Kay, J, Schneider, E.D,. 1994, "Embracing Complexity, The Challenge of the Ecosystem Approach", Alternatives Vol 20 No.3 pp.32- 38

REPUBLISHED AS:

Kay, J, Schneider, E.D,. 1995, "Embracing Complexity, The Challenge of the Ecosystem Approach", in Perspectives on Ecological Integrity, L. Westra, J. Lemons (eds) Kluwer, Dodrecht, pp. 49-59.

As environmental degradation and change continues, decision makers and managers feel significant pressure to rectify the situation. Scientists, in turn, find themselves under pressure to set out simple and clear rules for proper ecosystem management. The response has been one of frustration. Michael Soule and Laurence Slobdokin both loudly complain that ecology is an intractable science, immature and not very helpful. Kristin Shrader-Frechette and Robert Peters reproach ecologists for not producing simple testable hypotheses.¹ Meanwhile policy makers and managers clamour for a measure of ecosystem integrity whose value in different situations can be predicted by simulation models. The question on everyone's mind is "what does ecosystem science identify as the main, simple, basic, universal laws which will allow quantitative prediction of ecosystem behaviour and what are the resulting rules for ecosystem management?"

 All of these demands on ecology are predicated on a vision of science which assumes that it can provide firm knowledge, and that the only way of obtaining this knowledge is the scientific method. The standard scientific method works well with billiard balls and pendulums, and other very simple systems. However, systems theory suggests that ecosystems are inherently complex, that there may be no simple answers, and that our traditional managerial approaches, which presume a world of simple rules, are wrong-headed and likely to be dangerous.

 In order for the scientific method to work, an artificial situation of consistent reproducibility must be created. This requires simplification of the situation to the point where it is controllable and predictable. But the very nature of this act removes the complexity that leads to emergence of the new phenomena which makes complex systems interesting. If we are going to deal successfully with our biosphere, we are going to have to change how we do science and management. We will have to learn that we don't manage ecosystems, we manage our interaction with them. Furthermore, the search for simple rules of ecosystem behaviour is futile.

 Take for example the diversity-stability hypothesis.² This is a classic example of the kind of simple rule people are looking for. Students are taught that diversity in ecosystems is important because it maintains their stability. Yet, we know that to obtain an increase in diversity in ecosystems, we need only stress them.³ Daniel Goodman long ago dissected and refuted this hypothesis and yet we still see it being promoted as a guideline for management and policy.⁴ Why? Because we want simple answers to complex questions.

 The diversity-stability hypothesis illustrates this nicely. Examination of what is meant by diversity and stability quickly leads us into the quagmire of complexity. Is diversity to be measured by number of species? the relative abundance of species? their richness? Which species do you include? big ones that are easy to count? all the micro-organisms in the soil? Very quickly it turns out that there is no one correct way to measure diversity, and in the end, it is an observer-dependent phenomenon, dependent on which species the researcher decides to include.

 The notion of stability is even more slippery.⁵ The traditional approach developed by M. Lyapunov focuses on some numerical state function and whether that function has a constant value which the system tends towards and returns to when disturbed. But what state function should we measure? population numbers? biomass? productivity?... The list is endless and the problem doesn't end there. We may choose a function to represent the ecosystem and its stability, but we are now discovering that these functions are not stable.⁶ Instead ecosystems are dynamic and constantly changing. Stability gives way to the notion of a shifting steady mosaic.⁷ Thus, the diversity-stability hypothesis evaporates because the basic concepts of diversity and stability are just too simple to describe the complex reality of ecological phenomena.

 The same is true of the notion of "succession", the idea that an ecosystem develops through a series of dominant vegetation types and ultimately reaches a climax community. Robert McIntosh has documented the ongoing debate about succession and identifies six major schools of thought in the US alone.⁸ The thinking ranges from succession as an orderly pattern of development, which is reproduced time and time again, to succession as a myth, that is there is nothing but random assemblages of species with no underlying patterns. There is nothing approaching consensus about succession in the ecology community. In fact, ecologists bemoan that there is not one single "law" of the science of ecology. Why? Because we are asking the wrong questions.

