Complexity

A Philosophical Overview

Nicholas Rescher

• Publisher: Routledge
• Year: 2020
• ISBN:
  9781000675191
• Edition: 1
• Topic: Filosofia
• Subtopic: Storia e teoria della filosofia

Our world is enormously sophisticated and nature's complexity is literally inexhaustible. As a result, projects to describe and explain natural science can never be completed. This volume explores the nature of complexity and considers its bearing on our world and how we manage our affairs within it.
Rescher's overall lesson is that the management of our affairs within a socially, technologically, and cognitively complex environment is plagued with vast management problems and risks of mishap. In primitive societies, failure to understand how things work can endanger a family or, at worst, a clan or tribe. In the modern world, man-made catastrophes on the model of Chernobyl can endanger millions, possibly even risking the totality of human life on our planet. Rescher explains "technological escalation" as a sort of arms race against nature in which scientific progress requires more powerful technology for observation and experimentation, and, conversely, scientific progress requires the continual enhancement of technology. The increasing complexity of science and technology (and, in consequence, of social systems) along with problems growing faster than solutions confront us with major management and decision problems.
This study is the first of its kind. There have been many specialized studies of complexity in physics and computation theory, but no overall analysis of the phenomenon. Although Rescher offers a sobering outlook, he also believes that complexity entails mixed blessings: our imperfect knowledge provides a rationale for putting forth our best efforts. Rescher urges us to gear the conduct of life's practical affairs to the demands of a complex world. This highly readable and accessible volume will be of interest to those interested in philosophy, the philosophy of science, science policy studies, and future studies.

1
The Ways of Complexity

(1) A system’s complexity is a matter of the quantity and variety of its constituent elements and of the interrelational elaborateness of their organizational and operational make-up. (2) Ontological complexity accordingly has three main aspects: the compositional, the structural, and the functional. (3) As an item’s complexity increases, so do the cognitive requisites for its adequate comprehension, although, of course, cognitive ineptitude and mismanagement can manage to complicate even simple issues. All the same, our best practical index of an item’s complexity is the effort that has to be expended in coming to cognitive terms with it in matters of description and explanation. And this means that complexity can in principle make itself felt in any domain whatsoever.

1. What Complexity Involves

The world’s complexity is a fact of life that has profound and far-reaching implications for us. Complexity is first and foremost a matter of the number and variety of an item’s constituent elements and of the elaborateness of their interrelational structure, be it organizational or operational. Any sort of system or process—anything that is a structured whole consisting of interrelated parts—will be to some extent complex. Accordingly, all manner of things can be more or less complex: natural objects (plants or river systems), physical artifacts (watches or sailboats), mind-engendered processes (languages or instructions), bodies of knowledge, and so on. In greater or lesser degree, complexity is present throughout the domain of the real. And complexity appertains to the realm of fiction as well. Even as a “simple” person is naive and unsophisticated, so a “complex” character in a novel or drama will be one whose actions are not easy to explain and whose described motivations are varied and convoluted.

Some writers identify complexity with chaos. But they are mistaken because there are other, equally effective routes to complexity. Take holistic systems, for example. In chaotic systems we cannot predict the properties of wholes on the basis of details relating to their constituents because (1) we cannot determine (observe) very fine-grained details, and (2) these fine-grained details matter crucially for the large-scale results that emerge from them. In holistic system, by contrast, we cannot determine the properties of units on the basis of the details relating to their constituents because the constituents do not determine the properties of those wholes—their features emerge in autonomous independence of the behavior of the parts.¹ Thus, for example, nothing in the psychology of individuals enables us to predict various features of the social aggregates they constitute—the suicide rate, for example, or the rate of money circulation. When such holistic principles concern us we must study the relevant whole at large; there is no prospect of deriving the large-scale parameters that interest us from information about the comportment of individuals. With both sorts of systems, problem solving at the larger details of scale becomes a project separate from and additional to that of problem solving at the lesser levels. Both sorts of systems accordingly take a step in the direction of greater operational complexity. For example, the complexity of stories is holistic. It resides in the volume of their events, the intricacy of their plots, and the interweaving of the relationships among their characters: chaos, as such, has nothing to do with it.

