Get ebook downloads in full at ebookname.

com

Chaos and Complexity in Psychology The Theory of

Nonlinear Dynamical Systems 1st Edition Stephen J.
Guastello

https://ebookname.com/product/chaos-and-complexity-in-
psychology-the-theory-of-nonlinear-dynamical-systems-1st-
edition-stephen-j-guastello/

OR CLICK BUTTON

DOWNLOAD EBOOK

Explore and download more ebook at https://ebookname.com

Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

Chaos Complexity and Transport Theory and Applications

Cristel Chandre

https://ebookname.com/product/chaos-complexity-and-transport-
theory-and-applications-cristel-chandre/

Nonlinear Science and Complexity Albert C. J. Luo

https://ebookname.com/product/nonlinear-science-and-complexity-
albert-c-j-luo/

Nonlinear Dynamics and Chaos 2nd Edition J. M. T.

Thompson

https://ebookname.com/product/nonlinear-dynamics-and-chaos-2nd-
edition-j-m-t-thompson/

The Karate Mouse Geronimo Stilton 40 Geronimo Stilton

https://ebookname.com/product/the-karate-mouse-geronimo-
stilton-40-geronimo-stilton/
The Basic Aerobatic Manual With Spin and Upset Recovery
Techniques 1st Edition William K. Kershner

https://ebookname.com/product/the-basic-aerobatic-manual-with-
spin-and-upset-recovery-techniques-1st-edition-william-k-
kershner/

Millimeter Wave Communication System 1st Edition Kao-

Cheng Huang

https://ebookname.com/product/millimeter-wave-communication-
system-1st-edition-kao-cheng-huang/

Cultures of Expertise in Global Currency Markets 1st

Edition Leon Wansleben

https://ebookname.com/product/cultures-of-expertise-in-global-
currency-markets-1st-edition-leon-wansleben/

The Last Silent Picture Show William M. Drew

https://ebookname.com/product/the-last-silent-picture-show-
william-m-drew/

Death And Dying End Of Life Controversies 2006 Edition

Information Plus Reference Series Sandra Alters

https://ebookname.com/product/death-and-dying-end-of-life-
controversies-2006-edition-information-plus-reference-series-
sandra-alters/
Weighing goods equality uncertainty and time John
Broome

https://ebookname.com/product/weighing-goods-equality-
uncertainty-and-time-john-broome/
Chaos and Complexity in Psychology

Although many books have discussed methodological advances in nonlinear dynamical sys-
tems (NDS) theory, this volume is unique in its focus on NDS’s role in the development of
psychological theory. After an introductory chapter covering the fundamentals of chaos, com-
plexity, and other nonlinear dynamics, subsequent chapters provide in-depth coverage of each
of the specific topic areas in psychology. A concluding chapter takes stock of the field as a whole,
evaluating important challenges for the immediate future. The chapters are by experts in the
use of NDS in each of their respective areas, including biological, cognitive, developmental,
social, organizational, and clinical psychology. Each chapter provides an in-depth examina-
tion of theoretical foundations and specific applications and a review of relevant methods.
This edited collection represents the state of the art in NDS science across the disciplines of
psychology.

Dr. Stephen J. Guastello is a professor of psychology at Marquette University, Milwaukee,

Wisconsin, where his teaching and research specialize in industrial-organizational psychology,
human factors engineering and ergonomics, psychological testing topics, advanced statistics,
and applications of nonlinear dynamics in psychology and economics. He is the author of three
books and more than 100 journal articles and book chapters. He is the founding Editor in
Chief of the journal Nonlinear Dynamics, Psychology, and Life Sciences and served as president
of the Society for Chaos Theory in Psychology & Life Sciences.

Dr. Matthijs Koopmans is Senior Education Research and Evaluation Officer at the Center for
School and Community Services at the Academy for Educational Development in New York.
His expertise is in the use of advanced statistical modeling to analyze educational data and in the
applicability of nonlinear dynamical systems to educational and developmental phenomena.
He has published in numerous scholarly journals, including Nonlinear Dynamics, Psychology
and Life Sciences, Journal for the Education of Students Placed at Risk, and Evaluation and
Program Planning.

Dr. David Pincus is an assistant professor in the Psychology Department at Chapman Uni-
versity in Orange, California. Dr. Pincus’s teaching, clinical, and research interests focus on
applications of NDS theory to understanding interpersonal processes underlying healthy
exchanges of information. He has published journal articles, book chapters, workbooks, and
videos applying nonlinear dynamics to diverse subjects such as relationship development,
psychotherapy, family systems, pain management, and child weight management. Dr. Pincus
is a licensed psychologist and a member of the American Psychological Association and the
Association for Psychological Science.
Chaos and Complexity in Psychology
THE THEORY OF NONLINEAR DYNAMICAL SYSTEMS

Edited by
Stephen J. Guastello
Marquette University

Matthijs Koopmans
Academy for Educational Development

David Pincus
Chapman University
cambridge university press
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo, Delhi

Cambridge University Press

32 Avenue of the Americas, New York, NY 10013-2473, USA
www.cambridge.org
Information on this title: www.cambridge.org/9780521887267

c Cambridge University Press 2009

This publication is in copyright. Subject to statutory exception

and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without
the written permission of Cambridge University Press.

First published 2009

Printed in the United States of America

A catalog record for this publication is available from the British Library.

Library of Congress Cataloging in Publication Data

Guastello, Stephen J.
Chaos and complexity in psychology : the theory of nonlinear
dynamical systems / Stephen J. Guastello, Matthijs Koopmans, David Pincus.
p. cm.
Includes bibliographical references and index.
ISBN 978-0-521-88726-7 (hardback)
1. Psychology – Philosophy. 2. Chaotic behavior in systems. 3. System theory.
I. Koopmans, Matthijs. II. Pincus, David, 1969– III. Title.

BF38.G79 2009
150.1 1857–dc22 2008014866

ISBN 978-0-521-88726-7 hardback

Cambridge University Press has no responsibility for

the persistence or accuracy of URLs for external or
third-party Internet Web sites referred to in this publication
and does not guarantee that any content on such
Web sites is, or will remain, accurate or appropriate.
Contents

Contributors page xi
Preface xiii

1. Introduction to Nonlinear Dynamics and Complexity 1

Stephen J. Guastello and Larry S. Liebovitch 1
Elephants and Horses 1
General Systems Theories and Paradigms 2
Elements of NDS Functions 4
Attractors and Chaos 6
Bifurcations and Chaos 14
Fractals and Power Laws 16
Self-Organization 23
Catastrophe Theory 28
Computational Methods for NDS 31
Summary 35
References 36

2. Collective Intelligence: Observations and Models 41

William Sulis 41
Introduction 41
Models of Collective Intelligence 43
Observations of Social Insect Behavior 49
Conclusion 65
References 66

3. Neurodynamics and Electrocortical Activity 73

Tullio A. Minelli 73
Introduction 73
Basics of Cortical Activity and EEG 73

v
vi Contents

Digital EEG, Records, and Measure 75

Clinical and Cognitive Mathematical Phenomenology of EEG 77
Artificial Learning and Neural Networks 79
Freeman Cortical Networks and Other Sigmoidal Models 81
Hopfield’s Integrate and Fire Neurons: Networks of Spiking
Neurons 83
The Hodgkin–Huxley Circuit and Other Neurodynamic Models 85
Bifurcation, Chaos, and Stochastic Resonance in Neurodynamics 88
AP Propagation and TMS 92
Cortex Stimulation: Associative Memory and Binding 95
Concluding Remarks and Perspectives 97
Appendix: Spectral Coherence 98
References 99

4. Psychophysics 108
Robert A. M. Gregson 108
Introduction 108
The Ubiquitous Ogive 112
Transients 116
Hysteresis 118
Dimensions in Parallel 119
Extensions to Networks 122
Fast and Slow Dynamics 123
The Existence of Inverses 124
Variations in Response-Time Durations 125
Conclusion 127
References 128

