MATERIAL COMPLEXITY

Manuel DeLanda

For centuries the scientific study of the behavior of materials was performed in rela tively obscure engineering or metallurgy departments, away from the highly presti gious centers of 'pure' science. For physicists, it seems, the only property of materials that carried any weight was their mass; their strength was relegated to the more peripheral fields of applied science.Today this has all changed and the study of the complex behavior of matter has acquired legitimacy and its own name: materials sci ence and engineering.The long institutional struggle to gain respect, and the more sophisticated conceptual and technical tools that may be brought to bear on the study of material complexity, are full of lessons for the philosopher of science. Materialist philosophers, it is becoming increasingly clear, cannot afford to ignore the basic fact that the study of matter does matter. Cyril Stanley Smith, a metallurgist and historian of materials, has explored the devel opment of the philosophy of matter in the West and has concluded that for the most part the study of the complexity and variability of behavior of materials has always been the concern of empirically oriented craftsmen or engineers, not of philosophers or scientists.He argues that we may have inherited this generally condescending atti tude from the Greeks, who admired the handcrafted products produced by the black smith but who despised his material activities and his apparent unwillingness to engage in verbal exchanges on political or philosophical questions. As he puts it To those engaged in materials production and fabrication, it may be disconcerting to realize that for a fair fraction of human history their activities have been viewed with suspicion and downright distaste by social thinkers and the general public. The ancient Greek philosophers, who set the tone for many of the attitudes still prevalent throughout Western civilization, regarded those involved in the produc tion of material goods as being less worthy than agriculturalists and others who did not perform such mundane tasks ... Throughout ancient society the most menial tasks, especially those of mining and metallurgy, were left to slaves. Hence the common social attitude of antiquity, persisting to this day in some intellectual circles, was to look down upon those who work with their hands. Xenophon stated the case in this fashion, 'What are called the mechanical arts carry a social stigma and are rightly dishonored in our cities.For these arts dam age the bodies of those who work at them or who act as overseers, by com pelling them to a sedentary life and to an indoor life, and, in some cases, to spend the whole day by the fire.This physical degeneration results also in deterioration of the soul.Furthermore the workers at these trades simply do not have the time to perform the offices of friendship or citizenship. Consequently they are looked

14 Digital Tectonics

upon as bad friends and bad patriots, and in some cities, especially the warlike ones, it is not legal for a citizen to ply a mechanical trade. " Despite this negative attitude, Smith argues that Greek philosophers like Aristotle may have learned much from visiting workshops, given that practically everything about the behavior of metals and alloys that could be explored with pre-industrial technology was already known to craftsmen and blacksmiths for at least a thousand years. Indeed, as he says, the early philosophies of matter may have been derived from observation and conversation with those ' whose eyes had seen and whose fingers had felt the intricacies of the behavior of materials during thermal processing or as they were shaped by chip ping, cutting or plastic deformation'.2 For instance, Aristotle's famous four elements fire, earth, water and air - may be said to reflect an awareness of what today we know as energy and the three main states of aggregation of matter: the solid, liquid and gas states, all of which were very familiar to a metallurgist. As metaphysical speculation gave special meanings to these four elementary quali ties their original physical meaning was lost, and the variability and complexity of real materials was replaced with the uniform behavior of a philosophically simplified matter about which one could only speculate symbolically. It is true that sixteenth-century alchemists recovered a certain respect for a direct interaction with matter and energy, and that seventeenth-century Cartesian philosophers speculated intensely about the variable properties of different ways of aggregating material components. But these early attempts at capturing the complexity of physical transmutations and of the effect of physical structure on the complex properties of materials, eventually lost to the emer gent science of chemistry and its almost total concentration on simple behavior: that of individual components (such as Lavoisier's oxygen) or of substances that conform to the law of definite proportions (as in Dalton's atomic theory). There was, as Smith observes, an 'immense gain' in these simplifications, since the exact sciences could not have developed without them, but the triumph of chemistry was accompanied by a 'not insignificant loss'. In particular, the complete concentration of analysis at the level of molecules caused an almost total disregard for higher levels of aggregation in solids, but it is there where most complex properties of interest to today's material scientist occur.3 As is usual in the history of science there were several exceptions, such as Galileo, who studied the strength of materials in the sixteenth cen tury, but who may have derived his interest and even some insights from his visits to the Venetian arsenal, the largest military-industrial complex of its time and home to a large variety of craftsmen.How are we to theorise this ambivalent relation towards com plex materiality within science? Deleuze and Guattari offer a possible solution when they contrast two types of science, or two modes of conducting scientific research, a major and a minor mode: royal science and nomad science, the science of the royal societies and academies at the service of the state preoccupied above all with the discovery of abstract general laws, and the humbler science of those who built the laboratory instru ments and had the job of testing the validity of those laws in concrete physical situations.

