urges us to adopt a certain attitude when undertaking ethical decisions. However, unlike Kant, we argue for the provisionality of the ethical imperative, and further show that such an imperative is served by the mechanisms of provisionality, transgressivity, irony and the imagination (Preiser & Cilliers 2010). Although these mechanisms can help us to remain sensitive to the complexities that define the contexts in which we operate, there can be no guarantee that ethical actions will ensue. On the final count, the complexity of ethics (which is so intimately interlinked with the rich diversity of what it means to be human) can only be fostered through nurturing trusting relations and through an active recognition of, and engagement with, difference. This paper therefore concludes with a discussion on the importance of trust as a virtue to be cultivated within a complex world.

2 U  nderstanding complexity and the importance of the critical enterprise In order to better understand the claim that the ethics of complexity is a structural element of complexity thinking, it is useful to follow Edgar Morin (2007) in distinguishing between two perspectives on complexity, namely restricted complexity and general complexity. The central difference between these two paradigms concerns how we view the status of our practices. According to Morin (2007), the goal of a restricted approach to complexity is to study the multiple, interrelated processes that constitute complex systems, in order to retroactively uncover the rules or laws of complexity. This approach is



The ethics of complexity and the complexity of ethics 

 267

popular among the researchers at the Sana Fe Institute (which was founded in 1984, and which is dedicated to the study of complex systems), and much of the work conducted at the Institute is dedicated to discovering, comprehending and communicating the common fundamental principles in a variety of systems, which underlie many of the pressing problems currently facing science and society. Whilst research conducted at the Institute has no doubt led to important advances in formalisation and modelling, Jack Cowan (in Horgan 1995: 104), one of the Institute’s founders, notes that the major discovery to have emerged from the Institute is the insight that “it’s very hard to do science on complex systems”, if by science one understands the process of discovering and modelling the rules and laws that govern the behaviour of all phenomena. Such a view is informed by, what John Horgan (1995) calls, a seductive syllogism, which is based on the premises that since a computer that follows a simple set of mathematical rules can give rise to extremely complicated patterns, and since extremely complicated patterns exhibit in the world, simple rule must also give rise to complicated worldly phenomena. In their implicit acceptance of this syllogism, many researchers at the Institute, thus adhere to a restricted approach to complexity, in that it is believed that, with enough time and effort, we will be able to construct a unified theory of complexity – also referred to as the “Theory of Complexity” (TOC) or the “Theory of Everything” (TOE) (Chu, Strand & Fjelland 2003). In other words, the hope is that complex phenomena can be encapsulated in a precise definition or mathematical equation. We support a notion of general complexity and argue that it is impossible to construe a strict science of complex systems, if by science one understands the practice of uncovering the rules and laws that govern all phenomena. Although we cannot conclusively state that complexity is an ontological category of the world as opposed to merely a consequence of our epistemological limitations, this does not imply that we can relegate complexity to the status of a mere practical problem (which can ultimately be solved with enough computing power). One cannot simply “cut up” complex systems in order to understand them, since what is of interest is the dynamic, local interrelations that exist between the parts of a complex system, and which give rise to emergent phenomena (which are often not reducible to base laws). In this process, contingency – expressed in terms of both intra-systemic and extra-systemic conditions (Wimsatt 2008) – also plays a crucial role, which further frustrates any efforts to merely calculate the resultant effect of a certain configuration of parts. Therefore, in terms of general complexity, any attempt at formulating a TOC will necessarily fail because complexity itself is not accounted for (Morin 2007). Complex phenomena are irreducible, or, to elaborate in the words of the theoretical biologist, Robert Rosen (1985: 424), a system is complex precisely “to

268 

 Minka Woermann & Paul Cilliers

the extent that it admits non-equivalent encodings; encodings which cannot be reduced to one another”. Moreover, we argue that this view commits one to a self-critical rationality, which is defined as a rationality that makes no claim for objectivity, or for any special status for the grounds from which the claim is made. A self-critical rationality is therefore the outcome of acknowledging the irreducible nature of complexity. The fact that an engagement with complexity is not a purely objective exercise, does not however imply an “anything goes” approach. Indeed, as Peter Allen (2000: 93) states, “[a] representation or model with no assumptions whatsoever is clearly simply subjective reality” and therefore “does not concern systemic knowledge”. Acknowledging the irreducible nature of complexity means that we should “apply our ‘complexity reduction’ assumptions honestly” (94), rather than accept the defeatist attitude that limited knowledge commits us to relativism. In this context, intellectual honesty implies modesty, which denotes sensitivity to the levels and limits of our understanding. In other words, we must still be competent at performing the necessary calculations and considering the relevant information, but we should also recognise that doing the groundwork won’t resolve the complexity and that we still remain responsible for our modelling choices, since each choice gives rise to “a different spectrum of possible consequences, different successes and failures, and different strengths and weaknesses” (102). Knowledge acquisition is not the objective pursuit of truth, but rather a process of working towards finding suitable strategies for dealing with complex phenomena. However, since there is no final model, and since knowledge is a tumultuous building site, Morin (2007:21) argues that we must introduce a double conscience into our practices: “a conscience of itself and an ethical conscience”. The ethics of complexity is an expression of this double conscience, since it implies an acknowledgement of the implications that a general view of complexity holds for both the status of our models and for our attitude towards these models, and compels us to remain perpetually vigilant in the face of uncertainty. Vigilance demands a continual and critical evaluation and transformation of our claims and practices, and, in this context, David Wood’s (1999: 117) description of a Derridean notion of responsibility applies to our understanding of the ethics of complexity, which “is not quantifiably (or even inquantifiably) large and, therefore, not a basis of guilt through failure to live up to it. It is rather a recursive modality, an always renewable openness”. This renewable openness is safeguarded by a self-critical rationality, and the ethics of complexity therefore commits us to not just a general understanding, but also a critical understanding of complexity.