 There is a group of thinkers who argue that to deal with ecology requires an "ecosystem approach", an approach based on the notions of complex systems theory, the grandchild of Ludwig von Bertalanffy's general systems theory.⁹ It is a fundamentally different approach to knowing about the world, and it is, not surprisingly, complex itself. Any effort to study complex systems must look at them in the context of space, time, energy and information. We shall probe, in turn, each of these aspects of ecosystems as complex systems.

The sky is falling

Part of our trouble is that our conventional notion of science is based on understanding the temporal and sometimes spatial dynamics of systems in the context of their inertia (mass). We see the world as billiard balls from a Newtonian perspective. Ball A strikes ball B causing it to move. All activities of a system can be explained by mechanisms, in terms of the interactions of components, usually in a linear way. Component interactions are sufficient to explain all. So science focuses on establishing which components are responsible for what.¹⁰

 At the turn of the century, several insights changed how scientists look at the physical world. In terms of space and time, it was realized that there is not a preferred observer and that the relationship between observers, at least in four dimensions, is not linear. Space and time are curved. Furthermore, energy is quantized, mass is a form of energy, and we always lose information about things.¹¹ The world is running down. The sky is falling.

 These insights did not affect how we looked at the world on a day-to-day basis. Its direct impact has been on the development of things "nuclear" and things "solid state" (e.g. computer chips). As for the world running down, we already knew that. So scientific inquiry continued to follow the "scientific method", attempting to explain everything through mechanistic interactions of components. The logical extremes have been the elementary particles of physics, the selfish gene of biology, and "Newtonian ecology".

 However, the minute one leaves the physical sciences there is a paradox, a paradox whose resolution ultimately requires us to abandon the hypothesis that the reductionist, mechanistic, scientific method is sufficient for understanding the world. The paradox is that the second law of thermodynamics maintains that the world is running down, but the biological world is not running down. Quite the opposite is happening; life is proliferating. The sky is not falling! The same can be said of the systems studied by the social sciences.

 A revolution in science has occurred in the last two decades that is as profound as the one which occurred between 1890 and 1910 with the work of Ludwig Boltzmann, Albert Einstein, Josiah Gibbs, Max Planck, et al. The revolution of the turn of the century was about how we view the microscopic world. It did not change how we look at our world, day-to-day. The current revolution is about how we look at the macro world and it will profoundly affect our day-to-day living, our institutions and our decision making, including decisions on judicial matters.

 It is fitting that one cannot put this new set of insights down on paper in a nice linear way. The revolution emerges from the synergism of new insights in several fields. Since the prevalent world view is largely about mechanistic-reductionist predictions about space and time, it seems appropriate to start with the unravelling of this.

Space and time

Catastrophe theory describes the change in systems over time. It predicts that systems will undergo dramatic, sudden changes in a discontinuous way. The classic example is the failure of a structural beam under loading. The choice of the name of the theory is quite unfortunate because it implies abnormal nasty events, when in fact such events are normal and necessary for the continued ordinary functioning of many systems. Your heartbeat is a catastrophic event, as is the emptying of your bladder. Both are necessary for your continuing survival. Both are discontinuous events that occur suddenly.

 Furthermore, at the point where a system undergoes a catastrophic change several distinct changes are possible and actually occur - which one is not predictable. For example, dogs (in fact most animals) have a bubble of space around them which is their territory. Enter the space (the catastrophe threshold) and the dog will either retreat or attack, but it is not, a priori, possible to predict with certainty which of the two actions will occur.

 The general insight from catastrophe theory is that the world does not always change in a continuous and deterministic way. There are points in any system's development where several possible directions of radical change are open, and it is not possible to predict, with certainty, which one will occur.