Material Complexity 15

Indeed, the distinction between royal and nomad science is drawn more widely so that it does not coincide with the distinction between pure and applied science.In its minor mode, science deals with complex material behavior, liquids not solids, heterogeneous not homogeneous matter, turbulent not steady-state (or laminar) flow.(It also includes a preference for non-metric, projective or topological spaces, as well as for problematic rather than axiomatic logical structures.) 4 Not surprisingly, Deleuze and Guattari classify ancient metallurgy as 'minor science in person'.5 Although the distinction between the major and minor modes does not neatly divide individual scientists into two sharply drawn categories, we may illustrate it with two his torical characters: Isaac Newton and his contemporary archenemy Robert Hooke, who developed the first theory of material elasticity.As materials scientist James Edward Gordon has remarked, 'Unlike Newton, Hooke was intensely interested in what went on in kitchens, dockyards, and buildings - the mundane mechanical arenas of life ... Nor did Hooke despise craftsmen, and he probably got the inspiration for at least some of his ideas from his friend the great London clock maker Thomas Tompion
. . .'.6

The point

Gordon is trying to make is not that scientists may be divided into two separate classes since, after all, the same Newton who headed the Royal Society had also been an alchemist, but that in seventeenth-century England much more prestige was attached to scientific fields that were not concerned with the mundane mechanical arenas where materials displayed their full complex behavior. This may be one reason why conceptual advances in the study of materials, such as the key conceptual distinction between stress and strain (one referring to the forces acting on a material structure, the other to the behavior of the structure in response to those forces), were made in France where applied science was encouraged both officially and socially. Indeed, although Hooke is a perfect example of a minor scientist - his law of elasticity linking stress and strain (or more accurately, since these concepts did not yet exist, the load a structure bears and the deformation it undergoes under that load) was of the type characterising major science. In other words, his law postulated a linear relation between load and deformation, or what amounts to the same thing: a simple, well behaved, proportional response to a given cause.Although some materials do have a range of loads under which they respond linearly (a small increase in load producing a proportionally small deformation) and also reversibly (after removal of the load the defor mation disappears), even these simple materials display a critical threshold beyond which their behavior ceases to be elastic and becomes plastic: any further small load causes a large deformation and, moreover, the change of shape becomes irreversible. Plastic behavior, such as a permanent dent or bend in a metallic beam, is but one example of nonlinear behavior. Indeed, many materials behave nonlinearly even without critical loads. Organic tissues, for example, display a J-shaped reaction curve: a small load causes a large deformation at first, but then even large loads cease to have much effect. Rubber and other materials display an S-shaped reaction: a load fails to have any effect at all up to a point beyond which rubber stretches linearly but only to stop react ing to further loads beyond yet another point. Since a material's capacity to bear loads is directly related to its capacity to deform, rubber's inability to further deform under heavy loads makes it very brittle in those conditions. At any event, J- and S-shaped reaction curves, as well as many others, are examples of nonlinear, complex material behavior.7 Material complexity, however, needs more than just nonlinearity to become expressed.In particular, one may linearize a nonlinear system by studying it only under