The ethics of complexity and the complexity of ethics 

 269

3 The (im)possibility of ethics The previous section can be summarised as follows: a critical notion of complexity implies a radical or recursive understanding of ethics, which amounts to the insight that we cannot do away with ethical considerations because we can neither compute knowledge nor appeal to a priori principles to justify our knowledge claims. Every claim implies a choice. Furthermore, since our models must necessarily exclude some of the complexity, we are responsible for our choices, and we exercise this responsibility through critique, and moreover, self-critique. However, the ethics of complexity cannot do more than generate awareness of the fact that we are always in trouble. In other words, it cannot provide substantive guidelines for an ethical system. The question that now arises is whether we can move beyond this position, in order to say something more about the complexity of ethics. We argue that, despite not being able to provide a substantive ethics, it is possible to develop a type of meta-ethical position, which serves to highlight important considerations that underscore the ethical strategies that we employ when engaging in the particularities of situations. As a starting point, we turn to perhaps the most famous example of a meta-ethical position in the history of moral philosophy, namely Kant’s (1993) categorical imperative. The categorical imperative is a substantively-empty rule, in that it cannot generate contingent ethical principles, but can merely act as a yardstick for evaluating the morality of principles which already exist. This is because Kant wants his moral rule to be categorically applicable and, hence, universally valid. However, the only rule which conforms to this criterion is a purely abstract and formal rule, which says “always follow only universal rules” or, otherwise stated, “always follow only rules which you will want all other people to follow”. Thus, by combining a purely formal rule with the notion of universability, Kant can generate a formulation that actually does say something about ethics, namely that if certain contingent principles are universalisable, then the principles are deemed morally correct. Therefore, although the categorical imperative cannot indicate which principles are good, right and deserving of respect, it does provide a strategy for evaluating our contingent principles. As such, one can argue that Kant’s categorical imperative urges us to adopt a certain strategy when undertaking moral considerations (Preiser & Cilliers 2010). Next, we can try to apply the same logic Kant uses to the ethics of complexity, in order to say something about the complexity of ethics (in other words, in order to develop a meta-ethical position). From the analysis thus far we can construct the following argument: all knowledge (including self-knowledge) is limited because, in order to generate meaning, we need to reduce the complexity through modelling. Our models are radically contingent in time and space because they

270 

 Minka Woermann & Paul Cilliers

are the product of the resources at our disposal, the choices that we make, and the influences that act upon us (including the influences of others). Since all knowledge is contingent, it is also subject to revision, and therefore irreducibly provisional. Following Kantian logic, we can now capture the gist of the above argument in the following imperative: “When acting, always remain cognisant of other ways of acting”. Our meta-ethical position thus constitutes a provisional imperative (Preiser & Cilliers 2010). Note that on one reading, the idea of a provisional imperative is a contradiction in terms, since the logic of an imperative is absolute: either you follow the imperative or you don’t. The idea of a provisional imperative seems to suggest that the imperative itself is subject to change, and in this regard we seem to be advocating an impossible position. This is, to a large extent, exactly the point: we cannot do away with moral imperatives, but, if we take complexity seriously, we should also realise that our imperatives are the outcome of certain framing strategies or ways of thinking about the world, and are thus necessarily exclusionary. Thus the provisional imperative stipulates that we must be guided by the imperative, whilst simultaneously acknowledging the exclusionary nature of all imperatives. In terms of the actual content of the imperative, it must be noted that – unlike the Kantian imperative  – which tells us something about the rules for action, the provisional imperative says something about our state of mind or attitude when choosing rules for action. Again: it is impossible to say that “When acting, always choose rules that admit to the possibility of other rules”, since the logic of rules (as with the logic of imperatives) is absolute. In this regard, Jacque Derrida (1988: 116) notes that: Every concept that lays claim to any rigor whatsoever implies the alternative of ‘all or nothing’ [...] Even the concept of ‘difference to degree,’ the concept of relativity is, qua concept, determined according to the logic of all or nothing, of yes or no: differences of degree or no differences of degree. It is impossible or illegitimate to form a philosophical concept outside the logic of all or nothing.

In the above citation, Derrida is pointing to structural conditions of all concepts. We cannot do other than model and exclude. Yet, what the provisional imperative tells us is that when we act, we must be cognisant of this logic. In this regard, we argue that it makes a difference – and moreover, an ethical difference – whether one exercises this awareness. This is because if we remain open to other ways of modelling and other ways of being, we are more likely to practice a self-critical rationality, to respect diversity, to be willing to revise our models, and to guard against the naturalisation of these models. The provisional



The ethics of complexity and the complexity of ethics 

 271

imperative, therefore, provides us with a strategy for remaining open to complexity at the same time that we reduce complexity through our decisions and actions.

4 Four operations in service of the critical position Practically, it is very demanding to commit to a recursive or deconstructive view of ethics. Acknowledging that we are always in trouble can be daunting; and, in this regard, it is helpful to take note of four mechanisms that serve to reinforce and promote the critical attitude, namely provisionality, transgressivity, irony and imagination.

4.1 Provisionality Although the recursive or deconstructive view of ethics developed above necessarily implies provisionality, it is helpful to unpack the meaning of provisionality in more detail, in order to shed further light on the nature of our imperative. As stated above, provisionality is the outcome of the contingent nature of our knowledge claims. The source of this contingency is two-fold: firstly, the meaning of our claims is dependent on the context in which they are made. Language is iterable, which means that although concepts are repeatable and understandable across contexts, the meaning of the concept shifts every time the concepts are used; or, in the words of Mikhail Bakhtin (1984: 202): The life of the word is contained in its transfer [...] from one context to another context [...] In this process the word does not forget its own path and cannot completely free itself from the power of these concrete contexts into which it has entered.