5. Temporal and Spatial Patterns in Perceptual Behavior:

Implications for Dynamical Structure 132
Deborah J. Aks 132
Introduction 132
Shifts to the NDS Paradigm 133
Response Time and Eye Tracking 135
RT, Search, and Data Skew 136
Measuring Temporal Dependencies, Scaling, and Rates of Decay 143
Changing Behavioral Patterns Across Conditions: Previous Findings 148
Intrinsic versus Extrinsic Dynamics in Neural Activity 150
Assessing Image Complexity and Visual Search Dynamics 153
Results for Image Complexity and Visual Search Dynamics 159
General Discussion and Conclusions 163
References 170
Contents vii

6. Embodied and Embedded: The Dynamics of Extracting

Perceptual Visual Invariants 177
Patrice Renaud, Sylvain Chartier, and Guillaume Albert 177
Perceptual Stability and Lability 177
Visual Perception and Constancy of Position 177
Perception of Position Constancy 178
Fractal Perceptual–Motor Responses 180
Dynamics of Action–Perception Cycles 181
Intentional Dynamics and Invariance Extraction 183
Perceptual Constancy and Visual Search: Predictions and Methods 184
Results and Discussion for Perceptual Constancy and Visual Search 192
Conclusion 199
References 201

7. Origins of Order in Cognitive Activity 206

Geoff Hollis, Heidi Kloos, and Guy C. Van Orden 206
Origins of Order in Cognitive Activity 206
Order and Information Processing 207
Emergence of Order in Complex Systems 211
Evidence for Self-Organized Order in Cognition 218
Summary and New Questions 233
References 234

8. Nonlinear Complex Dynamical Systems

in Developmental Psychology 242
Paul van Geert 242
Introduction 242
Dynamical Systems and Explanatory Adequacy 243
Development and the Dynamics of Long-Term Change 248
Action and the Dynamics of Short-Term Change 252
Developmental Phenomena from the NDS Viewpoint 256
Fuzziness, Ambiguity, and the Developmental Construction of
Novelty 267
And Where’s the Brain? 270
Conclusion: The Complex Dynamics of Development 271
References 272

9. Developmental Psychopathology: Maladaptive and

Adaptive Attractors in Children’s Close Relationships 282
Erika S. Lunkenheimer and Thomas J. Dishion 282
Introduction 282
Developmental Psychopathology 283
viii Contents

Application of NDS Theory to Developmental

Psychopathology 287
Relationships as Nonlinear Dynamical Systems 288
NDS Analytic Methods in Developmental Psychopathology 290
New Directions in the Study of Attractors 295
Implications for Intervention Science 298
References 300

10. Psychopathology 307

Wolfgang Tschacher and Uli Junghan 307
Introduction 307
The Dynamical Disease Approach 308
The Complexity Theory Approach 314
Empirical Examination of Perceptional Organization and
Schizophrenia 317
Gestalt Studies, Complexity, and Schizophrenia 324
General Discussion and Conclusions 325
References 330

11. Coherence, Complexity, and Information Flow:

Self-Organizing Processes in Psychotherapy 335
David Pincus 335
The True Nature of Psychotherapy 335
From Roots to Branches: Historical Development
of the Process-Oriented Psychotherapies 337
The Nonlinear Elephant in the Office: The Common NDS
Assumptions of Therapy Process 350
Empirical Validation of the 5-R Model 356
Related Empirical Results Using the NDS Perspective 357
Summary and Conclusions 361
References 364

12. The Dynamics of Human Experience: Fundamentals

of Dynamical Social Psychology 370
Robin R. Vallacher and Andrzej Nowak 370
Introduction 370
The Dynamical Perspective 372
Dynamical Minimalism 381
The Dynamics of Social Influence 383
Dynamics of Interpersonal Synchronization 388
The Trajectory of Dynamical Social Psychology 393
References 395
Contents ix

13. Group Dynamics: Adaptation, Coordination,

and the Emergence of Leaders 402
Stephen J. Guastello 402
Introduction 402
Adaptive Behavior 403
Coordination Intersection Type 407
Coordination: Stag Hunt Type 414
Self-Organization and Leadership Emergence 418
Summary 425
References 427

14. Organizational Psychology 434

Kevin J. Dooley 434
Introduction 434
Why Do We Need Complexity Science? 435
How Has Complexity Science Been Used to Study Organizations? 440
Predictions About Future Developments in Organizational
Science 443
Summary 445
References 446

15. Complexity, Evolution, and Organizational Behavior 452

Peter Allen 452
Introduction 452
Evolutionary Drive 454
Ignorance and Learning 455
Complexity and Evolutionary Drive 459
Modeling Human Systems 462
Conclusions 470
References 472

16. Agent-Based Modeling Within a Dynamic Network 475

Terrill L. Frantz and Kathleen M. Carley 475
Introduction 475
History of Social Simulation and Network Analysis 477
Agent-Based Simulation as a Research Strategy 482
Applied Agent-Based Simulations 483
Designing an Agent-Based Dynamic Network Simulation 485
Simulation Software Platforms and Tools 494
Limitations 496
Implications for Researchers 497
References 498
x Contents

17. Epilogue: Psychology at the Edge of Chaos 506

Matthijs Koopmans 506
Introduction 506
Psychology as a Sleeping Beauty 507
The Roles of NDS in Psychology 509
Context, Evolution, and Growing Points 518
Future Challenges 522
References 524

Index 527
Contributors

Deborah J. Aks, Ph.D. Kevin J. Dooley, Ph.D.

Center for Cognitive Science (RuCCS) Department of Supply Chain
Rutgers University – New Brunswick Management
Piscataway, New Jersey Arizona State University
Tempe, Arizona
Guillaume Albert, Ph.D.
Department of Educational Psychology Terrill L. Frantz, Ed.D.
and Psychology School of Computer Science
University of Quebec at Outaouais Carnegie Mellon University
Gatineau, Quebec, Canada Pittsburgh, Pennsylvania
Robert A. M. Gregson, Ph.D., D.Sc.
Peter Allen, Ph.D.
School of Psychology
Complex Systems Management
Australian National University
Center
Canberra, ACT, Australia
Cranfield University
School of Management Stephen J. Guastello, Ph.D.
Bedford, England, Department of Psychology
United Kingdom Marquette University
Milwaukee, Wisconsin
Kathleen M. Carley, Ph.D.
School of Computer Science Geoff Hollis, B.Sc.
Carnegie Mellon University Department of Psychology
Pittsburgh, Pennsylvania University of Cincinnati
Cincinnati, Ohio
Sylvain Chartier, Ph.D.
Uli Junghan, M.D.
School of Psychology
University Hospital of Psychiatry
University of Ottawa
University of Bern
Ottawa, Ontario, Canada
Bern, Switzerland
Thomas J. Dishion, Ph.D. Heidi Kloos, Ph.D.
Child and Family Center Department of Psychology
University of Oregon University of Cincinnati
Eugene, Oregon Cincinnati, Ohio

xi
xii Contributors

Matthijs Koopmans, Ed.D. Patrice Renaud, Ph.D.

Academy for Educational Department of Educational Psychology
Development and Psychology
New York, New York University of Quebec at Outaouais
Montreal, Quebec, Canada
Larry S. Liebovitch, Ph.D.
Center for Complex Systems William Sulis, M.D., Ph.D.
Florida Atlantic University Department of Psychiatry and Behavioral
Boca Raton, Florida Neuroscience
McMaster University
Erika S. Lunkenheimer, Ph.D. Hamilton, Ontario, Canada
Child and Family Center
Wolfgang Tschacher, Ph.D.
University of Oregon
University Hospital of Psychiatry
Eugene, Oregon
University of Bern
Tullio A. Minelli, Ph.D. Bern, Switzerland
University of Padova Robin R. Vallacher, Ph.D.
Galileo Galilei Department of Department of Psychology
Physics Florida Atlantic University
Padova, Italy Boca Raton, Florida
Andrzej Nowak, Ph.D. Paul van Geert, Ph.D.
Department of Psychology University of Groningen
University of Warsaw The Heymans Institute
Warszawa, Poland Groningen, the Netherlands

David Pincus, Ph.D. Guy C. Van Orden, Ph.D.