16 Digital Tectonics

conditions near or at equilibrium. One of the oldest examples of this linearization, going back to Galileo, is the simple pendulum.The mathematical model of a pendulum shows that the relationship between the amplitude and the period of its swing is in fact non linear, that is, there is feedback or interaction between the two. But given that nonlinear equations were pretty much intractable until the advent of computers, this physical fact presented an obstacle to the modeling of this simple dynamical system.So what scientists did was to study the pendulum's behavior only for extremely small values of its amplitude, so as not to let the nonlinearities become too visible. As mathematician Ian Stewart puts it: Classical mathematics concentrated on linear equations for a sound pragmatic reason: it could not solve anything else ... So docile are linear equations that classical mathematicians were willing to compromise their physics to get them. So the classical theory deals with shallow waves, low-amplitude vibrations, small temperature gradients [that is, it linearizes nonlinearitiesl. So ingrained became the linear habit that by the 1940s and 1950s many scientists and engineers knew little else ...Linearity is a trap.The behavior of linear equations .. .is far from typical. But if you decide that only linear equations are worth thinking about, self censorship sets in.Your textbooks fill with triumphs of linear analysis, its failures buried so deep that the graves go unmarked and the existence of the graves goes unremarked.As the eighteenth century believed in a clockwork world, so did the mid-twentieth in a linear one.s Thus, nonlinear behavior needs non-equilibrium conditions to become manifest. Today's complexity theory depends as much on nonlinear mathematics as it does on far-from-equilibrium thermodynamics. The latter discipline studies systems which, unlike its nineteenth-century classical counterpart, are not closed to intense flows of matter and energy from the outside.While linear systems tend to be characterized by a single, global stable state, systems which are both nonlinear and non-equilibrium dis play multiple stable states and these may come in a variety of additional forms, not only steady-state but also periodic and chaotic.In addition, scientists have come to realize that these multiple stable states seem to characterize not only inorganic material behavior, but also organic and even social behavior.In other words, we are beginning to understand that any complex system, whether composed of interacting molecules, organic creatures or economic agents, is capable of spontaneously generating order and actively organizing itself into new structures and forms.It is precisely this ability of matter and energy to self-organize that is of greatest significance to the philosopher. This can be illustrated by an example from materials science. Long ago, practical metallurgists understood that a given piece of metal could be made to change its behavior, from ductile and tough to rigid and brittle, by hammering it while cold.The opposite transmutation, from hard to ductile, could also be achieved by heating the piece of metal again and then allowing it to cool down slowly (that is, by annealing it).Yet, although blacksmiths knew empirically how to cause these metamor phoses, it was not until a few decades ago that scientists understood the actual micro scopic mechanism.As it turns out, explaining the physical basis of ductility involved a radical conceptual change: scientists had to stop viewing metals in static terms, that is, as deriving their strength in a simple way from the chemical bonds between their com posing atoms, and begin looking at them as dynamical systems. In particular, the real cause of brittleness in rigid materials, and the reason why ductile ones can resist being

Material Complexity 17

broken, has to do with the complex dynamics of spreading cracks. A crack or fracture needs energy to spread through a piece of material and any mech anism that takes away energy from the crack will make the material tough. In metals, the mechanism seems to be based on certain line defects or imperfections within the component crystals, called dislocations. Dislocations not only trap energy locally but, moreover, are highly mobile and may be brought into existence in large quantities by the very concentrations of stress which tend to break a piece of metal. Roughly, if popula tions of these line defects are free to move in a material they will endow it with the capacity to yield locally without breaking, that is, they will make the material tough.On the other hand, restricted movement of dislocations will result in a more rigid material.9 Both of these properties may be desirable for different tools, and even within one and the same tool: in a sword or knife, for instance, the load-bearing body must be tough while the cutting edge must be rigid to be capable of holding on to its sharply triangular shape. What matters from the philosophical point of view is precisely that toughness or rigid ity are emergent properties of a metallic material that result from the complex dynami cal behavior of some of its components. An even deeper philosophical insight is related to the fact that the dynamics of populations of dislocations are very closely related to the population dynamics of very different entities, such as molecules in a rhythmic chemical reaction, termites in a nest-building colony, and perhaps even human agents in a market.In other words, despite the great difference in the nature and behavior of the components, a given population of interacting entities will tend to display similar collec tive behavior as long as the interactions are nonlinear and as long as the population in question operates far from a thermodynamic equilibrium. For materials scientists this commonality of behavior is of direct practical significance since it means that as they begin to confront increasingly more complex material properties, they can make use of tools coming from nonlinear dynamics and non-equilibrium thermodynamics, tools that may have been developed to deal with completely different problems. In the words of one author: . . . during the last years the whole field of materials science and related tech nologies has experienced a complete renewal. Effectively, by using techniques corresponding to strong non-equilibrium conditions, it is now possible to escape from the constraints of equilibrium thermodynamics and to process totally new material structures including different types of glasses, nano- and quasi-crystals, superlattices . . . As materials with increased resistance to fatigue and fracture are sought for actual applications, a fundamental understanding of the collective behavior of dislocations and point defects is highly desirable. Since the usual thermodynamic and mechanical concepts are not adapted to describe those sit uations, progress in this direction should be related to the explicit use of genuine non-equilibrium techniques, nonlinear dynamics and instability theory.'o An understanding of background is important in bringing the two strands of the argu ment together. The contemporary science of materials is an offspring of World War I I and the Manhattan Project. While prior to the war the field constituted a collage of minor sciences, engineers and metallurgists who had participated in wartime government pro jects finally unified and gave this discipline the prestige that it deserved. The study of material complexity is now the rule, and a new awareness of the self-organizing capa cities of matter is beginning to emerge in this field. In its more prestigious counterpart, royal or major science, on the other hand, the focus on linear and equilibrium behavior