This, as Derrida (1979: 81) notes, implies that “no meaning can be determined out of context”. A good illustration of this concerns how we understand the term “freedom”, used as the title of Jonathan Franzen’s (2010) recent book, compared to its use in the title of Mandela’s (1994) autobiography, A Long Walk to Freedom. In both counts, the term “freedom” is familiar, but in the former use it denotes a critical appraisal of contemporary American society; whereas in the latter use, the term is associated with the liberation struggle, and the story of Mandela’s own imprisonment. Secondly – and more radically – meaning is contingent because even within a given context we cannot fix meaning, since, as Derrida (1979:81) argues, “no context permits saturation”. Due to the complexities involved, every context is

272 

 Minka Woermann & Paul Cilliers

open to further description, and meanings change as the interpretation of the context changes. With regard to our example, one can argue that our understanding of “freedom” in Mandela’s autobiography is dependent on our own personal background. Not only does the understanding of the concept vary from one person to the next, but the very same book can never be reread in exactly the same way: what Mandela’s bibliography, and the importance of freedom meant to me ten years ago, will differ from what it will mean if I were to reread the book today. Provisionality therefore draws attention to both the spatial and the temporal dimensions of meaning (and by extension, of ethics). What the above implies is that “[t]here are no final meanings that arrest the movement of signification” (Culler 1983: 188). There is no universal origin that we can somehow access through applying ourselves. Unlike Plato (1987), who argues that the ideal form of the Good is the ultimate object of knowledge, a complex understanding of ethics posits the good as something which is necessarily subject to revision and deconstruction; and, hence as something which is provisional. Although pockets of consensus and relative stability are needed for the designation of right and wrong, it is important not to naturalise our ethical positions or arrest the movement of signification in the name of Truth. This is especially significant given the nature of today’s geopolitics in which Western ideals all too often pass as universal ideals. In this regard, it is useful to recall Derrida’s (2002c: 10) view of philosophy as something which is “no more assigned to its origin or by its origin, than it is simply, spontaneously, or abstractly cosmopolitical or universal”. He continues in arguing that: There are other ways for philosophy than those of appropriation or expropriation [...] Not only are there other ways for philosophy, but philosophy, if there is any such thing, is the other way.

In our context, we can substitute the term “philosophy” with “ethics”, since what lies at the heart of the provisional imperative is the belief that ethics is indeed the other way; or, more poignantly, the way which is still to come.

4.2 Transgressivity Preiser and Cilliers (2010) write that the critical position informed by complexity will have to be transgressive. It can never simply re-enforce that which is current, but – as the definition states – involves a violation of accepted or imposed boundaries. In this regard, transgressivity demands bold action. It can be argued that, on a literal level, transgressivity is at odds with modesty, which – as discussed



The ethics of complexity and the complexity of ethics 

 273

earlier  – also underscores the critical position. However, in order to practise transgressivity responsibly, one must be modest enough to recognise the limitations of one’s conceptual schema, and show a willingness to overcome these limitations. Modesty and transgressivity thus go hand in hand, since modesty acts as the impetus for transgressivity, in focusing attention on the possibility of other rules of action (as commanded by the provisional imperative). Moreover, being transgressive is not only an ethical move, but also a political move. It involves recognition of the importance of remaining vigilant and open to diversity and to the future, whilst simultaneously exercising choice and taking in a position. Derrida (2002d: 29) calls this the aporia of politics and ethics, which exhibits in the deconstructive nature of negotiation, defined as “a work of mediation [...] a to-and-fro between impatience and patience”. Transgressivity demands absolute engagement with both ethics and politics, since both are concerned with the here and the now, and require a thoughtful and urgent answer to the question “What should I do?” (Derrida 2002b: 296, 302). Transgressivity, like deconstruction, is therefore situated in the fold of the aporia between affirmation (ethics) and position (politics) (Derrida 2002d: 25). Furthermore, transgressivity is what gives ethics potence, since, as Derrida argues (25), “[a]ffirmation requires a position. It requires that one move to action and do something, even if it is imperfect”. We need to act, even though we know we cannot get it right (Preiser & Cilliers 2010: 270). The double logic of ethics and politics thus marks the heart of the critical position. Alain Badiou (2009) also hints at the importance of transgressivity in his moving tribute to Derrida. In this tribute, he describes Derrida as the opposite of a hunter, because unlike the hunter – who hopes that the animal will arrest its movements so that it can be shot – Derrida’s animal cannot cease fleeing. This is because locating the animal does not mean grasping; rather, it means suppressing it. This metaphor is intended to explain Derrida’s passion for doing justice to what Badiou terms the “non-existent”. Framing or modelling phenomena necessarily implies a reduction of complexity. In making choices, we leave out certain considerations from our models and, in the social realm, these considerations may include the interests of stakeholders, who – in terms of our model – become the non-existent or those who are not accounted for in our conceptual paradigms. However, the moment we try to do justice to the non-existent by accounting for their interests, we assimilate the alterity of the Other into our system of thought. In other words, the outside becomes inside. In terms of Badiou’s metaphor, the animal is the Other – to whom we must do justice, even though we cannot know the Other in her alterity. The endless flight, or ethics as a recursive modality, is therefore served by transgressivity.

