What are Complex Adaptive Systems?
The term 'complex adaptive systems' signals an approach which examines adaptive systems as non-linear, many-body, distributed, dissipative and self-organising, internally organised, dynamical systems. According to this approach, adaptiveness is the capacity to self-modify in the face of environmental change so as to preserve some set of features (typically functions) intact and adaptedness is the condition of sustaining those features. Acquiring either often involves many kinds of processes, including developmental and behavioural modification in response to the environment as well as Darwinian evolution.
To get a sense of the nature of complex adaptive systems, some examples are useful. Living organisms are paradigm complex adaptive systems, with fantastically intricate metabolisms, the ability to repair themselves, specialised organs (in the case of multicellular organisms), diverse behaviour repertoires (especially in animals), and the ability to evolve. At a lesser degree of complexity, some kinds of biochemical systems show forms of self-organization with properties that appear highly relevant to the emergence of life, and might be considered to be proto-adaptive systems. Moving up in scale, animal societies, and of course human economic and socio-political systems like businesses, cities and nations, can be complex adaptive systems. It is possible that some ecological systems should be counted as well, but that is whether the requisite internal organisation is present is often a difficult matter to determine. What is of importance is that a very rich and precise conceptual and principled account of complex adaptive systems is slowly being elaborated that will in future permit many more useful distinctions among, and understandings of, them than we can grasp today.
The significance of Complex Adaptive Systems Research
Complex adaptive systems research is a highly diverse field that poses many scientific and philosophical challenges. Complex adaptive systems are important because they are central to understanding life and mind. The way that we understand complex adaptive systems thus has implications that range from biochemistry to our philosophic understanding of nature and ourselves. It is also a matter of great practical importance that we learn how to manage complex adaptive systems so that we can live in a humane, ecologically and economically sustainable society.
Key points of interest in complex adaptive systems research:
Theory is in flux: new nonlinear far-from-equilibrium approaches to complex adaptive systems are challenging or reworking established theories such as Darwinian evolution theory, neoclassical economics and cognitivist psychology.
The new approaches suggest radically different concepts and metaphysics, changing deep assumptions about the nature of scientific explanation in the domains concerned (in particular, biological and social).
One of the key implications of complexity for the conduct of scientific research is the end of disciplinary autonomy. Research in biology and cognitive science is starting to reflect this, with a new breed of interdisciplinary research programs like 'evo-devo' and 'dynamical embodied cognition' coming to the forefront
The old guard and the new radicals
One of the reasons that complex adaptive systems research is particularly vital right now is that theoretical approaches to it are in flux. There are three great classical theories that have been used to understand what we are referring to as complex adaptive systems: Darwinian evolution theory, neoclassical economics, and cognitivist psychology. These theories are now being challenged by the advent of nonlinear, far-from-equilibrium, self-organising models, which are providing a radical new way to understand complex adaptive systems. Not only does the new kind of approach illuminate crucially important phenomena such as self-organisation and emergence, it provides a basis for critiquing the old theories. These look suspiciously as though they are treating complex systems like simple systems, for instance by assuming near-to-equilibrium dynamics.
But if there is a revolution afoot, it is a protracted and rather messy one. Whilst mathematical chaos made an impact on popular imagination during the 1980s, the percolation of complex systems ideas through the sciences has been less sweeping and rapid than some advocates might have hoped for. The established disciplines have produced well developed bodies of research to speak for their value. Moreover there is no lack of scientists and philosophers devoted to unpacking and elaborating their general theoretical and conceptual structure. In contrast, although growth in the essential use of complex systems models in science has been explosive, especially outside physics and chemistry, there are still few clear paradigms or guidelines for using the new complex systems ideas to frame empirical and theoretical research. And whilst there are many promising models of phenomena ranging from chemical self-organisation through genome modelling to ‘swarm intelligence', there is not yet a corpus of successful research to match that of the older disciplines.