18 Digital Tectonics

has led to a view of matter as an inert receptacle for forms imposed from the outside, a view with many similarities to Creationism and Platonism. Gilles Deleuze refers to this conception of the genesis of form as 'the hylomorphic model'. Artisans, craftsmen, and minor scientists in general, he argues, always had a different conception of the relation between matter and form, at least implicitly: they did not impose but teased a form out of an active material, collaborating with it in the production of a final product rather than commanding it to obey and passively receive a previously defined form.As Deleuze and Guattari write, the hylomorphic model: . . . assumes a fixed form and a matter deemed homogeneous. It is the idea of the law that assures the model's coherence, since laws are what submits matter to this or that form, and conversely, realize in matter a given property deduced from the form . . . [But the] hylomorphic model leaves many things, active and affective, by the wayside. On the one hand, to the formed or formable matter we must add an entire energetic materiality in movement, carrying singularities .. . that are already like implicit forms that are topological, rather than geometrical, and that combine with processes of deformation: for example, the variable undulations and torsions of the fibers guiding the operations of splitting wood. On the other hand, to the essential properties of matter deriving from the formal essence we must add variable intensive affects, now resulting from the operation, now on the contrary, making it possible: for example, wood that is more or less porous, more or less elastic and resistant. At any rate, it is a question of surrendering to the wood, then following where it leads by connecting operations to a materiality instead of imposing a form upon a matter. " The term 'singularities' in this quote refers to the multiple stable states which charac terize nonlinear systems. These may be singular points representing endogenous ten
dencies

towards a steady state, or tendencies towards endogenous oscillations, simple

or turbulent, represented by periodic and chaotic singularities, respectively. Since these singularities represent the state a system will tend to adopt in the long run, or the final state towards which it is attracted, they are referred to as 'attractors'. Singularity may also refer to the bifurcations or critical points at which a given attractor changes into another, such as the well-studied Hopf bifurcation which turns a steady state attractor into a periodic one. The term 'affects', on the other hand, refers not to tendencies but to capacities, the capacity of a material to affect and be affected. Bearing loads, for example, involves the capacity to be affected by a load, in the sense that a load-bearing structure must be capable of stretching if the loads are in tension, or of shrink'lng if they are in compression. 12 Any material, no matter how simple its behavior, has endogenous tendencies and capacities, but Deleuze argues that if the material in question is homogeneous and closed to intense flows of energy, its singularities and affects will be so simple as to seem reducible to a linear law. In a sense, these materials hide from view the full repertoire of self-organizing capabilities of matter and energy. On the other hand, if the material is far from equilibrium (or, what amounts to the same thing, if differences in intensity are not allowed to be canceled) or if it is complex and heterogeneous (that is, if the differences among its components are not canceled through homogenization) the full set of singularities and affects will be revealed, and complex materiality will be allowed to manifest itself.In other words, the emphasis here is not only, on the spon taneous generation of form, but on the fact that this morphogenetic potential is best

Material Complexity 19

expressed, not by the simple and uniform behavior of materials, but by their complex and variable behavior. In this sense, contemporary industrial metals, such as mild steel, may not be the best illustration of this new philosophical conception of matter. While naturally occurring metals contain all kinds of impurities that change their mechanical behavior in different ways, steel and other industrial metals have undergone in the last two hundred years an intense process of uniformity and homogenization in both their chemical composition and their physical structure. The rationale behind this process was partly based on ques tions of reliability and quality control, but it had also a social component: both human workers and the materials they used needed to be disciplined and their behavior made predictable. Only then the full efficiencies and economies of scale of mass-production techniques could be realized. But this homogenization also affected the engineers that designed structures using these well-disciplined materials. In the words of James E. Gordon: The widespread use of steel for so many purposes in the modern world is only . partly due to technical causes. Steel, especially mild steel, might euphemistically be described as a material that facilitates the dilution of skills . . . Manufacturing processes can be broken down into many separate stages, each requiring a mini mum of skill or intelligence . . . At a higher mental level, the design process becomes a good deal easier and more foolproof by the use of a ductile, isotropic, and practically uniform material with which there is already a great deal of accu mulated experience. The design of many components, such as gear wheels, can be reduced to a routine that can be looked up in handbooks. 13 Gordon sees in the spread of the use of steel in the late nineteenth and early twentieth centuries, a double danger for the creativity of structural designers. The first danger is the idea that a single, universal material is good for all different kinds of structure, some of which may be supporting loads in compression, some in tension, some withstanding sheer stresses and others torsional stresses. But as Gordon points out, given that the roles which a structure may play can be highly heterogeneous, the repertoire of materials that a designer uses should reflect this complexity. On the other hand, he points out that much as in the case of biological materials like bone, new designs may involve struc tures with properties that are in continuous variation, with some portions of the struc ture better able to deal with compression while others deal with tension. Intrinsically heterogeneous materials, such as fiberglass and the newer hi-tech composites, afford designers this possibility. As Gordon says, 'it is scarcely practicable to tabulate elaborate sets of "typical mechanical properties" for the new composites.In theory, the whole point of such materials is that, unlike metals, they do not have "typical properties", because the material is designed to suit not only each individual structure, but each place in that structure. ' 14 This is not to imply that there are no legitimate roles to be played by homogeneous materials with simple and predictable behavior, such as bearing loads in compression. And similarly for the institutional and economic arrangements that were behind the quest for uniformity: the economies of scale achieved by routinizing production and some design tasks were certainly very significant. As with the already-mentioned homogenizations performed by scientists in their conceptions of matter, there were undoubtedly some gains. The question is, what got lost in the process? To give the most obvious example of a hidden cost, the nineteenth-century process of transferring skills