274 

 Minka Woermann & Paul Cilliers

Furthermore, non-existence sits within the aporia of ethics and politics. On the one hand, to say that the “the non-existent is” fails to convey that it does not exist – at least not in terms of any explicit conceptual paradigm; and, on the other hand, to say that “the non-existent does not exist” fails to convey that it is (141) (i.e. that despite not being heard, the interests of unseen stakeholders, for example, continue to exist). Non-existence therefore simultaneously necessitates vigilance of the fact that not all the complexities of a situation are accounted for, and proactively attempts to give voice to the disenfranchised (which, for Badiou [141], is typified in the war cry of the Revolution, which reads “We are nothing, let us be all!”). Transgressivity, as an attempt to do homage to non-existence, is therefore fed by the complexity of ethics, but leads to the binaries of politics. Lastly, being transgressive implies that we acknowledge the fact that there is no way in which we can fully engage with the excess of meaning that results from complexity; that our context (i.e. our position) is defined by certain freedom and constraints, which acts as the necessary conditions for action and transformation; and, that we have to acknowledge and exercise choice. These three acknowledgements are interdependent: complexity is not the result of indeterminacy. Rather, as Derrida (1999: 79) states, the problem is that “there is too much determinacy”. There is no mystical basis for complexity. Complexity is only the result of the complex interrelations between the components of the system, and of our attempts to model this complexity (which, as Luhmann [2000] notes, creates further complexities due to the fact that our models are imperfect renditions of complex systems). As a result, Derrida (1999: 79) writes that “there are many meanings struggling with one another, there are tensions, there are over determinations, there are equivocations”. Paradoxically, it is these over determinations that generate freedom, since exercising choice and assuming a position is possible because of the constraints within which we operate. In this regard, Morin (2008: 113) writes that “the complex notion of self-organization permits us to conceive of beings that are relatively autonomous as beings while remaining subject to the necessities and hazards of existence”. Complexity is generated by both dependence and freedom, and in order to understand the nature of choice, these concepts must be thought simultaneously. Transgressivity does not imply an abandonment of everything that came before. Rather, in order to responsibly transgress current systems of meaning, we need to concede to the inextricable ways in which our lives are constituted by the systems of meaning in which we partake, whilst nevertheless remaining vigilant of the fact that we have both the duty to continually break open and transform these systems in order to account for the non-existent, as well as the duty to take responsibility for our positions – even when they have undesirable and unforeseen consequences. As Derrida (1981: 12) explains:



The ethics of complexity and the complexity of ethics 

 275

There is not a transgression, if one understands by that a pure and simple landing into a beyond of metaphysics [...] Now, even in aggressions or transgressions, we are consorting with a code to which metaphysics is tied irreducibly, such that every transgressive gesture reencloses us – precisely by giving us a hold on the closure of metaphysics – within this closure. But, by means of the work done on one side and the other of the limit the field inside is modified and a transgression is produced that consequently is nowhere present as a fait accompli. One is never installed within transgression, one never lives elsewhere. Transgression implies that the limit is always at work.

4.3 Irony Transgressivity is supported by an ironic outlook on life. Consider dictionary descriptions of irony as a means of expressing something other than the literal intention of words, as a demonstration of incongruence between what is expected and what is, and as an expression of that which is contrary to plan or expectation. In these descriptions, irony is defined as a means by which to subvert the idea of an objective reality. This is achieved by introducing an element of contingency and play into literal, objective language. In formulating the provisional imperative, we stated that the logic of a rule is absolute: either you follow the rule or not. Yet, the value of irony is that it draws attention to the supplementary complications that govern all rules and that point to the impossibility of concluding any general theory that rules give rise to (Derrida 1988). As such, irony affirms the necessity of improvising when faced with the limits of a binary logic. It is in this regard that irony can be interpreted as a generative creative task, akin to a form of improvisation, and one can further argue that there is “an important and potentially fruitful connection” between these skills and “the lived experience of complexity” (Montuori 2003: 238). Alfonso Montuori (238) elaborates in saying that “improvisation and creativity are capacities we would do well to develop in an increasingly unpredictable, complex, and at times chaotic experience”. It is specifically in relation to developing fruitful and responsible strategies for living that the ironic dimension of the critical task is indispensable. The theme of irony is not new to philosophy. Indeed, one of the earliest uses of irony is demonstrated in Plato’s dialogues, where the Socrates’ rhetorical technique “was [used] to pretend ignorance and, more sneakily, to feign credence in your opponent’s power of thought, in order to tie him in knots” (Williams 2003). Furthermore, the German Philosopher, Friedrich Schlegel, popularised the notion of romantic or philosophical irony, which he employed as a “more complex philosophical tool”, used to shed light on the divided self and a multiplicity of perspectives that could potentially unlock the truth of the whole (Williams 2003). The

276 

 Minka Woermann & Paul Cilliers

concept of irony is also prevalent in the work of Hegel, Kierkegaard, Nietzsche, and  – more recently  – Richard Rorty, the latter arguing that “irony is the only possible ethic of modern liberalism” (Colebrook 2004: 151). Despite all these wellknown accounts of irony in the extant literature, we turn to a figure that is not generally recognised as an ironist in the philosophical literature, namely Claude Levi-Strauss’s bricoleur, in order to further our analysis. A bricoleur is defined as a fiddler or tinker, and, by extension, as someone who makes creative and resourceful use of whatever materials are at hand. LeviStraus (1966) distinguishes the bricoleur from the engineer, arguing that the former approximates the savage mind that operates in a closed universe and is therefore forced to do with the means at hand, whereas the latter approximates the scientific mind that operates in an open universe because he can develop new tools, and thereby construct the totality of the universe. In “Structure, sign, and play”, Derrida (1978) deconstructs the distinction between the engineer and the bricoleur, by arguing that because there is no absolute origin (i.e. no objective reality that we can access through our practices), the engineer remains a theological idea; or, more radically, “the odds are that the engineer is a myth produced by the bricoleur” (285). Derrida’s point is that the world is a complex place, and pretending that it is otherwise is also merely a bricolage or strategy for dealing with the world. As such, Derrida (285) writes that: [...] as soon as we admit that every finite discourse is bound by a certain bricolage and that the engineer and the scientist are also species of bricoleurs, then the very idea of bricolage is menaced and the difference in which it took on its meaning breaks down.