 Chaos theory takes this one step further by noting that change in any dynamic system is ultimately not predictable, because individually small interactions between components accumulate.¹² This applies even to the balls on a billiard table and the planets in the heavens, those objects whose motion Newtonian mechanics is supposed to predict perfectly. Consequently our ability to forecast and predict is always limited, for example to about five days for weather forecasts, regardless of how sophisticated our computers are and how much information we have.

 These two bodies of insight into behaviour in space and time eliminate the possibility of precise, a priori, mechanistic, deterministic predictions of the future. Computers cannot substitute for crystal balls, except for very limited classes of problems that occur over short spatial and temporal dimensions.

Thermodynamics and open systems

The next insights concern energy, that is thermodynamics. Ilya Prigogine in his Nobel Prize winning work, showed that spontaneous coherent behaviour and organization (e.g. tornadoes) can occur and are completely consistent with thermodynamics.¹³ The key to understanding such phenomena is to realize that one is dealing with open systems with a constant flow of high quality energy. In these circumstances, coherent behaviour appears in systems almost magically. Prigogine showed that this occurs because the system reaches a catastrophe threshold and flips into a new coherent behavioural state. (This is evident for example in the vortex which spontaneously appears when draining water from a bathtub.)

 Prigogine's work can be taken one step further to explain the energetics of open systems.¹⁴ An open system with high quality energy pumped into it is moved away from equilibrium. But nature resists movement away from equilibrium.¹⁵ So the open system responds with the spontaneous emergence of organized behaviour that uses the high quality energy to maintain its structure, thus dissipating the ability of the high quality energy to move the system away from equilibrium. As more high quality energy is pumped into a system, more organization emerges to dissipate the energy.¹⁶

 This view of the world is radically different from that of a reductionist view which sees the world's workings in terms of mechanical interactions between components of a system. The emergence of organized behaviour, and even life, is now mandated by thermodynamics. This self-organization is characterized by abrupt changes that occur when a new set of interactions and activities emerge among components and the whole system.

 The form of expression this self-organization takes is not predictable in advance because the very process of self-organization is by catastrophic (in the catastrophe theory sense) change; it "flips" into new regimes. As noted earlier, one of the characteristics of catastrophic change is that systems may have several possible behavioural pathways available at a catastrophe threshold. Which pathway is followed is largely an accident of circumstances. A reductionist world view, which cannot deal with the reality of emergence and self-organization in non-equilibrium systems, cannot offer sufficient explanation of how the world works.

 An important observation about systems that exhibit self-organization is that they exist in a situation where they get enough energy, but not too much. If they do not get sufficient energy of high enough quality (beyond a minimum threshold level), organized structures cannot be supported and self-organization does not occur. If too much energy is supplied, chaos ensues in the system, as the energy overwhelms the dissipative ability of the organized structures and they fall apart. So self-organizing systems exist in a middle ground of enough, but not too much.

 Furthermore, these systems do not maximize or minimize their functioning. Rather their functioning represents an optimum, a trade-off among all the forces acting on them. If there is too much development of any one type of structure, the system becomes overextended and brittle. If a structure is not sufficiently developed to take full advantage of the available energy and resources, then some other more optimal (i.e. better adapted) structure will displace it. In sum, these systems represent a fine balancing act. Inevitably then, human management strategies that focus on maximizing or minimizing some aspect of these systems will always fail. Only management strategies which maintain a balance will succeed.

Middle number systems and observer dependence

The description of these self-organizing systems is known as the middle number problem. Small number problems involve a very controlled situation with very few components. (e.g. two billiard balls colliding). Such problems are usually well explained by traditional science. Large number problems involve so many objects interacting that they can be described by statistical means (e.g. the air molecules in a room). This is the domain of classical thermodynamics and statistical mechanics. Middle number problems involve many things interacting in ways that are not random (e.g. most real world problems).¹⁷

 This area of inquiry is the domain of system theorists. There are two important lessons to be learned from the study of middle number systems. First, such systems can only be understood from a hierarchical perspective. Neither a reductionist nor a holistic approach is sufficient. One must look at the system (e.g. a wetland or a woodlot) as a whole and as something composed of subsystems and their components. One must also look at the system in the context of its being a subsystem of a bigger system, which in turn is part of a wider environment. So, study of an animal population without reference to the individuals that make it up, the community it belongs to, and the environment it lives in, is not sufficient. This is not to say that population ecology is useless, but on its own, it cannot explain ecological phenomena.