Department of Psychology Department of Psychology
Chapman University University of Cincinnati
Orange, California Cincinnati, Ohio
Preface

Nonlinear dynamical systems (NDS) theory is the study of how complex pro-
cesses unfold over time and is sometimes known as chaos theory or complexity
theory. Perhaps it was one of those experiences in which “You had to be there,”
but the early days of NDS in psychology were rife with excitement. There was so
much potential for solving old and new problems and transforming the way psy-
chology was studied that everyone present knew it could occupy entire careers.
The days of saying, “Here’s what chaos and complexity can do!” were gone years
ago, however. NDS scholars have embarked on the less glamorous but ultimately
more important task of systematic model building and developing an empirical
research agenda. We estimate that about 50 books are published each year that
are relevant to some aspect of NDS, psychology, and the life sciences, as well
as numerous articles in peer-reviewed journals, such as Nonlinear Dynamics,
Psychology, and Life Sciences, signifying that the field is moving forward at a
rapid pace.
The waterfall of progress creates two dilemmas that we attempt to resolve by
composing this book. The first is to answer a simple question: “What should I
read first?” The second is how to reconcile the escalating gap between the state
of the science in NDS and the average level of awareness of its accomplishments
by professionals in psychology. At the same time, the number of university
courses in NDS for advanced undergraduates and graduate students has started
to grow, and thus the number of psychologists devoting concerted attention to
the topic is also growing. Yet with all the widely dispersed knowledge and dense
mathematical content in much of the literature, the question of what to read
first is difficult to answer with the resources that are currently available.
We wanted to be able to say, “Try this book! It’s a compendium of the
landmark developments and the state of the art in NDS science in psychological
theory and research.” To do so, we assembled a team of researchers who have
already expanded the frontiers of NDS in a wide range of topics that are of
central importance to psychology and who are actively engaged in empirical
research. Each chapter delineates where conventional thinking leaves off and
xiii
xiv Preface

NDS has been necessary to answer difficult questions. The chapters then move
on to capture the state of the art in each topic area.
The opening chapter explains the basic concepts and principles of NDS –
attractors, bifurcations, chaos, fractals, self-organization, catastrophes, and
agent-based and other simulation models – and how they interrelate. We also
encounter turbulence, sensitivity to initial conditions, the Lyapunov exponent,
entropy, information, and a few types of dimension. We made a point not to
become sidetracked into extensive expositions of NDS methods, although they
are sometimes fascinating in their own right. The most important points about
methods are covered in the last section of Chapter 1, and the authors of sub-
sequent chapters expand on the methods that have been particularly relevant
to the problems addressed in their topic areas. Although all the chapters in the
book review existing work in given areas of knowledge, Chapters 5 and 6 also
present some original data that illustrate the methodological issues inherent in
sensation and perception.
Subsequent chapters provide in-depth coverage of substantive areas, includ-
ing neuroscience, psychophysics, sensation, perception, cognition, and devel-
opmental, social, organizational, and clinical psychology. Taken together, these
chapters indicate that NDS can be meaningfully applied in all of those areas
and that it offers quite a few original perspectives on new questions as well as
old questions to which we thought we already knew the answers. Chapter 2
presents the principal theme of collective intelligence as studied in social insect
populations. The chapter could be just as readily positioned with the material
about human groups, but we thought the nonhuman social dynamics would
be a great way to start the psychological material, especially given that foraging
behavior pops up in human sensation and perception dynamics.
Chapter 3 reviews and consolidates what is now known about neural activity
and synchronization of neural circuits in the brain. Although the “binding
problem” (binding is the translation of neural activity to actual thoughts) has
not yet been solved, it is noteworthy that NDS has catalyzed progress in this
area, and there is no real competing linear theory at the present time.
Chapters 4 through 7 close in on the binding problem from the other direc-
tion. The divisions among psychophysics, sensation, perception, cognition, and
psychomotor response may have been historically convenient, but real-world
behavior involves synchronizations among these psychological processes. Thus
there is a preoccupation with perception-action sequence among neuroscien-
tists, cognitive scientists, and information-processing professionals, just for
starters. Given the nature of synchronization itself, it would be difficult for
mainstream psychological research to continue the way it has been going with-
out cheating a little bit and using NDS constructs and methods.
Chapters 8 through 11 address topics that have been traditional concerns
of clinical and counseling psychologists. Children and adults develop socially
Preface xv

and intellectually across the life span. Many of the influential theories in psy-
chology are stage theories in which discontinuous changes in schemata occur.
Chapter 8 examines the nature of developmental stages along with the litera-
ture that supports the notion of catastrophe models for those discontinuous
changes. Although nature-versus-nurture remains an important principle of
development, Chapter 9 on adolescent psychopathology makes the case that
adaptive and maladaptive responses by children and parents in microsocial inter-
actions play at least as great a role in shaping the functional autonomous pro-
cesses that we observe among adolescents. Chapter 10 concerns the major psy-
chopathologies. Here we encounter theories of dynamical disease and evidence
that shifts to normal perceptual gestalt processes underlie the various stages of
psychosis.
Chapter 11 on psychotherapy starts by considering the true differences among
the prominent therapeutic models currently in use. In the early 1990s, the Dodo
Hypothesis, “Everybody has won, and all must have prizes,” was used to char-
acterize the status of empirical evaluations of various types of psychotherapy;1
it was probably as accurate as it might have been offensive to the true believ-
ers in specific psychotherapy models. The potent moments in psychotherapy,
however, are evident to the dynamically aware therapist who engages the pro-
cesses of self-organization that can reconstruct the self and the interpersonal
relationships with which the self engages.
Chapter 12 covers some well-developed NDS themes in social attraction and
social influence. Other important themes include interpersonal and behavioral
synchronization, which carry over into the next chapters.
Chapters 13 through 15 address group dynamics and organizational behav-
ior. Chapter 13 notes that the current interest among organizational theorists
in adaptive behavior and emergence is laudable, but NDS developed the ideas
much further a decade earlier in the form of complex adaptive systems (CAS).
Important questions concerning coordination in work groups and the 50-year-
old enigma of how leaders emerge from leaderless groups were not unraveled
until the perspective changed to include NDS constructs. Similar issues are all
the more apparent in Chapter 14, in which the goal is to understand what an
organization is. Chapter 15 expands on CAS issues by considering evolutionary
behavior, the role of learning, and the impact of those dynamics on organiza-
tional behavior.
Chapter 16 describes the contributions of agent-based modeling and similar
techniques to the understanding of social influence networks and the behaviors
of groups and organizations. It could make us wonder whether the notion of
social building blocks that expand from dyads to groups to organizations was a
naı̈ve idea, and whether network dynamics are more fundamental after all. The
concluding chapter takes stock of the field as a whole and its growing points for
new developments.
xvi Preface

We are honored that Sir Michael Berry gave us permission to use one of his
graphics for the cover of this book. He describes “Chaotic Quantum Falling” to
us as follows:
Neutrons, all initially traveling at the same speed, are thrown upwards in random
directions and then fall under gravity. The wave shown here represents these quan-
tum particles, and its strength gives the probability of finding one. Red indicates
the brightest places in the wave, and the black snakes are the zero lines where the
particle will never be found. In classical physics, the particles would never get above a
certain height (here about two-thirds up the picture); in quantum physics, the waves
occasionally penetrate this “classical boundary” and reach greater heights. The study
of patterns like this is part of an international research effort in “quantum chaology.”

What we learn from this experience, among other things undoubtedly, is that it
takes a paradigm shift to break through a boundary. With that thought in mind,
we delay no further.