There can be no a priori basis from which to argue for the merits of one life strategy over another. However, it is important to assume responsibility for, and bear the consequences of, our choices and decisions. This is only possible if we are aware of the nature and status of our strategies, which Morin (2008: 96) refers to as “the art of working with uncertainty”. Irony is therefore a way of moving beyond a binary logic and of expressing the double movement involved in affirming a certain position or life strategy, whilst simultaneously undermining the absolutist status of that which we affirm through our lives. We are all improvisers who not only tell a story, but become a story. We create interwoven narratives, which together, constitute a tapestry of stories (Montuori 2003; also see Kearney 1988). The ethical moment lies in whether we concede to this or not, i.e. whether we accept – with irony and humour – our limited knowledge and fragile personal experiences, and focus these in the very moment that we are living in (Montuori 2003). To be able to improvise and to live with irony: requires a different discipline, a different way of organising our thoughts and actions. It requires, and at



The ethics of complexity and the complexity of ethics 

 277

best elicits, a social virtuosity which reflects our state of mind, our perceptions of who we are, and a willingness to take risks, to let go of the safety of the readymade, the already written, and to think, create, and “write” on the spot (244). Assuming an ironic disposition simultaneously draws attention to the status of our strategies, and lightens the burden of self-awareness. This is because those who live with irony find it easier to confess to the fact that their lives are not following a determinate course, but represent the outcome of choices and decisions. Irony – like transgressivity – needs modesty, which in this context means adopting a self-deprecating humour, and not taking oneself or one’s ideas too seriously, as this may prevent one from exercising the openness and tolerance1 needed to act responsibly in the face of complexity. Irony is a critical task, without which we potentially open the door to human evil. This is because, as Susan Sontag (2007: 227) suggests, it is exactly “this refusal of an extended awareness” (which she defines as taking in “more than is happening right now, right here”) that lies at the heart of “our ever-confused awareness of evil” and “of the immense capacity of human beings to commit evil”. In a sense then, it is irony that allows us to face up to the seriousness of our responsibilities, which is an insight which accords beautifully with the description of irony as a demonstration of incongruence between what is expected and what is.

4.4 Imagination Whereas irony assists us in forging strategies for successfully engaging within the constraints that characterise the present, imagination allows us to undertake the creative leap necessary to engage with a future that we cannot calculate. No one can contest the urgent need to move towards a more sustainable future, and in this regard, the role of the imaginative dimension of the critical task is crucial. This is because, as Allen (2000: 103) notes, “creativity” – and we argue, specifically imagination – “is the motor of change, and the hidden dynamic that underlies the rise and fall of civilizations, peoples, and regions, and evolution both encourages and feeds on invention”.

1 The notion of tolerance is increasingly viewed in a negative light in the extant literature, and is often understood as the passive acceptance of difference, and, hence, denotes an unwillingness to engage in other perspectives. Indeed, Rainer Forst (2007) goes as far as to suggest that tolerance could be translated as a kind of insult to difference, and that “recognition” is a better term for designating the positive qualities traditionally associated with the notion of tolerance. In the context of this paper, the notion of tolerance should be read in a positive light, as both respect for, and an engagement with, human diversity.

278 

 Minka Woermann & Paul Cilliers

In this respect, imagination constitutes the ability to generate variety and options, and to break out of one’s “closed or limited hermeneutic circles” (Verstraeten 2000: ix). Indeed, without imagination it is impossible to practise the provisional imperative, which commands us to take cognisance of other rules and other ways of being. In this regard, it is useful to distinguish between “requisite diversity”, which denotes the minimal level of variety needed for a system to cope with its environment, and “excess diversity”, which allows systems to experiment internally and thereby generate a number of strategies for operating in a given environment. Allen (2001) argues that excess diversity is needed for long-term systems survival, since the “fat” of excess knowledge and diversity is needed both for breaking out of our conceptual schema and for imagining, and thereby experimenting and innovating for the future. Todd May (1995: 145) notes that “[t]he terms in which one thinks of oneself and one’s possibilities, the practical parameters of those possibilities, and the ease or difficulty in realizing them, are all social as well as individual matters”. Therefore, imagination – defined as the generation of excess variety – is both a psychological and a social process. On the psychological level, we can develop our imagination by engaging in the arts, which – far from being a pleasurable diversion – is an important way in which to break out of our hermeneutic circles, or – otherwise stated – to “transform the framework we apply when apprehending the world” (Cilliers 2005: 264). This is because, as Morin (1999) notes, the creative arts foster awareness of human complexity, and draw attention to the full range of human subjectivity. As an example, Morin argues that fictional criminals  – such as the gangster kings of Shakespeare, the royal gangsters of films noirs, Jean Valjean and Raskolnikov – are portrayed in all their fullness in literature and film, rather than the least or worst part of themselves (as is often the case with real life criminals). Morin also uses the example of the movie tramp, Charlie Chaplin, in order to illustrate how films use psychological techniques of projection and identification, which brings us to understand and sympathise with people that we would normally find foreign or disgusting. As such, books and films help us “to learn the greatest lesson of life: compassion and true understanding for the humiliated in their suffering” (53). To this we add that the value of the arts lie not only in understanding human complexity, but also the complexities that define our situatedness in the world. As an example, consider Edward Burtynsky’s (2006) photos of manufactured landscapes, which document humanity’s impact on the world, and which persuaded millions to join a global conversation on sustainability. In this regard, imagination can help us to foster not only social sustainability, but also to think about what environmental sustainability might mean. Imagination – like irony – is therefore not just about seeing other ways of being, but also about creating new ways of being.