 Another property of these middle number systems is that everything is connected (at least weakly) to everything else. An analyst, in identifying the system to be studied, decides what to include and what to leave out. These decisions, about scale and extent and the hierarchical units to be studied, may be done in a systematic and consistent way, but they are necessarily subjective, and to some extent arbitrary. They reflect the viewpoint of the analyst about which connections are important to the study at hand, and which can be ignored. Thus the notion of a pristine objective scientific observer, is not applicable to the study of self-organizing systems.

 It is the observer-dependent nature of the study of self-organizing systems which is the most difficult point for traditional reductionist science to understand. Take for example the notion of an ecosystem. Because the world is made of living and non-living stuff with multitudes of interrelationships, any one defined ecosystem is just one package of stuff and relations. To describe one ecosystem is to take one of many possible perspectives on these entities.¹⁸ An ecosystem can refer to what's happening on our eyelashes, in our gut, or in Lake Ontario, or in the boreal forest. Where one draws the boundaries around an ecosystem depends on the scale and extent from which one needs to observe the whole, given the purpose of the study being undertaken. Different people looking at the same stuff are going to define the ecosystem differently, unless they agree on the inevitably subjective criteria for deciding on scale, extent and hierarchy.

 The response of traditional science to this is that ecosystems don't exist, since we cannot come up with an observer-independent way of defining them. One consequence of this logic is that ecosystem research is not considered proper "scientific" research by most North American granting agencies and is not a fit topic in American ecological journals. Luckily, Canadian and European journals do not have this problem. Complex systems theory represents a profound change in the paradigm for doing science, so profound that traditional science rejects it out of hand. The notion of ecosystem is a focal point for the clash between these paradigms of what science is about.

Information: The key to self-organization

The notions of observer dependence and hierarchical context lead us to discuss the last player of the space, time, energy and information quartet. The key question is: What information do systems need to self-organize successfully?

 All living systems go through cycles of birth, growth, death and renewal. We are all familiar with death and reproduction at the cellular level, and the birth-growth and death of individuals, but it is only recently that Buzz Holling has made us aware that this cycle occurs at many temporal and spatial scales.¹⁹

 Living systems must function within the context of the system and environment of which they are part. If a living system does not conform with the circumstances of the supersystem it is part of, it will be selected against. This process of selection functions at all levels. The supersystem imposes a set of constraints on the behaviour of the system, be it at the level of the cell, individual, population or community. Living systems that are evolutionarily successful have learned what these constraints are and how to live within them. (This is the painful process the human species is now undergoing, assuming it is not selected against).

 But this presents a problem. When a new living system is generating after the demise of an earlier one, it would make the self-organization process much more efficient if it were constrained to variations which have a high probability of success. At the level of cells to species, genes play this role. Genes constrain the self-organization process to those options which have a high probability of success. It is not that genes direct or control the process of development, rather they constrain it to forms which will respect the realities of the supersystem and environment. They are a record of successful self-organization. Genes are not the mechanism of development, the mechanism is self-organization. Genes put boundaries on the process of self-organization.²⁰

 At higher hierarchical levels other devices constrain the self-organization process. For example, some species will kill their young under certain conditions, and many tree species need specific micro-climate conditions to trigger self-organization.²¹ In some species, young are taught behaviours and individuals are banished from the group for inappropriate behaviour. Indigenous human cultures have taboos, morals and other cultural mechanisms that constrain behaviours to those which are sustainable in the context of specific ecosystems. Each of these devices acts, at a particular level of organization, as an information database about self-organization strategies that have an historical track record of success. They set out the boundaries of behaviour by self-organizing systems.