Stephen J. Gaustello
Matthijs Koopmans
David Pincus

Note
1. In Lewis Carroll’s Alice in Wonderland, the characters ran a “caucus race,” which
involved a racecourse that was approximately circular. The participants did not start
from the same place or at the same time. They stopped running when they felt like
doing so. The Dodo presided over the race, determined when it was over, and uttered
the pointed phrase about who won. The fairy-tale race was used as an analogy to
summarize the results of psychotherapy research, given that the various therapies
were used for different purposes and that the amounts and methods of study afforded
to any particular types were very irregular.
 1 Introduction to Nonlinear Dynamics and Complexity
stephen j. guastello and larry s. liebovitch

Elephants and Horses

Things change. Sometimes they change gradually, sometimes dramatically. The
prevailing concept of change in psychology consists of only one form of change,
linear change, which is simply undifferentiated, and with the assumption that
outcomes are proportional to inputs in a straightforward manner. The overre-
liance among psychologists and others on the general linear model as a statistical
tool for depicting change has only served to reinforce this monochrome concep-
tualization of change. Perhaps the most significant deviations from the concept
of linear change are the concepts of equilibrium and randomness. For most intents
and purposes, the concept of equilibrium has been used to describe places or
times when change stops occurring. Randomness suggests that the changes are
unpredictable and not explicable by any known concepts or predictors.
Nonlinear dynamical systems (NDS) theory significantly enriches our capa-
bility to conceptualize change, and it provides a rich array of constructs that
describe many types of change. The concept of equilibrium is no longer specific
enough to describe either the change or the events that surround the point where
change stops. The new constructs are the attractors, bifurcations, chaos, frac-
tals, self-organization, and catastrophes. As this chapter explains, each of these
constructs contains several more, including those associated with the “complex-
ity” of a system. Importantly, change is not proportional to inputs. Large inputs
sometimes produce small results, and a small input at the right time can produce
a dramatic result.
Psychology is not the first science to break out of the linear rut. According
to Stewart (1989), the physical sciences made the transition more than a half-
century ago:

So ingrained became the linear habit, that by the 1940s and 1950s many scientist[s]
and engineers knew little else. . . . [W]e live in a world which for centuries acted as
if the only animal in existence was the elephant, which assumed that holes in the
skirting-board must be made by tiny elephants, which saw the soaring eagle as a

1
2 Stephen J. Guastello and Larry S. Liebovitch

wing-eared Dumbo, [and] the tiger as an elephant with a rather short trunk and
stripes (1989, p. 83–84).

Some nonlinear phenomena, particularly chaos, force us to reconsider what it

means for an event to be random. According to Mandelbrot (1983), the word
random came into English by way of a medieval French idiom meaning “the
movements of a horse that the rider cannot predict” (p. 201). Later-day statistics
and experimental design have placed a great deal of emphasis on what the rider
does not control, resulting in the notion that variability in observations is either
“due to the model” or “due to error.” NDS places more emphasis, however, on
the horse’s point of view: There are reasons for the horse’s motions, and its rider
can get used to them or not. Less metaphorically, simple deterministic equations
can produce processes that appear random, but closer scrutiny may indicate that
they are not. It follows that a lot of so-called error variance can be accounted for
if we can identify the processes that generated the observations; those processes
are most likely to be overwhelmingly nonlinear.
In contemporary colloquial English, people speak of “random events” when
referring to events that occur without any apparent connection to prior events
or to any clues about which the speaker is aware. In NDS, these events are
called emergent phenomena. Their important features are their disconnection
with recent past events and that they occur without the deliberate action of
any person or agent. In other words, those phenomena are novel, and they can
sometimes be clear examples of nonlinear events and deterministic processes,
as will become apparent throughout this book.

General Systems Theories and Paradigms

One of the conceptual foundations of NDS is the general systems theory (GST),
an interdisciplinary theory that contains rules, propositions, and constructs that
extend beyond the confines of a single academic discipline. Within the realm of
GSTs, there is a strong representation of mathematically centered theories; this
approach is usually attributed to von Bertalanffy (1968) and Wymore (1967).
NDS is also an example of GST that is centered on principles of mathematical
origin.
A GST can be regarded as a metatheory and as an overarching methodolog-
ical approach in some cases. A metatheory is a theory that organizes concepts,
objects, or relationships inherent in several local (or specific) theories and places
those into an overarching conceptual framework. The development of such theo-
ries facilitates interdisciplinary cross-fertilization because at the metatheoretical
level, scholars from different disciplines will share a conceptual framework. For
instance, a theory that successfully solves a problem in macroeconomics with
principles of evolutionary biology would qualify as an example of such cross-
fertilization. Thus specific objects can be interchanged from one application to
Introduction to Nonlinear Dynamics and Complexity 3

another, but the relationships among those objects could remain approximately
the same. Another approach might show interchangeability of objects and rela-
tionships, but the “blueprint” that defines the metatheory would tell the scientist
where to look for objects and relationships that could be useful. GST and NDS
can both be seen as examples of metatheories. In fact, the same concepts that are
encountered in this book can also be found in texts that are concerned with the
subject matter of other academic disciplines, such as economics and biology.
When viewed as a methodology, a general systems analyst would typically
begin with a description of the phenomenon of interest and assess the relevance
of particular tenets of the GST to that phenomenon. For example, if one were
to observe a nonnormal frequency distribution with a sharp peak near the left
and a long tail to the right, one might hypothesize that a power law distribution
and the dynamics that usually go with it (see later in this chapter) were involved.
The next step would be to create a model to represent the system using the
tools and constructs of the GST plus any adaptations that are specific to the
application. The model-making process typically draws on the past successes and
failures encountered with applications of the general theory. If the GST is truly
meritorious, the knowledge gained from a successful application would increase
the knowledge about the core principles of the theory and thus facilitate the
hunt for further applications. Hence one good application serves as a metaphor
for another (Guastello, 1995, pp. 4–6), and NDS applications have indeed found
their way into many applications within physics, chemistry, biology, psychology,
economics, sociology, and political science. With repeated successes, one can
develop one’s intuition as to where dynamical events of different types might be
found.
Scholars have pondered whether NDS can be considered a paradigm of sci-
ence itself (Allen & Varga, 2007; Dore & Rosser, 2007; Fleener & Merritt, 2007;
Goerner, 1994; Guastello, 1997, 2007; Ibanez, 2007). A new scientific paradigm
would represent a new approach to a wide range of problems and ask entirely
different classes of questions. It would pursue its answers with its own set of stan-
dards and challenges. Thus a new paradigm unearths and explains phenomena
that could not have been approached through pre-paradigmatic means. Addi-
tionally, the new paradigm could be shown to provide better, more compact,
and more accurate explanations for existing questions.
NDS shows other symptoms of paradigmatic behavior beyond its new per-
spective on randomness, nonlinear structures, and system change. For one thing,
the concept of a dimension is not what it used to be. Although society at large
has assimilated the four basic dimensions of Euclidean space and time, mathe-
maticians have rendered 5 through N dimensions as comfortable abstractions,
and social scientists have extensively analyzed complex multidimensional data
spaces using factor analysis and multidimensional scaling. NDS offers a new
development: fractional dimensions. There are entities between lines and planes,
planes and volumes, and so on. Fractional dimensions, or fractals, are described
4 Stephen J. Guastello and Larry S. Liebovitch

later in this chapter. It is worthwhile to point out early that the fractals serve as
good indicators of the complexity of a system.
Our notions about the nature of a system have shifted as well. In Newton’s
view of a mechanical system, the function of the whole can be understood by
understanding the function of each of the parts. A correction of a flaw in a
system can be accomplished by removing and replacing one of its parts. In a
complex adaptive system (CAS), however, the whole is greater than the sum of its
parts. This description should surely ring familiar to psychologists because of
the Gestalt laws of perception. The parts of the system, perceptual or otherwise,
interact with each other, shape each other, and pass information around, and they
are not replaceable or removable without fundamentally altering the dynamics
of the system as a whole. Furthermore, attempts to correct “flaws,” or to change
otherwise a part in a CAS, often do not succeed because the parts adapt in such
a way as to protect the system from the intrusions of the outside tinkerer. By
the same token, a CAS can survive intrusions or assaults by the same means;
the tendency for brain tissue to pick up a function that was lost by a nearby
tissue area is an example. These self-organizing tendencies are a primary area of
interest in the study of NDS, and they are expanded further in this chapter and
throughout the remainder of this book.
Has NDS produced better explanations of phenomena? The answer in each
chapter that follows is “yes.” It has described phenomena that could not be
described in any other way, especially when temporal dynamics are concerned.
In some cases, they provide an organizing center where several partially relevant
theories were needed to describe a phenomenon; work-group coordination is
one such example. Although the explanations may appear more complex in one
respect, they become more efficient to the extent that the general principles
transcend many problems in psychology and also expand beyond the usual
confines of psychology.
Does it account for more of the data than conventional linear models? In
most cases reported in this book, there was no linear alternative available, so
the answer is a simple “yes.” In the cases in which it was possible to compare R2
coefficients for linear and nonlinear models, the cumulative advantage is 2:1 in
favor of the nonlinear models (Guastello, 1995, 2002). That is to say that about
50% of the explanation for a phenomenon comes from knowing what dynamics
are involved.