The salient fact of the matter is that the modes of complexity are multiple. The physicist Seth L. Loyal has computed an inventory of definitions of complexity—perhaps “standards” would be better. His list includes: information (Shannon); entropy (Gibbs, Boltzman); algorithmic complexity; algorithmic information; Renyi entropy; self-delimiting code length (Huffman, Shannon-Fano); error-correcting code length (Hamming); Chernoff information; minimum description length (Rissanen); number of parameters, or degrees of freedom, or dimensions; Lempel-Ziv complexity; mutual information, or channel capacity; algorithmic mutual information; correlation; stored information (Shaw); conditional information; conditional algorithmic information content; metric entropy; factual dimension; self-similarity; stochastic complexity (Rissanen); sophistication (Koppel, Atlan); topological machine size (Crutchfield); effective or ideal complexity (Gell-Mann); hierarchical complexity (Simon); tree subgraph diversity (Huberman, Hogg); homogeneous complexity (Teich, Mahler); time computations complexity; space computations complexity; information-based complexity (Traub); logical depth (Bennett); thermodynamic depth (Lloyd, Pagels); grammatical complexity (position in Chomsky hierarchy); Kullbach-Liebler information; distinguishability (Wooters, Caves, Fisher); Fisher distance; discriminability (Zee); information distance (Shannon); algorithmic information distance (Zurek); Hamming distance; long-range order; self-organization; complex adaptive systems; edge of chaos.² The possibilities are vast.

In a paper published more than a century ago, C. S. Peirce concluded “that there is probably in nature some agency by which the complexity and diversity of things can be increased.”³ For Peirce, this idea of the world’s ever unfolding complexity was closely linked to nature’s impetus to novelty: He held that as natural systems evolve they ongoingly develop new features that require ever more elaborate descriptive specification⁴—a condition of things exhibited in the emergence of new law levels in the course of cosmic evolution (as per the sequence of physics, chemistry, biology, sociology, etc.). And there is much to be said for such a view of the matter. In nature, certain factors—energy, matter, life, and complexity among them—appear to be self-potentiating: the more of them there is, the more powerful the impetus to the production of yet more. Left to its own devices without let or hindrance, such a tendency is liable to result in exponential growth,⁵ a condition illustrated in the initial phase of a big bang universe and that of the expansion of life through biological evolution in the organic realm. This sort of thing holds for complexity also,⁶ seeing that the operations of a complex system tend by their very nature to make for yet further complexity (whose management, cognitive or otherwise, becomes increasingly difficult⁷). Moreover, we encounter this phenomenon in the social realm as well as in the physical. One striking instance is the hypercomplexity of laws and regulations in the present era—an object of widespread complaint regarding which numerous proposals for remedy are astir, invariably to no avail.⁸ And similarly, complexity growth also characterizes the domain of human creativity—in art and literature, for example, which knows virtually no limits.

But why is it that the realms of nature and of artifact are so complex?

Four principal modes of explanation have been proposed here: the intelligent design theory, the inherent teleology theory, the chance-plus-self-perpetuation theory, and the automatic self-potentiation theory. Each of them deserves at least brief consideration.

The Intelligent Design Theory effectively represents the standpoint of traditional theology. It sees complexity as the work of intelligence, and views the world as the theater of operation of a powerful intelligent agency—a creative world-mind or world-spirit, if one likes, but one might as well call it God. This agency guides the course of evolution—physical and biological alike—in the direction of increasing functional sophistication. Complexity enhancement is thus impressed upon nature “from on high” so the speak.

With regard to design in biological evolution, the most sophisticated current exponent of this sort of position is Michael J. Behe, whose recent book reasons essentially as follows: Complexity in nature—and in particular in the biological realm—is best accounted for (or perhaps even can only be accounted for) by supposing the operation of an intelligent designer.⁹ A naturalistic developmental theory—such as Darwinian evolution by natural selection—encounters difficulties when it comes to explaining the development of the cell, seeing that many cellular systems are what can be characterized as “irreducibly complex.” That is, such systems need several duly coordinated components before they can function properly. A mousetrap, for example, built of several pieces (platform, hammer, spring, and so on). Such a system probably cannot be put together in the Darwinian evolutionary manner of improving its function bit by bit, one piece at a time. One cannot catch a mouse with just the platform and then catch a few more by adding the spring. All the pieces have to be functioning well in their duly coordinated collaborative roles before the trap can catch any mice at all. Whenever we encounter such interactive systems, we can (and should) assume that they are the products of intelligent activity. And this reasoning—so it is contended—should be extended to cellular systems apart from intelligence. The thesis is that we simply know of no other mechanism, including Darwin’s, which produces such operational complexity among collaborative components. Such, in essentials, is Behe’s position with regard to intelligence’s role in biological evolution. And the Intelligent Design Theory expands this position regarding the basis of complexity across a greater landscape.