 Given that living systems go through a constant cycle of birth growth, death and renewal again, at many temporal and spatial scales, a way of preserving information about what works and what doesn't so as to constrain the self-organization process is crucial for the continuance of life. This is the role of the gene. At a larger scale it is the role of biodiversity.

 Biodiversity is the information database for ecosystem organization. The ability of an ecosystem to regenerate, as part of the birth, growth and death and renewal cycle, is a function of the species available for the regeneration process. This, of course, is related to the biodiversity of the larger landscape that the ecosystem is part of. Thus preservation of biodiversity is important because we are in effect preserving the library used for regeneration of landscapes.²²

The ecosystem approach and integrity: A new perspective

So what are the implications of all this? The first is that we need to look at ecosystems from a hierarchical perspective with careful attention to scale and extent. Second, we must examine the spatial, temporal, thermodynamic and information aspects (dynamics) of these systems. This must be done in the context of behaviour which is both emergent and catastrophic. In other words, we must recognize that ecosystems are dynamic, not deterministic, that they have a degree of unpredictability and that they will exhibit phases of rapid change.

 This is not to say that ecosystem behaviour is chaotic or random and haphazard. On the contrary, ecosystem behaviour and development is like a large musical piece such as a symphony, which is also dynamic and not predictable and yet includes a sense of flow, of connection between what has been played and what is still to come, the repetition of recognizable themes and a general sense of orderly progression. In pieces such as symphonies or suites we know the stages (allegro, adagio, etc.) that the piece will progress through, even though we don't know the details of the piece. The same is true of ecosystems. Some behave in a very ordered way as does a Baroque suite, while others are full of improvisation as in modern jazz. And yet we know the difference between music and random collections of noise.

 Ecosystem self-organization unfolds like a symphony. Our challenge is to understand the rules of composition and the limitations and directions they place on the organization process, as well what makes for the ecological equivalent of a musical masterpiece that stands up to the test of time. However we should not expect to have a science of ecology which allows us to predict what the next note will be.

 We must always remember that left alone, living systems are self-organizing, that is they will look after themselves. Our responsibility is to not interfere with this self-organizing process or better yet, to enhance it. Of paramount importance, in this respect, is that we must not destroy the information needed for the regeneration process which is continually ongoing. A damaged ecosystem, left to its own devices, has the capability to regenerate if it has access to the information required for renewal, that is biodiversity; and if the context for the information to be used, that is the biophysical environment, has not been so altered as to make the information meaningless.

 Another important thing we need to do is to stop managing ecosystems for some fixed state, whether it be an idealistic pristine climax forest or a corn farm. Ecosystems are not static things, they are dynamic entities made up of self-organizing processes. Management goals that involve maintaining some fixed state in an ecosystem or maximizing some function (biomass, productivity, number of species) or minimizing some other function (pest outbreak) will always lead to disaster at some point, no matter how well meaning they are. We must instead recognize that ecosystems represent a balance, an optimum point of operation, and this balancing is constantly changing to suit a changing environment. And if this isn't radical enough we must bear in mind that all living systems from cells to communities face death and regeneration. This is required by the second law, it is a thermodynamic necessity.

 For us, the notion of serving ecological integrity means accepting all of this. If human activities maintain the integrity of the self-organizing entities that we call life, we will be all right. If they don't, we will be selected out of the systems. We have a simple choice, to be stewards of integrity or disrupters of integrity. There is no middle ground.

 But what exactly is ecological integrity? For an ecosystem, integrity²³ encompasses three major ecosystem organizational facets.²⁴ Ecosystem health, the ability to maintain normal operations under normal environmental conditions, is the first requisite for ecosystem integrity. But it alone is not sufficient. To have integrity, an ecosystem must also be able to cope with changes (which can be catastrophic) in environmental conditions; that is, it must be able to cope with stress. As well, an ecosystem which has integrity, must be able to continue the process of self-organization on an ongoing basis. It must be able to continue to evolve, develop, and proceed with the birth, growth, death and renewal cycle. It is these latter two facets of ecosystem integrity that differentiate it from the notion of ecosystem health.

