Complexity and the History of Economic Thought

David Colander, David Colander

• Publisher: Routledge
• Year: 2000
• ISBN:
  9781134785087
• Edition: 1
• Topic: Negocios y empresa
• Subtopic: Negocios en general

A new approach to science has recently developed. It is called the complexity approach. A number of researchers, such as Brian Arthur and Buz Brock, have used this approach to consider issues in economics. This volume considers the complexity approach to economics from a history of thought and methodological perspectives. It finds that the ideas un

Part I: INTRODUCTION TO COMPLEXITY AND THE HISTORY OF ECONOMIC THOUGHT

1: WHAT IS COMPLEXITY?

James Wible¹

Economic phenomena are complex social phenomena. In today’s intellectual milieu, this is hardly a novel statement. Dynamic and complex systems theories are now part of the intellectual landscape of the contemporary natural and social sciences. In economics, the mathematics of complex, dynamic systems modeling is now being developed for application to the problems of economic growth, increasing returns, and general equilibrium. Rather than exploring contemporary research in complexity in economics and some of the natural sciences in great detail, in this chapter the focus of attention will be on general aspects of a theory of complexity.² An account of complexity is provided which is broad enough to serve as a background for an exploration of various approaches and contributions to complexity theory in both the natural and social sciences.³ In the paragraphs which follow, the contributions of three mostly compatible conceptions of complexity will be explored. The ideas of natural scientists Gregoire Nicolis and Ilya Prigogine will be presented first, followed by an overview of F.A.Hayek’s theory of complexity, and the ideas of those at the Sante Fe Institute will be presented last.

Prigogine and Nicolis on complexity

In the last decade or so, several lines of scientific inquiry have coalesced to form an area of mathematical-scientific investigation now known as complex systems theory. One of its narrower sub-cases, chaos theory, also has gained widespread attention. As economists know, complexity theory is essentially a new stage of systems theory. Systems theory existed before the appearance of complexity theory.⁴ However, the idea of complexity has been around for a long time. Thus, in effect there may be older and newer versions of complexity theory. The newer complexity theory can be characterized best by exploring one of the major overviews by two prominent contributors to the field. Gregoire Nicolis and Ilya Prigogine (1989) have published an account of complexity in their book, Exploring Complexity. Their work, while it mostly focuses on the natural sciences, also provides some applications to society. In the natural sciences, their survey includes physics, chemistry, biology, materials science, climatic change, cosmology, information theory, and insect societies. They also consider information theory and self-organization in human systems.

Since the end of the nineteenth century, Nicolis and Prigogine maintain that complexity theory has dramatically revised our vision of science and the universe. Essentially, they claim that complexity theory is in the process of revolutionizing science, not suddenly, but over many decades. The boldness of their claim invites extensive quotation. In the prologue to their work, they portray the last century of science in the following way:

The two great revolutions in physics at the beginning of the century were quantum mechanics and relativity. Both started as corrections to classical mechanics…Today, both of these areas of physics have taken an unexpected “temporal” turn: Quantum mechanics now deals in its most interesting part with the description of unstable particles and their mutual transformations. Relativity, which started as a geometrical theory, today is mainly associated with the thermal history of the universe…

The history of science during the three centuries that followed the Newtonian synthesis is a dramatic story indeed. There were moments when the program of classical science seemed near completion: a fundamental level, which would be the carrier of deterministic and reversible laws, seemed in sight. However, at each such moment something invariably did not work out as anticipated…Today, wherever we look, we find evolution, diversification, and instabilities…

At the beginning of this century, continuing the tradition of the classical research program, physicists were almost unanimous in agreeing that the fundamental laws of the universe were deterministic and reversible. Processes that did not fit this scheme were taken to be exceptions, merely artifacts due to complexity, which itself had to be accounted for by invoking our ignorance, or lack of control of the variables involved. Now, at the end of this century, more and more scientists have come to think, as we do, that many fundamental processes shaping nature are irreversible and stochastic; that the deterministic and reversible laws describing the elementary interactions may not be telling the whole story. This leads to a new vision of matter, one no longer passive, as described in the mechanical world view, but associated with spontaneous activity. This change is so deep that we believe that we can truly speak of a new dialogue of man with nature.

