LIFE AND PHYSICS New Perspectives ILYA PRIGOGINE
Universit6 Libre de Bruxelles, Brussels, Belgium
INTRODUCTION The idea for a symposium on "The N e w Frontiers at the Crossing Between Physical and Life Sciences" comes at a time w h e n we are facing an entirely n e w situation in so far as these topics are concerned. I will focus my discussion here on the problem of the origins of life as it may be thought about using some newer conceptual tools of physics. As we all know, life is an evolutionary process. We are thus supposed to adopt an evolutionary paradigm in order to cope with biological time. It is n o w clear that physics had to go a long way to design an adequate conceptual frame: Classical physics stressed a quite atemporal description of p h e n o m e n a , and we are only in the first steps of the necessary reconceptualization. It should be emphasized that in classical physics all processes could in principle be reduced to the same laws of motion as those formulated by Newton: Once initial conditions are known, the outcome of every situation could in principle be predicted with absolute certainty. The two characteristics of this scheme that I would like to stress are as follows: The first corresponds to a deterministic description of nature: no place is left to any uncertainty. The second characteristic corresponds to a reversible description; Newton's laws do not imply any distinction between past and future. (1). This grandiose intellectual structure was not without problems. H o w to understand the position of life, of humankind, in a universe described as a giant automaton? H o w does one reconcile the two attitudes? Are h u m a n s outside the universe they describe? The famous French biologist Jacques Monod had 217
218
Prigogine
a way of making this contrast explicit by stating that the laws of biology were compatible with the laws of physics, but not included in them. In the classical version we inherited from the 17th century, nature appeared as passive, as submitted to the creative impulse of humankind. The history of m o d e r n culture is a dramatic story indeed. There were moments w h e n 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, something went wrong. The scheme had to be enlarged, and the fundamental level, in the sense of the classical program, remained elusive. Today, wherever we look we find evolution, diversification, and instabilities. A fundamental reconceptualization of science is going on. We have long k n o w n that we are living in a pluralistic world in which we find deterministic, as well as stochastic, phenomena; reversible, as well as irreversible processes. We observe deterministic phenomena, such as the frictionless p e n d u l u m or the trajectory of the moon around the earth; moreover, we know that the frictionless p e n d u l u m is also reversible. But other processes are irreversible, diffusion or many chemical reactions for instance: And we are obliged to acknowledge the existence of stochastic processes if we want to avoid the paradox of referring the variety of natural p h e n o m e n a to a program printed at the m o m e n t of the Big Bang. What has changed since the beginning of this century is our evaluation of the relative importance of these four types of phenomena. The artificial may be deterministic and reversible. The natural contains essential elements of randomness and irreversibility. This leads to a new vision of matter: No longer passive, as described in the mechanical world view, but associated with spontaneous activity. This change is so deep that I believe we can really speak about a n e w h u m a n dialog with nature. Even at the start of this century, continuing the tradition of the classical research program, physicists were almost unanimous in admitting that the fundamental laws of the universe were deterministic and reversible. Processes that did not fit this scheme were supposed to be exceptions, even artifacts resulting from some apparent complexity, which had itself to be accounted for by invoking our ignorance or lack of control of the variables involved. N o w that we are at the end of this century, more and more of us think that the fundamental laws of nature are irreversible and stochastic; that deterministic and reversible laws are applicable only in limited situations. It is interesting to inquire h o w such a change could occur over a relatively short time span. It is the outcome of unexpected results obtained in quite different areas of physics and chemistry, such as the fields of elementary particles and cosmology, or the study of self-organization in farfrom-equilibrium systems. Who would have believed, fifty years ago, that most, perhaps all, elementary particles are unstable? That we would speak about the evolu-
New Perspectives
219
tion of the universe as a whole? That far from equilibrium, molecules may communicate, to use anthropomorphic terms, as witnessed in the chemical clocks to which we shall later turn in more detail? All of these unexpected discoveries have also had a drastic effect on our outlook on the relation between the "hard" and "soft" sciences. According to the classical view, there was a sharp distinction between simple systems, such as those studied by physics or chemistry, and complex systems, such as those studied in biology and human science. Indeed, one could not imagine a greater contrast than the one that exists between the simple models of classical dynamics or the simple behavior of a gas or a liquid and the complex processes we discover in the evolution of life or in the history of the human societies. But now we find this gap being filled. Over the last decade, we have learned that in nonequilibrium conditions, simple materials, such as a gas or liquid, or simple chemical reactions can acquire complex behavior. This unexpected complexity leads to the appearance of broken temporal or spatial symmetry, to chaos, and to patterns of increasing diversity (2).