Elements of NDS Functions

It is no secret that the central concepts of NDS have mathematical origins, some
of which might qualify as exotic even among mathematicians. Fortunately, the
majority of the concepts can be represented in pictorial or graphic form, which
helps interpretations greatly. In psychological applications, we only need to use
Introduction to Nonlinear Dynamics and Complexity 5

the products of the mathematics; we do not need to reperform derivations, solve

differential equations, or anything of that nature. The principles are analytic
when applied to psychology, nonetheless. Some equations have a great deal of
meaning, and they serve to codify the dynamical principles concisely. There are
some novel computations that are often used in NDS research, and some of
them can be rendered statistically, which is perhaps more compatible with the
analytic habits of social scientists.
The equations that we do encounter are all functional and sometimes struc-
tural as well. Some functions that we encounter state that a dependent measure
Y is a nonlinear function of another variable X. Some functions are itera-
tive, in which Y at Time 2 (Y2 ) is a function of Y1 and other variables in the
system.
In an iterative function, a time series of Y is produced by running Y1 through
the equation to produce Y2 ; Y2 is then run through the equation to produce
Y3 , and so on, thereby producing a forecast of future behavioral patterns. Many
psychological phenomena that occur over time are iterative in nature, such as
the flow of a conversation or series of conversations between two people; see
Chapters 9 and 11 through 13 in this volume. If we do not know what equation
is the best representation for the phenomenon, it helps to recognize when an
iterative function could be implied.
Structural equations are similar to functional equations, but they have the
added feature of allowing, if not encouraging, the user to plug complex functions
into a position that might have been signified by a single variable. For instance,
we might have a function containing a bifurcation effect called c, but c might
consist of several psychologically defined variables. Sometimes an element that
looks as simple as c could turn out to be a constant, a variable, or an entire
complex function.
Control parameters are essentially independent variables, with the important
difference that they can act in ways that are often more interesting than the
simple additive relationships that are found in conventional research designs.
Several distinct types of control parameters are described in a later section of this
chapter and in subsequent chapters. Order parameters are essentially dependent
measures in the social scientist’s worldview. There may be more than one order
parameter in some complex dynamical systems, however. Order parameters
within a system might be completely independent of each other, or they might
interact with each other as they evolve over time.
The basic principles of nonlinear dynamics are presented next. In addition
to the references that are interspersed in the text, we recommend Abraham and
Shaw (1992), Bassingthwaighte, Liebovitch, and West (1994), Kaplan and Glass
(1995), Korsch and Jodl (1999), Liebovitch (1998), Nicolis and Prigogine (1989),
Puu (2000), Scott (2005), Sprott (2003), and Thompson and Stewart (1986) as
broad technical references to the mathematical literature.
6 Stephen J. Guastello and Larry S. Liebovitch

The chapter concludes with an overview of the analytic techniques that one
might apply to real data to determine which types of dynamics are present.
The authors of the subsequent chapters elaborate on their uses of particular
techniques.

Attractors and Chaos

An attractor can be seen as a box of space in which movement could take place
or not. When an object, represented by a point, enters the space, the point does
not leave, unless a strong enough force is applied to pull it out. An attractor, like
a magnet, has an effective range in which it can draw in objects, known as its
basin. Some attractors are stronger than others, and stronger attractors have a
wider basin.
An attractor is regarded as a stable structure because all the points within it
follow the same rules of motion. There are four principal types of attractors: the
fixed point, the limit cycle, toroidal attractors, and chaotic attractors. Each type
reflects a distinctly different type of movement that occurs within it. Repellors
and saddles are closely related structures that are not structurally stable.

Fixed-Point Attractors
As its name suggests, a fixed-point attractor is one where the trajectories of points
that are pulled within it gravitate toward a fixed point. Figure 1.1A illustrates
points that are being pulled in from all directions toward the epicenter. A point
might start its path by shooting by the attractor, and might even have a chance
of breaking free, but is pulled into the attractor, where it rests. Figure 1.1B is a
spiral attractor, also known as a sink; the point spirals into the epicenter rather
than moving in directly.
Attractor behavior that is viewed in a time series would look similar to the
trends shown in Figure 1.1D, which would correspond most closely to the case
where points are being pulled in from all directions. The structural equation
that generates that type of motion is:
Z 2 = θ 1 exp(θ 2 Z 1 ). (1.1)
Equation 1.1 shows the function in iterative form, where the order parameter
Z2 is a function of itself at a previous point in time (Z1 ). θ i are empirical weights
that can be determined through nonlinear regression. θ 1 is a proportionality
constant (this is likely to become important when two or more attractors are
involved in the dynamical field); θ 2 is the more important of the two weights
because it reflects how quickly the points are converging onto the fixed point.
Notice that the trajectories of the points in the figure all started with different
initial values but ended up on the same attractor. If θ 2 is a negative number,
Introduction to Nonlinear Dynamics and Complexity 7

Figure 1.1. Gallery of basic dynamics: A, fixed point attractor (Eqs. 1.1 and 1.2); B, spiral
attractor (Eqs. 1.3); C, trajectory of a control point in a field containing two fixed points;
D, time series of points gravitating to a fixed point; E, limit cycle (Eqs. 1.4 and 1.5);
F, time series for a dampened oscillator (Eq. 1.6); G, limit cycle created by organizing
fixed points into a circle; H, system of three repellors; I, three coupled oscillators; J, saddle
point; K, saddle created by a perturbed pendulum.
8 Stephen J. Guastello and Larry S. Liebovitch

the points are indeed converging; otherwise the points’ trajectories would be
expanding exponentially.
Another equation for a fixed-point attractor that is used often is:

Z 2 = K /[1 + θ 1 exp(θ 2 Z 1 )]. (1.2)

A particular value of Z in Eq. 1.2 is recognized as the asymptotic point, which

is designated as K. The regression weights would indicate how quickly the points
were reaching K. A close variation on Eq. 1.2 is one in which the researcher wants
to determine whether an exogenous independent variable X is responsible for
the movement of Z to K. In that case, one would substitute X1 for Z1 .
A spiral attractor is a more complicated phenomenon to analyze, but we can
describe it nonetheless. Imagine that the trajectory of points on the spiral has
the usual X and Y coordinates, and both X and Y are changing over time, t. We
would need to write equations for both the X and Y functions (Puu, p. 21):

X = θ1 exp[θ 2 t cos(θ 3 t + θ4 )] (1.3)

Y = θ5 exp[θ 2 t cos(θ 6 t + θ7 )].