The ethics of complexity and the complexity of ethics 

 279

On the social level, we follow Timothy Hargrave (2009: 87) in defining moral imagination as a process, which “emerges through dialectical processes that are influenced by actors’ relative power and political skill”. Again, moral imagination is a skill that needs to be fostered and exercised within “pluralistic processes in which multiple actors with opposing moral viewpoints interact, and [where] no single actor is in control” (90). An element of conflict is also always present in imaginative activities because of the “lived tensions between contradictory perspectives” (91). Although Hargrave views moral imagination in terms of a collective action model, his analysis of moral imagination also has implications for individuals. In this regard, Hargrave (91) argues that “morally imaginative actors recognise and integrate contradictory moral viewpoints, and also integrate moral sensitivity [...] [of] contextual considerations”. Since these characteristics are also hallmarks of critical thought, one can argue that moral imagination is itself a critical activity (Woermann 2010). Another characteristic of moral imagination is that it involves an element of uncertainty or risk. Far from being a form of creative abandonment, moral imagination necessitates that we critically project and plan for the future (Woermann 2010). However, since this future cannot be known, and since uncertainty involves a real property of situations, we have to respond with judgement (Luntley 2003). Since the future is uncertain we should be tolerant of each other’s opinions, and also tolerant of failure. Edmund Husserl (in Mensch 2003: 143) explains that tolerance means that I affirm “his ideals as his, as ideals which I must affirm in him, just as he must affirm my ideals – not, indeed, as his ideals of life but as the ideals of my being and life”. James Mensch (2003) explains that, in Latin, tolerance has the sense of supporting or sustaining, rather than enduring or suffering. He further states that “it can be understood as the attitude that actively sustains the maximum number of compatible possibilities of being human” (142). As such, tolerance should be understood as the ideal of human fullness. The reason why Mensch views human fullness as an ideal is because it demands more than can be achieved by a single individual (for example, I cannot simultaneously realise the possibility of being a professional weightlifter and a sprinter). Therefore, for Mensch, tolerance “appears when we acknowledge our finitude in attempting to embody this ideal” (142–143), as well as when we recognise the uniqueness and singularity of human beings. Tolerance is thus essential in allowing the personal and social imagination to flourish, since without it, it is impossible to generate the excess variety needed to productively engage in one’s environment. From this argument, we can also deduce that tolerance is the acceptance of human complexity, even though this complexity can never be fully understood, but only imagined. In this regard, tolerance is nurtured by the creative arts and flourishes in diverse human societies,

280 

 Minka Woermann & Paul Cilliers

in which freedom and other aspects of disorder are accepted, and in which innovation and creativity blossom in defiance of perspectives that try to frame societies as fixed, homogenous systems (Morin 2008). Imagination and tolerance is therefore that which safeguards us from “the well-managed dystopia of the brave new world” (Cilliers 2005: 264).

5 Practicing the Provisional Imperative The above analysis demonstrates how provisionality, transgressivity, irony and imagination can help us to remain cognisant of other ways of acting (and hence, help us to remain faithful to the demands of the provisional imperative). In this last section, the importance of the provisional imperative for our relations with others is investigated. We begin by referring to an anecdote cited by the cybernetican, Heinz von Foerster (1990): I have a friend who grew up in Marrakech. The house of his family stood on the street that divide [sic] the Jewish and Arabic quarter. As a boy he played with all the others, listened to what they thought and said, and learned of their fundamentally different views. When I asked him once, ‘Who was right’ he said, ‘They are both right.’ ‘But this cannot be,’ I argued from an Aristotelian platform, ‘Only one of them can have the truth!’ ‘The problem is not truth,’ he answered, ‘The problem is trust.’

This anecdote points both to the limitations of a binary logic, and the need to overcome these limitations, in order to foster trust. The emphasis on trust, as opposed to truth is an interesting choice, and if we look at the basic characteristics of trust, we see why it is an important virtue for governing relations in a complex world. In her celebrated paper, entitled “Trust and Antitrust”, Annette Baier (1986) argues that trust creates vulnerability, but that this vulnerability is inevitable, because, following two simple Socratic truths, we need the help of others in creating, and caring for the things that we value, and, therefore, have no choice but to place ourselves in a position where others can harm us. Furthermore, because of the richness of human diversity, it follows that not everyone places value on the same things. In complexity terms, one can say that there are many different ways in which to model our world, and these models are interrelated. This also holds important implications for understanding identity. Our identities are neither a priori nor static. Rather, identity is constituted in a complex network, and must be contextualised as both a temporal process of becoming, and as a point in a nexus of relationships (Cilliers 2010). We act on one another in



The ethics of complexity and the complexity of ethics 

 281

ways that give rise to our personal identities as well as the identities of our social practices, and that leads to the transformation of these identities (Woermann 2010). We are therefore not only vulnerable with regard to the things that we value, but also with regard to our very identities. My state depends on the state of others (Preiser & Cilliers 2010). Here, one must draw a distinction between objects of entrustments and trust relations. Whereas the former implies a measure of discretionary responsibility (B knowing what is entrusted (C) by A), and, in some cases, a form of relative power (where A is dependent in some sense upon B’s goodwill) (Baier 1986); the latter is characterised by tolerance (that B will act in such a way so as to allow A to flourish). Whereas, entrustment implies responsibility for something, trust relations imply a responsibility towards someone (Painter-Morland 2006). We can only exercise our responsibility toward others if we show a fundamental respect for difference, even when our actions reduce difference. The only way in which to exercise this responsibility is to heed to the provisional imperative, and to always remain cognisant of other ways of acting. Of course not all differences are good, and in this regard, the provisional imperative – like the Kantian imperative – can be used as a yardstick to evaluate our and other people’s choices and actions. Since the provisional imperative is substantively empty, it should be supplemented by concrete ethical positions that can inform our notions of concrete morality. However, when these ethical positions lead to the homogenisation of “the different perspectives until everybody thinks, believes, and acts the same way” (von Foerster & Poerksen 2002: 36), they will necessarily fail the demands imposed on us by the provisional imperative and should therefore be morally condemned. Indeed, Derrida (1988: 119) warns against those who wish to simplify at all costs, calling them “dangerous dogmatists and tedious obscurantists”, and Morin (2008: 57) refers to the ravages caused by simplifying visions, arguing that “[m]uch of the suffering of millions of beings resulted from the effects of fragmented and one-dimensional thought”. In practice, the provisional imperative thus leads to action that “enlarges the field of vision, opening up new possibilities and revealing the abundance” (Von Foerster & Poerksen 2002: 36), and in this way, allows us – at minimal – to remain open to the truths of others, and trust that others will remain open to our truths.