FAR FROM EQUILIBRIUM In its evolution toward complexity, with all its connotations of irreversibility and stochasticity, thermodynamics has played an important role. It may be useful to contrast the description of a system in terms of classical dynamics with the thermodynamical one. In the former, we consider, typically, a given number of points interacting through some type of potential. A typical example would be the system formed by the sun and earth. Of course, there are also the other planets and stars, but they are treated as a kind of perturbation. In contrast, thermodynamics is based on the concept of entropy. Entropy has quite unique behavior. Its changes with time can be split into two additive terms: d S, corresponding to the flow of entropy that the system exchanges with the outside world; and d i S , corresponding to the production of entropy inside the system. Because diS is always positive or zero, irreversibility can only create entropy, but not destroy it. For an isolated system, there is no entropy flow, and S changes only as a result of entropy production: Entropy increases monotonously as the system approaches equilibrium, corresponding to a maximum of entropy. For a long time interest in thermodynamics was concentrated on isolated systems at equilibrium. Today, our interest is shifting to nonequilibrium systems interacting with their surroundings through entropy flow. Let us emphasize an essential difference with the dynamical description. In thermodynamics, we are dealing with embedded systems. The interaction with the outside world through entropy flow plays an essential role. This immediately brings us closer to situations like towns or living systems, which can only survive because of embedding in their respective environments.
220
Prigogine
From the start, the thermodynamical point of view is one of interaction, we could say a "holistic" one. But there is more: Suppose that in the gravitational example quoted above, we have some foreign body approaching the earth. This would lead to a deformation of the earth's trajectory that would then remain so deformed forever. This class of systems has no way of forgetting perturbations. This is no longer the case when we include the effects of dissipation. A damped pendulum will reach a position of equilibrium, whatever the initial perturbation. Again, let us emphasize how much closer to life is the thermodynamic description. If I impose large oscillations on an undamped pendulum, its period slows down and remains so slowed forever. However, when I run, my heart beat increases; but when I rest, it returns to its normal rate. In thermodynamics, perturbations, may be forgotten. In the thermodynamic description that includes dissipation, we have attractors. Without attractors, our world would be chaotic. No general rules could ever have been formulated. Every system would pose a problem in itself. We can now also understand in quite general terms what happens when we drive a system far from equilibrium. The attractor that dominated the behavior of the system near equilibrium may become unstable as a result of the flow of matter and energy that we direct at the system. Nonequilibrium becomes a source of order; new types of attractors, more complicated ones, may appear and give the system new remarkable and space-time properties. Let us illustrate these general statements with two examples that are widely studied today. We first consider the so-called B6nard instability. It is a striking example of instability in a stationary state, giving rise to the phenomenon of spontaneous self-organization; the instability is a result of a vertical temperature gradient set up in a horizontal liquid layer. The lower face is maintained at a given temperature, higher than that of the upper. As a result of these boundary conditions, a permanent heat flux is set up, moving from bottom to top. For small differences of temperature, heat can be conveyed by conduction, without any convection; but when the imposed temperature gradient reaches a threshold value, the stationary state (the fluid's state of "rest") becomes unstable: Convection arises, corresponding to the coherent motion of a huge number of molecules, increasing the rate of heat transfer. Under appropriate conditions, the convection produces a complex spatial organization in the system (3). There is another way of looking at the B6nard phenomenon. There are two elements involved: heat flow and gravitation. At equilibrium conditions, the gravitational force has hardly any effect on a thin layer on the order of 10 mm. In contrast, in the B6nard instability, gravitation gives rise to macroscopic structures. Nonequilibrium matter becomes much more sensitive to outer world conditions than matter at equilibrium. I like to say that at equilibrium, matter is blind; far from equilibrium, it may begin to see.