 This understanding of the behaviour of complex self-organizing systems provides a framework for the investigation of environmentally induced changes in ecosystem organization and integrity.²⁵ It establishes that ecosystems can respond to changes in the environment in five qualitatively different ways:

*The system can continue to operate as before, even though its operations may be initially and temporarily unsettled.

*The system can operate at a different level using the same structures it originally had (for example, a reduction or increase in species numbers).

*Some new structures can emerge in the system that replace or augment existing structures (for example, new species or paths in the food web).

*A new ecosystem, made up of quite different structures, can emerge.

*The final, and very rare possibility, is that the ecosystem can collapse completely and no regeneration occurs.

 This enumeration of possible ecosystem responses to environmental change is far richer than the simple classical notion, which holds that stress temporarily displaces an ecosystem from its climax community, to which it eventually returns. In fact, an ecosystem has no inherent single preferred state for which it should be managed.

 While this identifies the ways in which an ecosystem might re-organize in the face of environmental change, it does not indicate which re-organization constitutes a loss of integrity. It could be argued (and often is) that any environmental change that permanently alters the normal operations of an ecosystem affects its integrity. Ecosystem integrity would then be defined as the ability to absorb environmental change without any permanent ecosystem change. Thus the final four distinct ecosystem responses described above would constitute a loss of integrity, even though all but the last option (collapse) are responses in which the ecosystem reorganizes itself to mitigate the environmental change. However, the reorganized ecosystem is usually just as healthy as the original, even though it may be different. There is no scientific reason that an existing ecosystem should be the only one to have integrity in a situation, just because of its primacy.

 At the other extreme, it could also be argued that any ecosystem that can maintain itself without collapsing has integrity. Utter collapses have been rare, desertification being one of the few examples. This definition would encompass almost all ecosystems, including ones whose organization has changed radically in response to major stress.

 Neither of these definitions of integrity is operationally useful. The definition which accepts only temporary change is too restrictive in most situations, and reflects a desire to preserve the world as it is currently.²⁶ This denies the fundamental dynamic nature of ecosystems and leads to disastrous mismanagement (e.g. the complete suppression of forest fires, which eventually results in catastrophic conflagrations). But the latter definition, which accepts all responses except collapse, does not help managers because it restricts loss of integrity to a situation that rarely occurs and that is clearly undesirable. Hence this definition would be trivial.

 In between these two extremes of definition lies a third option, which holds that some changes in ecosystems are undesirable, and therefore represent a loss of integrity. This option promises to be the most useful but it embraces many possibilities and requires difficult choices. In particular it requires the value-laden selection of criteria for determining which changes are desirable and which are not. The science of ecology can, in principle, inform us about the kind of ecosystem response or reorganization to expect in a given situation. It does not provide us with a scientific basis for deciding that one change is better than another, except possibly in the two extreme cases just discussed.²⁷

 Here again the insight into ecological integrity gained from complex systems theory is that the physical and biological sciences can describe and, even to a limited extent, predict human-induced changes in the biosphere, but they alone cannot determine which changes are acceptable. Ultimately, any evaluation of the ecological acceptability of a human activity, will depend on value judgments about whether the resulting changes in the affected ecosystem are acceptable to the human participants.

 It should be noted that it is exactly this conclusion that leads classical scientists to reject this whole mode of reasoning as unscientific, soft and useless except as a parlour game. The complaint most often spoken is that such a treatment of ecology is not defensible in court, because there are no black and white answers, no linear causes and effects, no definitive mechanisms and no one person to blame. In short this treatment does not lead to a scientific conclusion that this behaviour is good and that behaviour is bad.