References Allen, P.M. 2000. Knowledge, ignorance, and learning. In: Emergence 2(4): 78–103. Allen, P.M. 2001. A complex systems approach to learning, adaptive networks. In: International Journal of Innovation Management 5(2): 149–180.

282 

 Minka Woermann & Paul Cilliers

Badiou, A. 2009. Jacques Derrida. In: Pocket Pantheon. transl. Macey, D. London: Verso, 125–144. Baier, A. 1986. Trust and antitrust. In: Ethics 96: 231–260. Bakhtin, M. 1984. Problems of Dostoevsky’s Poetics. (ed. and transl.) Emerson, C. Minneapolis: Minnesota University Press. Burtynsky, E. 2006. Manufactured Landscapes. A film by J. Baichwal, Canada: Mongrel Media. Cilliers, P. 2005. Complexity, deconstruction and relativism. In: Theory, Culture & Society 22(5): 255–267. Cilliers, P. 2010. Difference, identity, and complexity. In: Philosophy Today, (Spring 2010): 55–65. Cilliers, P., De Villiers, T. & Roodt, V. 2002. The formation of the self. Nietzsche and Complexity. In: South African Journal of Philosophy 21(1): 1–18. Chu, D., Strand, R. & Fjelland, R. 2003. Theories of complexity: common denominators of complex systems. In: Wiley Periodicals 8(3): 19–30. Colebrook, C. 2004. Irony: The New Critical Idiom. London: Routledge. Culler, J. 1983. On Deconstruction: Theory and Criticism after Structuralism. London: Routledge. Derrida, J. 1978. From restricted to general economy: a Hegelianism without reserve. In: Writing and Difference, transl. Bass, A. Chicago: University of Chicago Press, 251–277. Derrida, J. 1979. Living On: Border Lines, transl. Hulbert, J. In: Bloom, H.I. (ed.). Deconstruction Criticism. New York. Continuum, 75–175. Derrida, J. 1981. Positions. In: Positions. transl. Bass, A. Chicago: University of Chicago Press, 37–96. Derrida, J. 1988. Afterword. In: Limited Inc. (ed. Graff, G., transl. Weber, S.) Evanston: Northern Western University Press, 111–160. Derrida, J. 1999. Hospitality, justice and responsibility: a dialogue with Jacques Derrida. In: Kearney, R. & Dooley, M (eds.). Questioning Ethics: Contemporary Debates in Philosophy. London, New York: Routledge, 65–83. Derrida, J. 2002a. As if it were possible, “within such limits”... in Negotiations: Interventions and Interviews, 1971–2001. (ed. and transl.) Rottenberg, E. Stanford: Stanford University Press, 343–370. Derrida, J. 2002b. Ethics and politics today. In: Negotiations: Interventions and Interviews, 1971–2001, (ed. and transl.) Rottenberg, E. Stanford: Stanford University Press, 295–314. Derrida, J. 2002c. The right to philosophy from a cosmopolitical point of view: the example of an international institution. In: Ethics, Institutions, and the Right to Philosophy. (ed. and transl.) Trifonas, P.P. Maryland: Rowman & Littlefield Publishers, 1–18. Derrida, J. 2002d. Negotiations. In: Negotiations: Interventions and Interviews, 1971–2001. (ed. and transl.) Rottenberg, E. Stanford: Stanford University Press, 11–40. Derrida, J. 2005. The principle of hospitality, transl. Thompson, A. In: Parallax 11(1): 6–9. Franzen, J. 2010. Freedom. Toronto: HarperCollins Canada. Forst, R. 2007. “To tolerate means to insult”: toleration, recognition, and emancipation, In: Van den Brink, B. & Owen, D. (eds.). Recognition and Power. Cambridge: Cambridge University Press, 215–237. Hargrave, T.J. 2009. Moral imagination, collective action, and the achievement of moral outcomes. In: Business Ethics Quarterly 19(1): 87–104. Horgan, J. 1995. From complexity to perplexity. In: Scientific America, June: 104–109. Kant, I. 1993. Grounding for the Metaphysics of Morals, 3rd ed. transl. Ellington, J.W. Indianapolis: Hackett.