(Nicolis and Prigogine 1989:2–3)

Although Nicolis and Prigogine profess to offer no simple or precise definition of complexity or complex systems, they begin with the following idea. They begin with a suggestion that complexity implies:

a pluralistic view of the physical world, a view that can accommodate a world in which different kinds of phenomena coexist side by side as the conditions to which a given system is submitted are varied.

(Ibid.: 5–6)

They continue in the next paragraph elaborating a view of the physical world as composed of many systems and sub-systems rather than one grand unified order or system of order. The obvious contrast would be to compare the pluralism and multiplicity of systems of complexity theory with the reductionism of Newtonian mechanics. Metaphorically, the Newtonian theory has been portrayed as implying that the world is one vast clockwork mechanism. And if this clock runs perfectly and if its laws of motion are known with extreme precision, then the future of the world can be predicted with unfailing accuracy indefinitely into the future.

Another key aspect of complexity which is not as clearly highlighted as the side-by-side pluralism of systems in Nicolis and Prigogine is the idea of levels of systems. They portray multiple systems of complexity in the physical, biological, and social worlds and only inadvertently imply that systems may be subject to a more general principle of organization into what one might call levels of ontological complexity. The very vocabulary of the specialized sciences such as physics, biology, chemistry, and the social sciences suggests a notion of levels of partially autonomous systems of phenomena. Furthermore, within each level there may be many partially autonomous systems which function side by side. For example, the various sub-systems and organs which make up the human body are sufficiently autonomous to be identified as separate entities. Yet they all function side by side in organismic unity within that very complex entity known as the human body. The human body exists within complex biological and social systems which must provide enough resources to sustain life itself; and for life beyond survival, there must be a whole realm of a great variety of social systems, processes, and institutions for human civilization to have evolved into its present configuration.

Other than systems existing side by side and on many different levels, another important feature for complex systems is their ability to self-organize. Self-organization involves the creation of new patterns of activity by the entities in the system. A system such as a thermal system may go from a homogeneous state of uniform temperature to one of intermeshing circles of convection. The pattern of convection can be quite complicated relative to that of a gas without any temperature variation. Two different patterns of rotation are possible and the pattern of convection which occurs is quite unpredictable. Besides self-organization, according to Nicolis and Prigogine (1989:26), complex systems also are characterized by bi-stability and hysteresis, by oscillation, and by spatial patterns. Thermal convection exhibits all of these features.

Self-organizing complex systems are by their very nature dynamic systems. A dynamic system is one which endures and evolves over time. If there is change in the system, important aspects of such change can be described by dynamic equations which capture the laws of motion of the system evolving through time. There are conservative and dissipative dynamic systems. Conservative systems have something significant which remains constant or invariant throughout the period of change. Dissipative systems typically have no conservation rules or principles. Conservative systems are reversible in time while dissipative systems are irreversible. The best-known conservative system is Newtonian mechanics, which conserves total energy, total translational momentum, and total angular momentum. One of the simplest forms of dissipative processes is friction. Friction is a form of resistance which eventually affects the functioning of a Newtonian mechanical system. Other than friction, heat conduction and diffusion are dissipative and so are social processes and many natural and biological processes.

Complex dynamic systems require new conceptions of order besides equilibrium. In a mechanical system, equilibrium is attained when all of the forces in the system sum to zero. There is either no motion or the motion occurs in a path that is reversible in time. The motion of the planets around the sun is an example of time-reversible motion. In a non-mechanical system, an order or pattern is defined as a steady state, a macroscopic property of an entire system which is not a property of the constituent entities of the system. The thermal steady state does not imply that the individual molecules of a gas are motionless or that their movement is patterned through time. Motion of an individual particle may be quite random. A thermal steady state refers to the uniform mixing of the particles in a gas mixture. The very use of microscopic and macroscopic terminology such as equilibrium and steady state implies a distinction between at least two levels of functioning for thermal systems.⁵

It is also possible for complex systems to be in states far from equilibrium. Such states may persist indefinitely and are called states of non-equilibrium. Non-equilibrium states are those in which self-organizing new patterns may emerge. Complex interlocking circles of thermal convection are an example of self-organizing patterns in non-equilibrium. Nicolis and Prigogine suggest that the universe, since the moment of creation, is a self-organizing structure in a sustained non-equilibrium pattern of evolution. Other non-equilibrium processes of self-organization are all biological organisms, life itself, climate and the weather, most social processes, and the origin and continued evolution of the universe on a cosmic scale. Non-equilibrium is pervasive and seems to be the more general case than mechanical equilibrium, which appears in quite limited and very rigid contexts.