New Perspectives
221
This can also be illustrated by the example of chemical oscillations. The Belousov-Zhabotinski reaction is a well-known chemical clock. The striking results obtained by Zhabotinski became available in the late 1960s. I cannot discuss here the detailed mechanism (see refs. 3,4). But I would like to point out some general features. Ideally speaking, we have a chemical reaction whose state we control through appropriate injection of chemical products and elimination of waste products. Suppose that two of the intermediate components are formed by red and blue molecules, respectively, in comparable quantities. We would expect to observe some kind of blurred color with, perhaps occasionally, some flash of red or blue spots. This is, however, not what actually happens. Under appropriate conditions, we see--in sequence--the whole vessel become red, then blue, then red again; we have a chemical clock. In a sense, this violates all of our intuition about chemical reactions. We were used to speaking of chemical reactions as being produced by molecules moving in a disordered fashion and colliding at random. In contrast, the existence of chemical clocks shows that far from being chaotic, the behavior of the intermediate species is highly coherent. In a sense, the molecules have to be able to "communicate" in order to synchronize their periodic changes of color. In other words, we are dealing here with new supermolecular scales, both in time and space, produced by chemical activity. In a stirred chemical reactor, we can only observe temporal behavior (which may be periodic or "chaotic"). In other cases, we may also see the appearance of space structures. Another important concept is bifurcation. A system presenting a multiplicity of solutions may be characterized by a bifurcation diagram. We see that at critical points new types of solutions emerge. There is also another elment I would like to emphasize. Near a bifurcation point, the system has a choice between the two branches; we may therefore expect a stochastic element. Quantum mechanics appears to us so revolutionary because it introduces a stochastic description in the microscopic world. The new developments we have summarized here suggest that this stochastic type of behavior may also appear in the macroscopic world.
CLIMATE AND CHIRALITY The concepts that we outlined in the previous section are today diffused throughout a large domain, including biology, the social sciences, and climatology. One of the main problems we confront is to understand the large-scale regularities that shape our environment. Let me give two examples of what I have in mind. Consider first the problem of climate. We know now that climate has fluctuated violently over the past. Climatic conditions that prevailed during the last two or three hundred million years were extremely different from those found at present. During these periods, with the exception of the quaternary
222
Prigogine
era (which began about two millions years ago) there was practically no ice on the continents, and the sea level was higher than its present level by about 80 m. A striking feature of the quaternary era is the appearance of a series of glaciations, with an average periodicity of one hundred thousand years, on which is superimposed an important amount of noise. What is the source of these violent fluctuations that have obviously played an important role in our history? There is no indication that the intensity of solar energy may be responsible for these fluctuations. Let us now consider a problem by which, on the contrary, we are stunned at the remarkable uniformity and lack of fluctuation. In chemistry, we deal often with chiral molecules. For example, in most amino acids, the central carbon is asymmetric: There are two configurations called the D and L isomers. The relation between the two is somewhat similar to left- and right-handed spirals. The curious fact is that amino acids in proteins are always in the L configuration. A somewhat similar example is the case of particles and antiparticles. From the point of view of quantum theory, they play nearly identical roles. Still, the world is composed of particles, whereas antiparticles are produced in accelerators. We now have the tools to begin to understand these two extreme situations: Climate is described by highly nonlinear equations, involving both the absorption and the emission of solar light. In short, this leads to two types of climate: a cold one and a moderate one. There are, of course, fluctuations around these two climates. However, these fluctuations taken alone would not be sufficient to tilt the climate from one type to another. But, in addition, the eccentricity of the earth's orbit leads to a very slight periodocity in the solar influx. This has a major consequence, which was derived by two independent research groups (5,6). Such a very small outside perturbation leads to amplification of the transition probability from one climate to the other. We begin in this way to understand the glaciation periodicity of one hundred thousand years. The point that I want to emphasize here is that the internal noise may amplify minute external perturbations, leading to major changes in climate. Let us turn to the second question: We have seen that matter under nonequilibrium conditions may become very sensitive to outside perturbations. This is especially true in the neighborhood of bifurcation points. A very minute difference, which may exist in the stability of right- or left-handed molecules (leading for example, to differences in the energy of activation on the order of 10 15) may suppress the stochasticity inscribed in a "symmetrical" bifurcation and lead to a preferential selection of one of the states (7). This may lead to applications in a wide range of fields. We may construct new types of highly sensitive switches. However, a price has to be paid; these switches are very slow. We have to maintain the bifurcation
New Perspectives
223
parameter near the bifurcation point for long periods (the exact value of which may vary from hours to minutes, according to the kinetic equations).
