Reviewed by
Julia Schindler
Free University, Amsterdam
The book by John H. Miller can be understood as an emphatic plea and an attempt to scientifically refocus on the unifying principles that organize our world. Motivated by the increasing fragmentation of science during the past decades, the author argues that the discipline-transcending science of complexity can and should be used to identify features and processes that complex adaptive systems have in common. Using illustrative examples from the biological to the socio-economic and from the micro- to the macro-scale, the author makes an attempt to translate this call into action. Actual and potential commonalities across a range of diverse systems are highlighted, organised by common themes such as interaction, feedback, heterogeneity, networks, and intelligence. In addition, potential candidates for shared laws are presented, culminating in a fundamental theorem for complex adaptive systems.Although current monographs on complex systems treat such aspects and concepts in depth (e.g. Mobus and Kalton, 2015) or in relation to other disciplines (e.g. White et al., 2015; Byrne and Callaghan, 2013), the strength of this book, in my view, lies in making the focal shift from the characteristics of a system towards the role or function that they fulfil for a system, accessible to a wider audience. It makes the reader get a vivid idea of how the focus on the functions and their typical ways of emergence, rather than their content, make potential similarities between systems visible. Current existing monographs on complex systems vary in their perspective, e.g. by conceptual, theoretical, disciplinary, or methodological perspective. While each focus has its purpose, the purpose of this book is obviously to motivate defocusing the perspective from the detailed to the holistic. It does so by taking the reader by the hand on a colourful sightseeing tour of illustrative examples, contrasted with each other and aligned along common complexity concepts and themes in meaningful ways.
In the introductory Chapter 1, the author elaborates his motivation and plea, and gives an outline of the subsequent chapters. He argues that the inherent quest of science to define the research frontier by consistently publicly disclosing, evaluating and correcting ideas, has proven a powerful engine to produce knowledge, but contributes little to putting the pieces together and discovering their hidden patterns. He suggests that what he calls the core principles of complexity science can be a useful tool to do so. In the remainder of the chapter, he then illustrates each of these core principles with examples, thereby giving an outline of the coming chapters, each devoted to one principle.
The presentation of core principles is opened by introducing the notions of interaction and emergence in Chapter 2. The major point in this chapter is that simple (and local) interactions can lead to complex emergent phenomena, but that vice versa complex interactions can also lead to simple emergent phenomena. The author illustrates this difference with intuitive examples such as the pattern of a sea shell and market equilibria.
In Chapter 3, the author takes the logical next step of introducing feedback, a consequence of interactions. While mentioning the stabilizing function of negative feedback, this chapter focuses on illustrating the destabilizing and potentially devastating power of positive feedback loops. As an impressive example of such a loop, it is described how a minor flaw in the algorithm of a single trading computer caused price crashes on the stock exchange.
The diversity of the functions of heterogeneity in complex systems and how these functions are facilitated by interaction and feedback is addressed in Chapter 4. In particular, the stabilising function of heterogeneity in bee-hives with regard to bee-hive temperature is contrasted with the destabilising function of social heterogeneity in triggering socio-political revolts.
The next chapter addresses noise, which is regarded as the system's ability to allow for errors. Based on an impression how interaction, feedback, and heterogeneity contribute to forming a system's state space, it is suggested that one of the functions of noise is helping a complex system to find its optimum configuration by allowing for sub-optimal moves with the state space. This is accompanied, among others, by an intriguing example of medicine development using simulated annealing.
With Chapter 6 then, the existence of molecular intelligence is emphasised, illustrated by the ability of single-cell organisms like bacteria to act on their own behalf. Being acquainted to the internal mechanisms of some of these small-scale organisms, the reader dives into a perspective of how intelligence between humans and bacteria might compare in functional terms, irrespective of its respective biological basis, i.e. neurons vs. molecules.
In logical consequence, Chapter 7 is then devoted to how group intelligence can emerge from interactions among lower-level intelligent organisms. This is, among other examples, impressively illustrated by a perspective on a bee-hive, viewing it as a reproducing, subsisting, and acting organism on its own.
The topic of networks is touched in the next chapter. Here, based on examples of housing and neighbourhood, the reader gets an impression of the diversity of network functions, e.g. speeding up message passing or segregation, how these functions are supported by the particular network structure, and how very different systems might resemble in terms of their network structures.
Chapter 9 opens the floor for speculation about potential universal laws, focusing predominantly on scaling laws. Scaling laws describe how the features of a system change with the system's scale. Examples for scaling laws that seem valid across a range of contexts and systems are highlighted, including potential joint causes, such as physical constraints.
The discussion on scale issues is followed by a delineation of how cooperation among organisms can emerge, often leading to beneficial behaviour at system level, in spite of incentives to compete. This is illustrated by an analysis of a rice-terrace system in Bali and an abstract computer model based on genetic algorithms and game theory. The latter offers an alternative explanation of cooperation to kin selection based on the common genetic basis of a population.
The last of the main chapters provides a fascinating historical excursion to nuclear reaction research, the first programmable electronic computer, and the development of Markov Chain and Monte Carlo methods. A finding that bounds these themes together is the discovery of a phenomenon that the author calls the fundamental theorem for complex adaptive systems. It is an interesting statement about the relationship between the probability distribution of the possible states of a system and the criterions the system elements use for decision-making. The validity of this relationship for a particular system would have important implications, such as the functional ability of systems to avoid getting stuck in local optima.
Written in a nonchalant manner and with wit, I find the general style of the book very pleasant. Though the complexity and descriptive granularity of the examples presented differ to some extent, most of them are well-selected and illustrative for the intended purpose, nevertheless. Each of the chapters has a clean-cut theme, and all-together they provide the reader with a coherent structure which makes reading really comfortable. However, the overall theme binding the chapters together could have been more pronounced. One of the binding themes, and with it one of the book's strengths to me, is that the book assumes and invites for a functional and evolutionary perspective on systems in order to make systems comparable. This or any other intended overall binding theme, along with the chapters' main conclusions, could have been made more explicit.
I think the author's plea for daring a more systematic discovery of structural, functional, and evolutionary patterns to present a big picture is an important one. The obvious increasing fragmentation of science and knowledge clearly calls for a better organisation of knowledge. In this regard, Balietti et al. (2015) could show, based on simulation experiments, how scientific fragmentation can limit scientific progress and that, vice versa, decreasing fragmentation can have positive effects on scientific progress. A related finding by these authors was that progress is increased when views and approaches are frequently confronted with empirical data. In this regard, the book is a good showcase how ground truth can be espoused with abstraction.
With regard to the workout of the book, the chapters and their passages are heterogeneous in challenge and detail. Apart from a few difficult passages, the good readability and conciseness of the book (272 pages overall) makes it a nice motivational introduction to the topic for beginners. Scientists reading the book might miss references. But maybe the target audience is not the point of the book. I think the major point of the book is to make the reader curious, be it scientists or average readers, and to reward them with an inspiring feeling of which laws in the universe might still be there to be discovered.
ReferencesBYRNE, D., and Callaghan, G. (2013), Complexity theory and the social sciences: The state of the art. London: Routledge.
MOBUS, G. E. and Kalton, M. C. (2015), Principles of systems science. New York: Springer.
WHITE, R., Engelen, G. and Uljee, I. (2015), Modeling Cities and Regions as Complex Systems: From Theory to Planning Applications. Boston: MIT Press.
Return to Contents
of this issue