Thus, some of the most crucial problems of complex adaptive systems research lie in developing the new complex systems ideas as frameworks for robust, empirically productive scientific research programs, and with relating the new approaches to the older theories. This issue is far from straightforward, because although the relationship is unsettled, a sceptical view is that the new approaches might not be as revolutionary as they seem. According to this view, the established theories can absorb some of the implications of the ‘complexity' disciplines as peripheral emendations and steam on barely troubled. However, a more likely result is that there will be a protracted period of mutual readjustment as researchers struggle to weld together old and new ideas into new syntheses. Of the three classic theories, cognitivist psychology is clearly in significant trouble. Traditional AI has stalled over intractable difficulties like the frame problem and a failure to produce effective simulations of real world intelligence, and there is a plethora of new 'dynamical embodied' approaches which are, by comparison, making great progress. Neoclassical economics is also experiencing serious challenge: for example new ‘innovation' theories argue that one of the principal drivers of economic growth - technological innovation - arises precisely because markets are imperfect and not at equilibrium. This is not a peripheral emendation.
In contrast with the other two, the standing of Darwinian evolution theory seems stronger than ever. It has no serious rivals, and appears if anything to be extending it's reach into new domains, including economics. However there are controversies which point to some deep problems, most notably the debate over the 'gene-centered' version of evolution theory promoted by Richard Dawkins. Darwin had no theory of inheritance, and his theory of evolution is quite neutral with respect to the nature of the causality that underlies evolution. The New Synthesis version of evolutionary biology, which took shape in the 1940's, added to Darwin's theory a mechanistic theory of population genetics based on Mendelian inheritance. Population genetics treats genes as the 'units' of inheritance, and models evolution abstractly in terms of changes in gene frequency within a population. Dawkins carries this picture through to it's logical extreme, elevating genes as the prime causal agents in evolution. However, in retrospect this will probably turn out to be a mistake. It is becoming increasingly apparent that the causal processes of organism development and inheritance are highly distributed, and genes simply aren't the sole controlling agents that they have been painted as. They are one kind of element in a complex network of processes. Ultimately the neutrality of Darwin's theory of evolution with respect to underlying causal processes may be a great strength, allowing it to survive a transition from simple systems models of those processes to complex systems models, perhaps even being enriched along the way.
The shock of the new
So, contrary to the sceptical view, the impact of nonlinear 'complexity' approaches is proving wide ranging, if sometimes quite subtle, and still has some way to run. Their potential can be briefly indicated by outlining some important general implications:
A new kind of metaphysics: Complex adaptive systems are characteristically made up of distributed, self-organising dynamical processes. This contrasts with traditional metaphysical assumptions in science and philosophy, which typically posit localised structures as the prime causal agents (e.g. basic particles, genes in the body, representations in the mind). The new type of metaphysics has implications for the basic structure of scientific theories, including default assumptions about what should be explained, and how. In nonlinear far-from-equilibrium systems, processes are typically more fundamental than structures, and change rather than stasis is often the norm.
Theoretical pluralism and interdisciplinary research programs: The interaction of many kinds of processes across multiple spatial and temporal scales is a characteristic feature of complex adaptive systems, and it poses central challenges to the conventional disciplinary structure of science. The Newtonian ideal that a scientific theory should be both comprehensive in its domain and quantitatively precise is unattainable for biological and social domains, both because the phenomena are so ferociously complex that simplification cannot be avoided, and because the domains ‘bleed' into one another (for instance your neuronal processes mutually interact with your social relations). Because of this, disciplinary autonomy is being trampled as scientists are increasingly forced to grapple with the problem of coordinating many partial theories and models to explain biological and social phenomena. The result is the emergence of a fascinating new kind of synthetic cross disciplinary research program. In biology, for example, evolution theory, developmental and molecular biology are being integrated to form the highly successful ‘evo-devo' interdisciplinary research program (and some argue that ecology should be added to create ‘eco-evo-devo'!). In cognitive science, psychology has taken some tentative steps towards integration with evolutionary biology, and arguably this should be extended to include development, neuroscience and society.
Like no time since the early stages of modern science, then, there is an urgent need to revisit the fundamental conceptual and metaphysical assumptions guiding science. And, indeed, philosophy. Basic assumptions about causality and scientific explanation need to be critically re-examined. And if the Newtonian goal for theory that is both comprehensive and precise is impossible, there is all the more need for scientists and philosophers to develop qualitative ‘big picture' sketches. These sketches can do much to assist in framing and coordinating the new interdisciplinary research programs; research programs that arise directly from the problem of tackling complexity head on.