One aspect of complex systems which has garnered a great deal of attention is chaos.⁶ Chaotic phenomena mostly occur in dissipative systems and can be described in part as the appearance of turbulence in the system. Weather and climate are two good instances of thermal turbulence. The movement of some particles may show chaos. Chaotic trajectories would be the time paths of entities in phase space when a system is confronted with a sizable perturbation. Chaotic, turbulent behavior is irregular and non-periodic. Often it is both patterned and unpredictable. One very simple model of chaos, the Rössler model, gives rise to unpatterned rotational chaos. This model describes the movement of a particle in three dimensions around an unstable focus using three dynamic equations. Two of the equations are linear and the third is non-linear.⁷ The variable rotational path of the equations of evolution of the system means that the behavior of the system is non-periodic. Furthermore, the type of chaos depends on the initial conditions of the system. If the system is started on one side of the vertical plane through the origin, screw chaos appears. If started on the other side, the system exhibits spiral chaos.

Hayek’s theory of complex phenomena

Before complexity theory of Nicolis and Prigogine, there was another relatively brief, explicit account of a world with complexity. In the mid-1960s, Hayek (1967a, 1967b, and 1967c) wrote several essays on complexity. Perhaps the most important idea in Hayek’s “Theory of Complex Phenomena” is the conception of different levels of existence. Hayek portrays the world and many of the entities and processes in that world as complexities of many levels:

It has occasionally been questioned whether the phenomena of life, of mind, and of society are really more complex than those of the physical world. This seems largely due to a confusion between the degree of complexity of a peculiar kind of phenomenon and the degree of complexity to which, by combination of elements, any kind of phenomenon can be built up. Of course, in this manner physical phenomena may achieve any degree of complexity.

(Hayek 1967a:25–26)

In a long footnote to this passage, Hayek relates the idea of complexity to von Neumann’s work. In that note, Hayek suggests one way of imagining the number of entities in any level of order. He describes the levels in exponential terms:

It may be useful to give here a few illustrations of the orders of magnitude with which biology and neurology have to deal. While the total number of electrons in the Universe has been estimated at 10⁷⁹ and the number of electrons and protons at 10¹⁰⁰, there are in chromosomes with 1,000 locations [genes] with 10 allelomorphs 10¹⁰⁰⁰ possible combinations; and the number of possible proteins is estimated at 10²⁷⁰⁰.

(Ibid.: 25, n. 8)

For complex phenomena, Hayek maintains that it is difficult to predict individual events. Instead he argues that many events must be observed. Often the events being observed form a pattern. It is the pattern which is potentially predictable rather than individual events. For many patterns, prediction may give way to a weaker criterion. At most what may be achieved is an explanation of the principle or the central factors which give a systemic pattern its characteristic appearance. Hayek is also critical of statistics. Essentially, he maintains that statistics focuses on the wrong level of analysis. Statistics focuses on classes of phenomena at a lower level of explanation than the one on which the pattern appears. Consequently, knowledge of the structure generating a pattern of complex phenomena could be lost through a preoccupation with a-theoretical statistics.

As an example of complex patterns and his theory of complexity, Hayek (1967a:31) chooses Darwinian evolution. With the theory of evolution, an understanding of the growth and functioning of various classes of organisms only can be understood in patterns above the level of the individual creature. The theory deals with “pattern building forces” and not with conditions which determine the condition of specific individual entities. Because the appearance of new creatures cannot be predicted, evolution does not meet the narrow, orthodox criterion of prediction if it is taken as a hallmark of the scientific method. There are possible courses of events which the theory forbids, so that aspects of the theory of evolution are conceivably falsifiable in Hayek’s view.

For Hayek, pattern prediction is what is mostly possible at higher levels such as those of mind and society. Social structures are patterns of regularities which appear among the individuals in society. Again, it is difficult to translate knowledge of the patterns of social structures into predictions of specific events as more positivistic conceptions of science might require. For Hayek, economics and linguistics are two social sciences which have succeeded in building up coherent theories of the patterns of social phe...