Equations 1.3 allow for the full range of possibilities that the spiral can be
perfectly circular or stretched (elliptical) by some amount in the vertical or
horizontal direction by changing θ 1 , θ 3 , θ 5 , or θ 6 . θ 4 and θ 7 denote the lower
boundaries of the cosine functions in real numbers. Notice that the forcing
function, θ 2 , is common to both axes. If θ 2 were negative, the system would
be converging to the fixed point and thus stable. If θ 2 were positive, however,
the points would be floating away from the epicenter, and the spiral would
be unstable. The spiral attractor functions become more interesting when we
consider repellors and saddles a bit later.
Figure 1.1C illustrates what can happen in a dynamic field containing two
fixed-point attractors. A point that might have been entered into the dynamic
field in the neighborhood of the upper attractor makes a few irregular turns
around the attractor and then travels to the other attractor. Nothing prevents
the point from revisiting the first attractor before landing on the second one,
where it appears to be spending most of its time. The convoluted pathways take
different forms, depending on the relative strength of the two attractors, their
proximity to each other, and the exact location where the traveling point is
entered.
Figure 1.1C is, in essence, the three-body problem that got the entire study of
nonlinear dynamics going. Henri Poincaré made a profound contribution in the
1890s by identifying the problem that we did not have the mathematics available
to characterize the path of a “speck of dust” in the neighborhood of two celestial
bodies, that is, two strong attractors (Stewart, 1989). Those trajectories were
eventually characterized as chaotic several decades later.
Introduction to Nonlinear Dynamics and Complexity 9

Limit Cycles
A limit cycle attractor is also known as a periodic attractor. Its behavior is cyclic,
in the same way that the earth orbits the sun or the moon orbits the earth. The
motion of the point is thus depicted in Figure 1.1E. Once again, if a traveling
point enters the basin of attraction, it does not leave. Points entering from the
outside of the limit cycle are drawn into the attractor. Points that happen to be
inside it are pulled outward into the orbit but do not move toward the epicenter
in the way that fixed-point or spiral attractors do. Instead, there is an ongoing
oscillation around the epicenter value. Oscillating functions are common in
biology (May, 2001) and economic business cycles (Puu, 2000), although both
types of oscillators tend to become more complex by the presence of other
oscillators in the niche or economy.
A classic limit cycle can be generated by a simple function:
Z = sin(X). (1.4)
Again, one might substitute time, t, for the exogenous variable, X. One might
characterize the function in both an iterative and statistically tractable form as:
Z 2 = θ1 sin(Z 1 − θ2 ) − θ3 cos(Z 1 − θ4 ). (1.5)
Fourier analysis or spectral analysis is often employed to separate compound
oscillators. This is a common practice in auditory signal processing where the
sinusoidal components are additive. The generic structure of a compound oscil-
lator is defined as a sine function, a subtracted cosine function, plus another sine
function, and so on. θ 1 and θ 3 are forcing constants that adjust the amplitude
of the sinusoidal functions. θ 2 and θ 4 are lower limits of the oscillating variable
Z and depend on the scale used to measure Z. Note that if θ 2 or θ 4 becomes a
variable, then f(Z) becomes aperiodic.
Another way to obtain a limit cycle is to place a number of point attractors
in a circular configuration. If a point is injected into the neighborhood of this
configuration, it will exhibit the behavior of a limit cycle (Fig. 1.1G).
We can also have dampened oscillators, which exhibit periodic behavior, but
the amplitude of the fluctuations gradually becomes smaller until the order
parameter gravitates to a fixed point. The temporal signature of a dampened
oscillator would look like the time series that is shown in Figure 1.1F. A dampened
oscillator might not be easily discerned from a spiral fixed point. One would be
looking for a control variable that induced the dampening effect. Equation 1.6
would be a likely place to start:
Z 2 = θ1 exp[θ 2 Xcos(θ 3 Z 1 + θ4 )], (1.6)
where θ 2 X is the variable that induces the dampening with its regression coeffi-
cient. Note that Eq. 1.6 is structurally the same as Eq. 1.3, with a substitution of
X and Z1 for t.
10 Stephen J. Guastello and Larry S. Liebovitch

In light of the potential similarity of spiral points and dampened oscillators,

psychologists who might encounter these dynamics should sample a long enough
time series to view the full dynamics of the system. The concept of “restriction of
range” is probably familiar to most psychologists from linear statistical analysis.
A comparable concept here is restriction of topological range, meaning that the
full range of topological events or movements needs to be captured in the
data set to draw reliable conclusions about the dynamics of the model under
study (Guastello, 1995). This principle will recur later in conjunction with other
dynamics.
We can also couple oscillators so that two or more are linked to each other.
Again using Eq. 1.6 as a starting point, X would be a second order parameter, and
it would also have an equation that is structured like Eq. 1.6. We can also imagine
a third oscillator Y in the system that is responsible for the sinusoidal forcing of
X. To close this system of three variables, let Z be the forcing parameter of Y:
Z 2 = θ1 exp[θ2 Xcos(θ3 Z 1 + θ4 )],
X 2 = θ5 exp[θ6 Y cos(θ3 X 1 + θ4 )], (1.7)
Y2 = θ1 exp[θ2 Zcos(θ3 Z 1 + θ4 )].
According to a theorem by Newhouse, Ruelle, and Takens (1978), three
coupled oscillators are minimally sufficient to produce chaos. Chaos, for present
purposes, is a highly unpredictable time series of numbers or events that appears
random but that is actually the result of a deterministic process. Chaos in the
form of coupled oscillators can be readily demonstrated by joining one pen-
dulum to another (Fig. 1.1I), so that they all swing freely. Swing the pendulum
marked “1,” and watch the highly volatile motion of the others.
Not all combinations of three oscillators will produce chaos, however. Puu
(1993) examined different combinations that could occur in intertwined eco-
nomic cycles and concluded that although some combinations produce chaos,
some serve to dampen the volatility of the others in the system. It is also possible
to take two or more oscillators that are oscillating at different speeds, couple
them, and synchronize the whole set (Strogatz, 2003).

Repellors
A repellor is also a box of space, but it has the opposite effect on traveling points.
Any point that gets too close to it is deflected away from the epicenter. It does
not matter where the traveling point goes, so long as it goes away. Thus repellors
characterize an unstable pattern of behavior. We already encountered a type of
repellor when we considered the spiral with a positive coefficient.
Another type of repellor is more akin to the fixed point. A system containing
three such repellors is shown in Figure 1.1H. A beam of points coming in from
the left hits the edge of an attractor basin and is deflected outward, as indicated
by the thin lines around the repellor regions. In some cases, the repellors deflect
Introduction to Nonlinear Dynamics and Complexity 11

the point into another repellor’s basin, whereupon it is deflected again until it
gets free of the repellor region.

Saddles
Saddle points have properties of both attractors and repellors. On one hand, they
attract traveling points. On the other, the points do not stay on the saddle point
very long before they are deflected in another direction. The saddle phenomenon
is similar to the flow of people to and from an information booth in an airport
or other major facility. The information booth attracts people, but the booth is
not their destination. The visitors remain there only long enough to find out
what they need to know to get to where they are going. Again, because the points
are following different rules of flow, the system is not stable.
Figure 1.1J shows one type of saddle node where the vectors cross at an
intersection. The intersection is the saddle. Other trajectories, shown to the sides,
are more pathways where the traveling points approach the intersection and then
turn. This is the classic busy urban intersection that facilitates the placement of
all sorts of businesses. The businesses, if they are interesting enough, form an
attractor in their own right, but that is another matter.
Figure 1.1K shows a saddle that is obtained by perturbing a pendulum. The
arc indicates that the pendulum is swinging in the normal fashion on one axis,
but the oblique vector indicates that a force has been placed on the pendulum so
that it also swings closer to and further from the page. The result is a figure-eight
pattern; the saddle point is marked “S.”

Toroidal Attractors
A toroidal attractor is the result of a limit cycle that is cycling along two axes
rather than one. The result is a bagel-shaped set of pathways. Figure 1.2 shows
two snapshots of the evolution of a toroidal attractor early and late in its regime.
The movement of the traveling point is not systematically “eating” the bagel.
Rather, it is following its travel rule (equations) that takes it along unpredictable
local pathways. It is not until the process has evolved far enough that one is able
to see the full structure of the attractor take shape.
Because of the unpredictable motion of the traveling point, a toroidal attractor
might be mistaken for chaos. It is not quite as complex as chaos, however. Rather,
it is known as an aperiodic attractor, which is an important transition from limit
cycles to chaotic attractors, which are considered next.