 Scientific judgments about right and wrong seemed possible when we viewed the world as a set of billiard balls, and it is this mechanistic, reductionist worldview that our court system assumes. Unfortunately, this worldview with its approach to governance and law does not recognize, and will not help us deal with, the realities of complex systems. And here we have the crux of the issue. If we are truly to use an ecosystem approach, and we must if we are to have sustainability, it means changing in a fundamental way how we govern ourselves, how we design and operate our decision-making processes and institutions, and how we approach the business of environmental science and management.²⁸ This is the real challenge presented by an ecosystem approach. o

James J. Kay is a professor in Environment and Resource Studies at the University of Waterloo. Eric Schneider was formerly cheif scientist of the US National Oceanic and Atmospheric Administration.

Thanks to Henry Regier, George Francis, and Laura Westra for their support of James Kay's research through their Donner, NSERC and SSHRC grants, Marie Lagimodiere for her extensive search for literature on ecosystem and complex system thinking, and Nina Marie Lister for her work on biodiversity and information.

Notes

¹ L.B. Slobodkin, "Intellectual Problems of Applied Ecology," Bioscience, 38:5 (1988), pp. 337-342.

² The diversity-stability hypothesis arose from Robert MacArthur, "Fluctuations of Animal Populations and a Measure of Community Stability," Ecology, 3 (1955), pp. 533-535) in which he proposed that the diversity of a food web was a measure of community stability. G.E. Hutchinson, "Homage to Santa Rosalia Or Why Are There So Many Kinds of Animals?" American Naturalist, 93 (1959), pp. 415-427, mistook this paper as a proof that species diversity explains community stability. Ramon Margalef, "On Certain Unifying Principles in Ecology," American Naturalist, 97 (1963), pp. 357-374; and Ramon Margalef, Perspectives on Ecological Theory (Chicago: University of Chicago Press, 1968) elaborated a theory of ecosystem development which held that species diversity was the cornerstone of the emergence of a stable system. This hypothesis was "codified" as dogma by the Brookhaven Symposium of 1968 in Diversity and Stability in Ecological Systems, G.M. Woodwell, H.H. Smith, eds. (Brookhaven National Laboratories Symposium #22, 1969). It is a very pleasing and simple to understand hypothesis based on the notion that "you don't put all your eggs in one basket." In the early 70s a number of empirical counter-examples to this hypothesis were presented. Daniel Goodman, "The Theory of Diversity-Stability Relationships in Ecology," Quarterly Review of Biology, 50:3 (1975), pp. 237-366, systematically examined the literature and demonstrated clearly that there was no scientific basis for the diversity-stability hypothesis.

³ For example, in southwestern Ontario the most diverse ecosystems can be found in the area between urban development and rural lands. For more discussion see P.S. Petraitis, R.E. Latham, and R.A. Niesenbaum, "The Maintenance of Species Diversity by Disturbance," Quarterly Review of Biology, 64:4 (1989), pp. 393-418.

⁴ See P.J. Burton, et al., "The Value of Managing for Biodiversity," The Forestry Chronicle, 68:2 (1992), pp. 225-237, "... the diversity within a biological community confers some measure of stability to that community," p. 229.

⁵ J.J. Kay, "A Non-Equilibrium Thermodynamic Framework for Discussing Ecosystem Integrity," Environmental Management, 15:4 (1991), pp. 483-495.

⁶ C.S. Holling, "The Resilience of Terrestrial Ecosystems: Local Surprise and Global Change, Sustainable Development in the Biosphere, W.M. Clark and R.E. Munn, eds. (Cambridge: Cambridge University Press, 1986), pp. 292-320; C.S. Holling, "Cross-scale Morphology, Geometry, and Dynamics of Ecosystems," Ecological Monographs, 62:4 (1992), pp. 447-502; and Kay, "Non-equilibrium" [note 5].

⁷ F. Bormann, G. Likens, Pattern and Process in a Forested Ecosystem (New York: Springer-Verlag, 1979).