HUMAN BEINGS AND ENVIRONMENT Living beings are undoubtedly the most complex and organized objects found in nature, in view of both their morphology and their functioning. They already serve as prototypes from which physical scientists can get the motivation and inspiration needed for an understanding of complexity. They are literally historical structures, since they have the ability to preserve the memory of forms and functions acquired in the past, during the long period of biological evolution. Moreover, they function definitely under conditions far away from equilibirum. An organism as a whole receives continuous fluxes of energy and matter, it transforms into quite different waste products that are evacuated into the environment. These nonequilibrium conditions are observable at all scales of the organism. It is then legitimate to inquire whether some of the above features of biological systems can be attributed to transitions induced by nonequalibrium constraints and appropriate destabilizing mechanisms similar to chemical autocatalysis. This is probably one of the most fundamental questions that can be raised in science. No exhaustive answer can be claimed today, but one can mention some examples in which the connection between physicochemical selforganization and biological order is striking: The reciprocal relations between chemotaxis and amplification of inhomogeneity in the aggregation process of Dictyostelium discoideum, the accrasial amoeba; the threshold behavior of the immune system, induced by a cellular dynamics similar to classical equations of enzyme kinetics; adaptative phenomena in social insects, whose behavior presents highly probabilistic features at the level of the individual, and coherent patterns at the scale of the colony. These examples, which I cannot develop here, are promising in the sense that they permit us to believe that in the future the gap between biology and physics will narrow as we design a new approach to complexity (8).
CONCLUSION Is what we traditionally called fundamental physics the basic description of nature, or is it instead a simplified description of a reality to which we only now begin to gain some access through the study of complex processes? It would be presumptuous of us to try to answer yet, but the very fact that we can ask such questions shows that our perspectives have changed drastically since the time when, at least on our level, the laws of nature seemed to have reached their final formulation.
224
Prigogine
REFERENCES 1. Prigogine, I., and Stengers, [., (1979), La nouvelle alliance, Paris, Gallimard (1979). English translation Order out of Chaos, Bantam Books, NY (1984). 2. Prigogine, I., (1980), From Being to Becoming. Time and Complexity in the Physical Sciences, Freeman, CA. 3. Nicolis, G., and Prigogine I., (1977) Self-Organization in Nonequilibrium Systems--From Dissipative Structures to Order through Fluctuations, Wiley, NY. 4. P. de Kepper and I. Epstein, (1982) Journal of the American Chemical Society, 104, 9, 2657. 5. Nicolis, C., (1982), Stochastic Aspects of Climatic Transition-Responses to a Periodic Forcing, Tellus 34, 1-9. 6. Benzi, R., Parisi, G., Sutera, A. Vulpiani, A. (1982) Stochastic Resonance in Climatic Change, Tellus 34 10-16. 7. Kondepudi, D. K., and Nelson, G. W., (1983) Physical Review Letters, 5On, n ~ 14 1023-26. 8. Nicolis, G., and Prigogine, I., (1985) Understanding Complexity, to appear Piper Vlg.