Chaos and Chaotic Attractors

As with other attractors, chaotic attractors can be described as points that are
pulled into and stay within the space, even though they are allowed to move
12 Stephen J. Guastello and Larry S. Liebovitch

A B

C D
Change in Driving Speed

Extroverts Introverts

Figure 1.2. Gallery of basic dynamics: A, toroidal attractor in its early iterations;
B, toroidal attractor in its later iterations; C, one of many possible strange attractors;
D, simple bifurcation, pitchfork or cusp type.

within the space. The internal motion within the attractor space is more com-
plex than is the case with the other attractors discussed here. Chaos has three
primary features: unpredictability, boundedness, and sensitivity to initial condi-
tions (Kaplan & Glass, 1995). Unpredictability means that a sequence of numbers
that is generated from a chaotic function does not repeat. This principle is per-
haps a matter of degree, because some of the numbers could look as though they
are recurring only because they are rounded to a convenient number of decimal
points. Rounding of this nature is going to happen in psychological applications
because most psychological measurements simply do not have much meaning
beyond a couple of decimal places; very small differences in most measure-
ments are drowned in measurement error. Boundedness means that, for all the
unpredictability of motion, all points remain within certain boundaries.
The principle of sensitivity to initial conditions means that two points that start
off as arbitrarily close together become exponentially farther away from each
Introduction to Nonlinear Dynamics and Complexity 13

other as the iteration process proceeds. This is a clear case of small differences
producing a huge effect.
Chaotic motion itself is characterized by both expansion and contraction.
When a point veers too close to the edge of the attractor, it is pulled inside. If
it gets too close to the center, it steers outward. A school of fish behaves in the
same manner (Semovski, 2001).
Figure 1.2C is one of many known chaotic attractors. It was chosen for
display simply because it looks good, as well as to provide some comparison
with other attractors. The typical chaotic attractor is operating in more than
two dimensions. Most often, it is generated by two or more equations joined
together, although there is a famous exception, which is considered next. Each
equation reflects the behavior of an order parameter; each order parameter is
affected in part by the behavior of other order parameters; this is not unlike the
structure of Eqs. 1.7. Control parameters are usually present.
The different types of chaotic functions generate time series that may vary in
complexity. The meaning of the complexity depends on the nature of the system
and the theory behind it. For instance, a higher-level complexity could denote
greater adaptability of a living system, whereas lower complexity could denote
rigidity in the system. Two indicators of complexity in a dynamical time series
that is potentially chaotic are the Lyapunov exponent and the fractal dimension.

Lyapunov Exponents and Turbulence

The Lyapunov exponent is based on the concept of turbulence or entropy, such as
might be found in the motion of water or air (Ruelle, 1991). In signal processing,
entropy is the amount of change in a system over space or time that cannot be
predicted by available information (Shannon, 1948). In dynamical systems,
however, a turbulent system is generating information (Nicolis & Prigogine,
1989; Prigogine & Stengers, 1984); information generated at one point in time
might then be used to predict the state of the system at the next point in time.
From either perspective, the exponent reflects the rate at which information that
allows a forecast of a variable y is lost. The faster the information is lost, the
greater the entropy. It is calculated (Kaplan & Glass, 1995, pp. 334–335; Puu,
2000, p. 157) by taking pairs of initial conditions y1 and y2 and their iterations
one step ahead in time, which would be y2 and y3 . If the ratio of absolute values
of differences,

L ≈ |y 3 − y2 |/|y 2 − y1 |, (1.8)

is less than 1.0, the series is contracting. If the value of L is greater than 1.0,
the function is expanding, and sensitive dependence is present. The Lyapunov
exponent, λ, is thus

λ ≈ ln[L ]. (1.9)
14 Stephen J. Guastello and Larry S. Liebovitch

For an ensemble of trajectories in a dynamical field, exponents are computed

for all values of y. If the largest value of λ is positive, and the sum of λi is negative,
then the series is chaotic in the sense that the trajectories are unpredictable,
bounded, and sensitive to initial conditions.
The calculation of Eq. 1.9 is made on the entire time series and averaged over
N values, where N is the last entry in the time series:


N
λ = (1/N) |ln(L )|. (1.10)
N=1

As N increases to infinity, Eq. 1.10 generalizes as:

y = eλt , (1.11)
which is actually insensitive to initial conditions. A positive value of λ indicates
an expanding function, which is to say, chaos. A negative λ indicates a contracting
process, which could be a fixed point or limit cycle attractor.

Bifurcations and Chaos

A bifurcation is a pattern of instability in which a system attains greater com-
plexity by accessing new types of dynamical states (Nicolis & Prigogine, 1989).
It could take the form of a split in a dynamical field in which different dynamics
are occurring in each part. It can also denote a change in an attractor itself, such
as the Hopf bifurcation, in which a fixed point attractor becomes a limit cycle. A
bifurcation structure could be as simple as a critical point, or it could be a more
complex structure such as the bifurcation diagram that appears in Figure 1.2D.
Bifurcation mechanisms are usually observed when a control parameter is chang-
ing value beyond a given threshold value. A bifurcation effect is probably present
if a modulus by n appears in an equation for an f(y), such as Eq. 1.12 that follows;
in those cases, b is the control parameter, and y is the order parameter.

Pitchfork or Cusp Bifurcations

The bifurcation shown in Figure 1.2D is known either as a pitchfork or a cusp,
depending somewhat on the context. The example was first introduced by
Zeeman (1977) as an explanation for some curious results of the impact of alco-
hol consumption on driving speed. Participants in the study drove an automobile
simulator before and after the consumption of sufficient quantities of alcoholic
beverage. They completed a measurement of extroversion–introversion before
the driving tasks. After alcohol consumption, extraverted drivers were able to
maintain the same speed, but introverts were divided into two groups, those
who increased their speed and those who drove more slowly. The change in
speed for the introverts was verified as two attractor states (Cobb, 1981). The
Introduction to Nonlinear Dynamics and Complexity 15

Bifurcation point
Order Parameter

Period doubling

Period 1 Period 2 Period 3

Fixed point Limit Cycle Chaos

Control Parameter
Figure 1.3. Logistic map showing its various dynamics as the control parameter increases
in value (Eq. 1.12).

configuration of two attractor states and the bifurcation variable (introversion–

extraversion) forms a cusp catastrophe model, which is considered later in the
chapter.

Logistic Map
The logistic map (Fig. 1.3) is one of the more famous bifurcations in NDS and
was studied extensively by May (1976). It represents the transition from a fixed
point attractor to a limit cycle, and from a limit cycle into chaos all through one
equation:
Y2 = BY1 (1 − Y1 ), 0 < Y1 < 1, (1.12)
where Y is the order parameter, and B is the control parameter. In the region
of Figure 1.2 labeled as Period 1, the system is globally stable, with dynamics
that are characteristic of a fixed point attractor. The bifurcation point marks the
transition from Period 1 to Period 2 where the attractor becomes a limit cycle
(Hopf bifurcation) as B increases in value.
Toward the end of Period 2, where B has increased further in value, the bifur-
cation pattern bifurcates again, dividing the system into four smaller regions.
Here one observes cycles within cycles. This is the phenomenon of period dou-
bling, which later became also known as a Feigenbaum bifurcation. When B
16 Stephen J. Guastello and Larry S. Liebovitch

becomes larger still, the system bifurcates again, so that there are oscillations
within oscillations within oscillations again.
The system enters Period 3, which is full-scale chaos, after one more set of
bifurcations. The behavior of Y bears little resemblance to the relative order of
Periods 1 or 2. Of further interest are the windows of relative order that striate
the chaotic region. The windows contain ordered trajectories that create a path,
with additional bifurcations, from one chaotic period to another.
The logistic map can be expanded to reflect a dynamic that has the option of
going into period doubling or not, such as the cubic variant in Eq. 1.13:
Y2 = BY (1 − Y1 )(1 − Y1 ). (1.13)
The logistic map function generalizes to an exponential function:
Y2 = BY1 exp(Ay1 t), (1.14)
which has seen a good deal of application in population dynamics (May & Oster,
1976), where B is the birth rate of the organisms and A is a crowding parameter
of ecological importance.