The ethics of complexity and the complexity of ethics 

 283

Kearney, R. 1988. The Wake of Imagination: Towards a Postmodern Culture. Minneapolis: University of Minnesota Press. Levi-Strauss, C. 1966. The Savage Mind. Chicago: University of Chicago Press. Luhmann, N. 2000. Why does society describe itself as postmodern?. In: Rasch, W. & Wolfe, C. (eds.). Observing Complexity: Systems Theory and Postmodernity. Minneapolis, London: University of Minnesota Press, 35–49. Luntley, M. 2003. Ethics in the face of uncertainty: judgement not rules. In: Business Ethics: A European Review 12(4): 325–333. Mandela, N. 1994. Long Walk to Freedom: The Autobiography of Nelson Mandela. London: Little, Brown. May, T. 1995. The Moral Theory of Poststructuralism. Pennsylvania: Pennsylvania State University Press. Mensch, J.R. 2003. Ethics and Selfhood. New York: State University of New York Press. Montuori, A. 2003. The complexity of improvisation and the improvisation of complexity: social science, art and creativity. In: Human Relations 56(2): 237–255. Morin, E. 1999. Seven complex lessons in education for the future. trans. N. Poller. UNESCO Publication. Morin, E. 2007. Restricted complexity, general complexity. transl. C. Gershenson. In: Gershenson, C., Aerts, D. & Edmonds, B. (eds.). Worldviews, Science and Us: Philosophy and Complexity. Singapore: World Scientific, 5–29. Morin, E. 2008. On Complexity. transl. Kelly, S.M. Cresskill: Hampton Press. Painter-Morland, M. 2006. Redefining accountability as relational responsiveness. In: Journal of Business Ethics 66: 89–98. Plato. 1987. The Republic, 2nd ed., transl. Lee, D. London: Penguin Books. Preiser, R. & Cilliers, P. 2010. Unpacking the ethics of complexity: concluding reflections. In: Cilliers, P. & Preiser, R. (eds.). Complexity, Difference and Identity. Dordrecht: Springer, 265–287. Rosen, R. 1985. Anticipatory Systems: Philosophical, Mathematical and Methodological Foundations. Oxford: Pergamon Press. Sontag, S. 2007. At the same time: the novelist and moral reasoning: the Nadine Gordimer lecture. In: Dilonardo, P. & Jump, A. (eds.). At the Same Time: Essays and Speeches. New York: Farrar, Straus and Giroux, 210–230. Verstraeten, J. (ed.) 2000. Business Ethics: Broadening the Perspective. Leuven: Peeters, xii–xiv and 1–8. Von Foerster, H. 1990. Ethics and second-order cybernetics, Opening address for the international conference, Systems and Family Therapy: Ethics, Epistemology, New Methods, Paris, 4 October 1990. Von Foerster, H. & Poerksen, B. 2002. Understanding Systems: Conversations on Epistemology and Ethics. New York: Kluwer Academic/Plenum Publishers. Williams, Z. 2003. The final irony, The Guardian, Saturday 28 June. Wimsatt, W.C. 2008. Aggregativity: reductive heuristics for finding emergence. In: Bedau, M.A. & Humphreys, P. (eds.). Emergence: Contemporary Readings in Philosophy and Science. Cambridge, MA: The MIT Press, 99–110. Woermann, M. 2010. Corporate identity, responsibility and the ethics of complexity. In: Cilliers, P. & Preiser, R. (eds.). Complexity, Difference and Identity. Dordrecht: Springer, 167–192. Wood, D. 1999. The experience of the ethical. In: Kearney, R. & Dooley, M. London (eds.). Questioning Ethics: Contemporary Debates in Philosophy. New York: Routledge, 105–119.

Author Index Allen, Peter 7, 17, 207, 259, 268, 277, 278 Badiou, Alain 238, 262, 273–274 Bak, Per 63, 70, 129, 200 Bataille, George 13, 230–231, 245–247, 251, 254–256, 259–260 Bauman, Zygmunt 188, 193, 213 Byrne, David 1, 17, 245, 246 Chomsky, Noam 23, 42, 44, 56 Cornell, Drucilla 64, 171, 181, 189, 205 Derrida, Jacques xv, 3–5, 7–9, 12–13, 23, 27–31, 39, 43–44, 47, 50–52, 55–56, 60, 62, 74, 78, 107, 118, 134, 140, 144–149, 153–168, 171, 180, 182, 188–189, 194–195, 198, 201–202, 204–205, 208, 214, 226, 228–232, 235, 237–238, 240, 245–247, 251, 256–260, 263, 265–266, 270–276, 281 Emmeche, Claus 85–88, 141 Fodor, Jerry 39, 43–48 Freud, Sigmund 3, 4, 23, 25–27, 29, 30–32, 37, 39, 44, 50, 59, 130, 135, 197 Gasché, Rudolph 203–205, 234 Habermas, Jürgen 77, 107, 129, 136, 146–147, 156, 158–159, 203–204, 233 Holland, John 1, 17, 57–59, 64, 141 Juarrero, Alicia 80, 89, 109, 248 Kant, Immanuel 33, 61, 171, 183, 188, 233–234, 266, 269–270, 281 de Saussure, Ferdinand 3–5, 23, 27–29, 37, 39, 42–44, 50, 56, 123, 196–198, 225–226, 228

Luhmann, Niklas 18, 90, 101, 111, 249, 274 Lyotard, Jean-Francois 7, 60, 78, 107, 118, 124–130 McClelland, James 34–36, 44–45 Mitchell, Melanie 6 Montuori, Alfonso 275–276 Morin, Edgar 12–13, 16, 18, 194, 198, 226, 228, 237–238, 245–247, 249, 252–254, 260, 265–268, 274, 276, 278, 280–281 Nowotny, Helga 213–214 Parkins, Wendy 11, 214, 220 Pylyshyn, Zenon 39, 43–48 Poli, Roberto 7, 16 Preiser, Rika 13, 14, 262, 263, 265–266, 269–270, 272-273, 281 Pribram, Karl 25, 32 Putnam, Hilary 43, 47 Richardson, Kurt 17, 141, 212 Rosen, Robert xv, 12, 17, 106, 113, 267 Rumelhart, David 34–36, 44–45 Simon, Herbert A 92–93 Taylor, Mark 150, 217, 219 Thrift, Nigel 1, 85 Turing, Alan 23, 33, 37, 42–44, 47, 105 Von Foerster, Heinz 280–281 Wilden, Anthony 24, 200 Woermann, Minka 14, 279, 281 Wood, David 8, 147, 268