20 Digital Tectonics

from the human worker to the machine, and the task of homogenizing metallic behav ior went hand in hand. As Cyril Stanley Smith remarks, 'The craftsman can compensate for differences in the qualities of his material, for he can adjust the precise strength and pattern of application of his tools to the material's local vagaries. Conversely, the con stant motion of a machine requires constant materials."5 If it is true that much of the knowledge about the complex behavior of materials was developed outside science by empirically oriented individuals, the de-skilling of craftsmen that accompanied mecha nization may be seen as involving a loss of at least part of that knowledge, since in many cases empirical know-how is stored in the form of skills. Additionally, not only the production process was routinized in this way; so, to a lesser extent, was the design process. Many professionals who design load-bearing structures have lost their ability to design with materials that are not isotropic, that is, that do not have identical properties in all directions. But it is precisely the ability to deal with com plex, continuously variable behavior that is now needed to design structures with the new composites. Hence, we may need to nurture once again our ability to deal with vari ation as a creative force, and to think of structures that incorporate heterogeneous ele ments as a challenge to be met by innovative design. To conclude, the historical processes of homogenization and routinization have pro moted the 'hylomorphic schema' as a paradigm of the genesis of form. Conversely, it is partly thanks to the new theories of self-organization that the potential complexity of behavior of even the humbler forms of matter-energy has been revealed. We may now be in a position to think about the origin of form and structure, not as something imposed from the outside on an inert matter, not as a hierarchical command from above as in an assembly line, but as something that may come from within the materials, a form that we tease out of those materials as we allow them to have their say in the structures we create.

Notes

1 Melvin Kranzberg and Cyril Stanley Smith, 'Materials in History and Society' in The Materials Revolution, ed Tom Forrester, (Cambridge, Mass: MIT Press, 1988). p. 93. 2 Cyril Stanley Smith, 'Matter Versus Materials: A Historical View', in A Search for Structure (Cambridge, Mass: MIT Press, 1992). p. 115. 3 Ibid, pp. 120-1. 4 Gilles Deleuze and Felix Guattari, A Thousand Plateaus (Minneapolis: University of Minnesota Press, 1980) pp. 361-2. 5 Ibid, p. 411. 6 James Edward Gordon, The Science of Structures and Materials (Scientific American Library, 1988) New York, p. 18.
7 Ibid, pp. 20-1.

8 Ian Stewart, Does God Play Dice? The Mathematics of Chaos (Oxford: Basil Blackwell, 1989). p. 83. 9 James Edward Gordon, The Science of Structures and Materials, p. 111. 10 D. Walgraef, 'Pattern Selection and Symmetry Competition in Materials Instabilities' in New Trends in Nonlinear Dynamics and Pattern-Forming Phenomena, eds Pierre Coullet and Patrick Huerre (Plenum Press, 1990) New York, p. 26. 11 Gilles Deleuze and Felix Guilttari, A Thousand Plateaus, p. 408. 12 See Manuel DeLanda, Intensive Science and Virtual Philosophy (London: Continuum Press, 2002) for a full discussion of singularities and affects. 13 James Edward Gordon, The Science of Structures and Materials, p. 135. 14 Ibid., p. 200. 15 Cyril Stanley Smith, 'Reflections on Technology and the Decorative Arts in the Nineteenth Century' in ' A Search for Structure, cit. p. 313.

Material Complexity 21