The Inherent Teleology Theory sees nature itself as the bearer of an intrinsic, self-engendered penchant towards increasing complexity. The most dedicated philosophical exponent of this position is C. S. Peirce, who put the matter as follows in one of the many discussions on this theme:

Evolution means nothing but growth in the widest sense of that word. Reproduction, of course, is merely one of the incidents of growth. And what is growth? Not mere increase. Spencer says it is the passage from the homogeneous to the heterogeneous—or, if we prefer English to Spencerese—diversification. That is certainly an important factor of it. Spencer further says that it is certainly an important factor of it. Spencer further says that it is a passage from the unorganized to the organized; but that part of the definition is so obscure that I will leave it aside for the present. But think what an astonishing idea this of diversification is! Is there such thing in nature as increase of variety? Were things simpler, was variety less in the original nebula from which the solar system is supposed to have grown than it is now when the land and sea swarms with animal and vegetable forms with their intricate anatomies and still more wonderful economies? It would seem as if there were an increase in variety, would it not? And yet mechanical law, which the scientific infallibilist tells us is the only agency of nature, mechanical law can never produce diversification. That is a mathematical truth—a proposition of analytical mechanics; and anybody can see without any algebraical apparatus that mechanical law out of like antecedence can only produce like consequents. It is the very idea of [mechanical] law. So if observed facts point to real growth, they point to another agency, to a spontaneity for which infallibilism provides no pigeonhole.¹⁰

On such a view, growth, diversification, and complexification are simply intrinsic developmental tendencies in nature. The physical universe is seen as autoteleological in this regard, propelling itself through its own resources into ever greater complexity.

The Chance-Plus-Self-Perpetuation Theory is a third major approach to the explanation of increasing complexity. It takes the line that (1) the merely chance fluctuation of things occasionally brings manifolds of greater-than-prevailing complexity into being, and that (2) complexity once present tends to be self-perpetuating. Such a situation makes for a drift towards an ever increasing complexity, on its basis chance ongoingly engenders new complexity which is henceforth continued through complexity’s capacity to maintain its new gains over time. Such a view of complexity enhancement is entirely naturalistic and nonteleogical. It contemplates a supranatural intelligence nor an intranatural teleology to provide for its explanatory mechanism, but rests content with the purely natural processes of chance fluctuation and self-maintenance. (Perhaps the main obstacle that such an otherwise attractive theory faces is that of the question whether the actual historical course of complexity-enhancement in nature may not have been too rapid to be explained satisfactorily by the operations of mere chance.)

The Self-Potentiation Theory turns on the idea that complexity is in its own inherent nature self-propagating—that even as intelligence, for example, pushes itself to higher levels through the impetus of its own operation, so also does complexity. Complexity is by nature expansive; once present it clambers up upon itself to achieve yet higher levels of complexity. In this way, complexification can develop “from the bottom up.” Sentences make up books, books make up libraries, and so on. But such productivity is a two-way street that also goes “from the top down.” Where there are books, the production of further sentences is greatly facilitated. Where there are libraries, books come to multiply. (Indeed without libraries many sorts of books—catalogues, indices, citation indices, etc.—would not exist at all.) Even as increases in power in personal or institutional affairs set the stage for yet further increases in power—at any rate up to some point fixed by the context of the case—so increases in complexity function in an analogous manner. More complex systems generally not only permit but actually demand the addition of yet further complexities. As is illustrated by the evolution of organisms and of machines—be they mechanical or electronic—increases in complexity engender yet further increases.¹¹ And the theory of complexity self-potentiation sees this as a universal phenomenon.

These, then, are the principal approaches to the explanation of complexity enhancement in nature. But irrespective of the mechanism, the upshot is sub...