Other Bifurcations
The bifurcations that have been discussed to this point, including all those asso-
ciated with catastrophe theory, are known as subtle bifurcations. There are two
other classes of bifurcations known as catastrophic and annihilation bifurcations
(Abraham & Shaw, 1992). In catastrophic bifurcations, new attractor regimes
pop in and out of the dynamical field as a control parameter changes. In annihila-
tion bifurcations, a control parameter induces two or more dynamical structures
to collide, producing a dynamical field that did not exist previously. Neither of
these types of bifurcations has surfaced in psychological applications, and to our
knowledge, an annihilation bifurcation only surfaced once in sociology in an
application to urban growth and renewal (Guastello, 1995).

Fractals and Power Laws

Fractals are geometric structures with noninteger dimensions. They have some
important properties that characterize nonlinear events: their scale–scale prop-
erty, scaling relationships, and a characteristic probability density function
(PDF).
Look outside your window. Out there are trees in the grass, clouds in the sky,
roads on the land, and people in cities. A normal (or Gaussian) distribution has
a few small things, a lot of things of about an average size, and a few big things.
Trees, clouds, roads, and people are not at all like this, however. The tree has
an ever larger number of ever smaller branches. There is an ever larger number
of ever smaller clouds in the sky. There is an ever larger number of narrower
Introduction to Nonlinear Dynamics and Complexity 17

roads (from Interstate 95, to State Road 808, to 20th Street, to the path to
the Behavioral Sciences Building outside our window). There is an ever larger
number of tiny villages than large cities. Objects in space, or processes in time,
or sets of numbers, with this kind of statistical relationship that we so often see
in nature, are called fractals.

Self-Similarity
A property that all fractals have in common is that they are self-similar, which
means that the small pieces are like their larger pieces. For a purely mathematical
fractal, the small pieces can be exact smaller copies of the larger pieces, which is
called geometric self-similarity. However, the small tree branches, clouds, roads,
and villages are not exact smaller copies, but are only kind of like their larger
branches, clouds, roads, and cities. This is called statistical self-similarity. It
means that the statistics describing the small pieces are similar to the statistics
that describe the larger pieces.

Scale-Free
If I hide the rest of the tree, it is quite hard for you to tell whether you are looking
at a small branch or a large branch. If I hide the rest of the sky, it is quite hard for
you to know whether you are looking at a small cloud or a large cloud. Thus these
images are independent of the scale at which they are seen. They seem to be the
same at all scales; they are scale-free. There is no one scale that best characterizes
the size of the tree branches or clouds or the roads or the number of people in
a city. The PDF for the normal distribution is
1 (x−m)2
PDF(x) = √ e 2s 2 . (1.15)
2πs 2
The values of the variate x are characterized by two parameters: the mean, m,
and the standard deviation, s.
For a fractal, however,

PDF(x) = Ax −b , (1.16)

the values are characterized by the shape parameter, b, rather than m or s. The
values extend over many scales, and so no one scale is a good way to characterize
the data. The property of statistical self-similarity described earlier – namely, that
the statistics of the small pieces at size x are similar to the statistics of the larger
pieces size Ax – is formally that PDF(x) has the same shape as PDF(Ax). Equa-
tion 1.16 is a power law distribution. Two examples are shown in Figure 1.4A,
each with a different (negative) shape parameter. The shape parameter is not
affected by s for the variate x (z = x/s in Fig. 1.4A).
 Another Random Document on
Scribd Without Any Related Topics
If the second copy is also defective, you may demand a refund
in writing without further opportunities to fix the problem.

1.F.4. Except for the limited right of replacement or refund set

forth in paragraph 1.F.3, this work is provided to you ‘AS-IS’,
WITH NO OTHER WARRANTIES OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR ANY PURPOSE.

1.F.5. Some states do not allow disclaimers of certain implied

warranties or the exclusion or limitation of certain types of
damages. If any disclaimer or limitation set forth in this
agreement violates the law of the state applicable to this
agreement, the agreement shall be interpreted to make the
maximum disclaimer or limitation permitted by the applicable
state law. The invalidity or unenforceability of any provision of
this agreement shall not void the remaining provisions.

1.F.6. INDEMNITY - You agree to indemnify and hold the

Foundation, the trademark owner, any agent or employee of the
Foundation, anyone providing copies of Project Gutenberg™
electronic works in accordance with this agreement, and any
volunteers associated with the production, promotion and
distribution of Project Gutenberg™ electronic works, harmless
from all liability, costs and expenses, including legal fees, that
arise directly or indirectly from any of the following which you
do or cause to occur: (a) distribution of this or any Project
Gutenberg™ work, (b) alteration, modification, or additions or
deletions to any Project Gutenberg™ work, and (c) any Defect
you cause.

Section 2. Information about the Mission

of Project Gutenberg™
Project Gutenberg™ is synonymous with the free distribution of
electronic works in formats readable by the widest variety of
computers including obsolete, old, middle-aged and new
computers. It exists because of the efforts of hundreds of
volunteers and donations from people in all walks of life.

Volunteers and financial support to provide volunteers with the

assistance they need are critical to reaching Project
Gutenberg™’s goals and ensuring that the Project Gutenberg™
collection will remain freely available for generations to come. In
2001, the Project Gutenberg Literary Archive Foundation was
created to provide a secure and permanent future for Project
Gutenberg™ and future generations. To learn more about the
Project Gutenberg Literary Archive Foundation and how your
efforts and donations can help, see Sections 3 and 4 and the
Foundation information page at www.gutenberg.org.

Section 3. Information about the Project

Gutenberg Literary Archive Foundation
The Project Gutenberg Literary Archive Foundation is a non-
profit 501(c)(3) educational corporation organized under the
laws of the state of Mississippi and granted tax exempt status
by the Internal Revenue Service. The Foundation’s EIN or
federal tax identification number is 64-6221541. Contributions
to the Project Gutenberg Literary Archive Foundation are tax
deductible to the full extent permitted by U.S. federal laws and
your state’s laws.

The Foundation’s business office is located at 809 North 1500

West, Salt Lake City, UT 84116, (801) 596-1887. Email contact
links and up to date contact information can be found at the
Foundation’s website and official page at
www.gutenberg.org/contact
 Section 4. Information about Donations to
the Project Gutenberg Literary Archive
Foundation
Project Gutenberg™ depends upon and cannot survive without
widespread public support and donations to carry out its mission
of increasing the number of public domain and licensed works
that can be freely distributed in machine-readable form
accessible by the widest array of equipment including outdated
equipment. Many small donations ($1 to $5,000) are particularly
important to maintaining tax exempt status with the IRS.

The Foundation is committed to complying with the laws

regulating charities and charitable donations in all 50 states of
the United States. Compliance requirements are not uniform
and it takes a considerable effort, much paperwork and many
fees to meet and keep up with these requirements. We do not
solicit donations in locations where we have not received written
confirmation of compliance. To SEND DONATIONS or determine
the status of compliance for any particular state visit
www.gutenberg.org/donate.

While we cannot and do not solicit contributions from states

where we have not met the solicitation requirements, we know
of no prohibition against accepting unsolicited donations from
donors in such states who approach us with offers to donate.

International donations are gratefully accepted, but we cannot

make any statements concerning tax treatment of donations
received from outside the United States. U.S. laws alone swamp
our small staff.

Please check the Project Gutenberg web pages for current

donation methods and addresses. Donations are accepted in a
number of other ways including checks, online payments and
credit card donations. To donate, please visit:
www.gutenberg.org/donate.

Section 5. General Information About

Project Gutenberg™ electronic works
Professor Michael S. Hart was the originator of the Project
Gutenberg™ concept of a library of electronic works that could
be freely shared with anyone. For forty years, he produced and
distributed Project Gutenberg™ eBooks with only a loose
network of volunteer support.

Project Gutenberg™ eBooks are often created from several

printed editions, all of which are confirmed as not protected by
copyright in the U.S. unless a copyright notice is included. Thus,
we do not necessarily keep eBooks in compliance with any
particular paper edition.

Most people start at our website which has the main PG search
facility: www.gutenberg.org.

This website includes information about Project Gutenberg™,

including how to make donations to the Project Gutenberg
Literary Archive Foundation, how to help produce our new
eBooks, and how to subscribe to our email newsletter to hear
about new eBooks.