⁸ R.P. McIntosh, "The Relationship between Succession and Recovery Process in Ecosystems," The Recovery Process in Damaged Ecosystems, J. Cairns, ed. (Ann Arbor Science, 1980), pp. 11-62.

⁹ See for example T.F.H. Allen, T.W. Hoekstra, Toward a Unified Ecology (New York: Columbia University Press, 1992).

¹⁰ This way of looking at the world spills over into our judicial system, where we strive to determine who is responsible, and who is guilty. This is based on the assumption that the observed behaviour can be explained by simple linear interactions between the components. Somebody is responsible for something happening.

¹¹ In the sense that Ludwig Boltzmann spoke of randomization rather than the modern Jaynesian interpretation of information, see E.T. Jaynes, "Where Do We Stand on Maximum Entropy," The Maximum Entropy Formalism, R. Levine and M. Tribus, eds. (Cambridge, Massachusetts: MIT Press, 1979), pp. 15-118.

¹² In classical analysis, small interactions between components (such as friction), interaction due to spherical imperfections (billiard tables which aren't perfectly flat), etc. are ignored. It turns out that these interactions, after some time, actually determine the system's behaviour as much as anything. But these interactions are essentially noise and unpredictable.

¹³ G. Nicolis, I. Prigogine, Exploring Complexity (New York: W.H. Freeman, 1989). Prigogine showed that such systems do not violate the second law that entropy must increase, even though they increase order or organization.

¹⁴ E.D. Schneider and J.J. Kay, "Life as a Manifestation of the Second Law of Thermodynamics," Advances in Mathematics and Computers in Medicine (1994 in press).

¹⁵ This is the second law of thermodynamics restated for non-equilibrium situations.

¹⁶ More formally, from Schneider and Kay, "Life" [note 14], "the thermodynamic principle which governs the behaviour of self-organizing systems is that, as they are moved away from equilibrium, they will utilize all avenues available to counter applied gradients (high quality energy flows). As an applied gradient increases so does a system's ability to oppose further movement from equilibrium." This seems to be the natural principle behind the emergence of life.

¹⁷ See Gerald M. Weinberg, An Introduction to General Systems Thinking (New York: John Wiley and Sons, 1975).

¹⁸ A.W. King, "Considerations of Scale and Hierarchy," Ecological Integrity and the Management of Ecosystems, S. Woodley, J.J. Kay, G. Francis, eds. (Delray, Florida: St. Lucie Press, 1993), pp. 19-46. An ecosystem is a collection of interacting biological entities combined with the physical environment in which they live, which is perceived to act as a whole.

¹⁹ Holling, "Resilience" [note 6].

²⁰ J.J. Kay, "Self-Organization in Living Systems" (PhD thesis, Systems Design Engineering, University of Waterloo, 1984), pp. 85-88.

²¹ For example, jack pine cones require heat from a forest fire to open.

²² N.M. Lister, "Biodiversity: Socio-Cultural and Scientific Perspectives With Reference to Decision Making in the Great Lakes Basin," (unpublished 1994).

²² Ecosystem integrity is about the integrity of ecosystems versus ecological integrity which refers to the integrity of life at all ecological levels including ecosystems. In what follows the focus is on ecosystem integrity.

²³ Kay, "Non-equilibrium" [note 5]; and J.J. Kay, "On the Nature of Ecological Integrity: Some Closing Comments," Ecological Integrity, Woodley, Kay and Francis [note 18], pp.201-212.

²⁴ Kay, "Non-equilibrium" [note 5].

²⁵ Of course one may wish to preserve an ecosystem as an example or specimen of a specific type.

²⁶ To return to the musical composition analogy, the two extreme cases correspond to the playing of the same piece over and over with minor variations or to no music at all. The third option allows for different compositions, but not all compositions. As in music, the question of taste and need plays an important role in deciding which compositions are acceptable.

²⁸ For an early version of some practical and institutional aspects see H.A. Regier, A Balanced Science of Renewable Resources (Seattle: University of Washington Press, 1978).

  Back to JK publication page