Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

The evolution of ecological structure

In recent years there has been some talk about the ecological restructuring of industry. But before we can discuss this coherently, we need to know what precisely ecological structure is. What is the organizational principle which underlies it? If we cannot answer this, then we surely cannot hope to organize human systems in an "ecological" manner. So, first, let us consider the only example that we have of sustainable structure: an ecosystem.

The key issues concern such questions as why the ecosystem is as it is. Why this number of populations? Why not more or less? Why these connections and not others? What would happen if we interfered with the system? Are the feeding relationships necessary, or do they merely reflect proximity and convenience?

None of these questions can be answered from the flow diagram of figure 2. Not only that, but, if we build a model based on the appropriate mechanisms of birth and death which change the numbers of each population, we might imagine that we could then use a computer to predict the behaviour of the system and perform simulations for policy analysis.

Unfortunately, this is not the case. When we run our computer model, it simplifies down to just a few species, because there are parallel paths through the system. Some of these are more "effective" than others; this leads to the elimination of apparently inferior paths through the action of (un)"natural selection."

But, in reality this does not occur. The system remains complex. Some source of diversity successfully opposes the tendency to simplify down that is apparent in our computer simulation. And this is the key to the understanding of ecological structure, the evolutionary process, and sustainability itself. The organizing principle that underlies sustainable systems is the presence, the maintenance, and the production of microscopic diversity in the system! These ideas have been developed in a series of recent papers (Allen and McGlade, 1987a, 1989; Allen, 1990; Allen and Lesser, 1991).

Ecological structure results from the working of the evolutionary process, and this in turn results from the nature of ecological structure. We can understand the ecological structuring of human activities by considering a "possibility space" representing the technologies and options that could potentially arise. In practice, of course, this is a multidimensional space of which we would only be able to anticipate a few of the principal dimensions. Ecological structure emerges over time, as the types of behaviour present in our possibility space increase and become more complex over time, and this is what we have successfully modelled.

The possibility space will be explored if the methods or techniques that firms use are influenced by new scientific knowledge and new ideas or by information and perceptions concerning others. New no tions must be either generated within a company or may be copied or miscopied from others. Either way, cost and effort are expended in finding, filtering, and adapting ideas. New ideas still involve an element of risk when implemented. Physical constraints automatically ensure that some techniques do better than others, and so there is a differential rate of survival and of profitability.

Fig. 2 The Crystal River estuarine ecosystem (Source: Homes and Kemp, 1983)

The possibility space is filled with an "evolutionary landscape," with hills representing high performance. Our simulations show how behaviours that include "exploration" in possibility space, although loss-making in the short term, will in the longer term eliminate behaviours that are fixed. Although exploration is costly in the short term, a small fraction of the initiatives tried are better then previous practice, and it is the gradual amplification of these, and the suppression of the less successful tries, that allow an adaptive progress to higher performance.

It is the presence of variations in the behaviour which, though costly, provides the capacity to "climb" the hills of the adaptive landscape as a result of the differential success and failure of different variants. The landscape expresses the pay-off that would be experienced by an actor or company in competition with the behaviours used by its competitors at that time. But, of course, the landscape is not really fixed, because as soon as a new technology or technique is found to be successful, and a firm moves up a hill, the other firms will respond and change their behaviour, moving the hill away again. In addition, improvements in competitors' technology will also have the effect of pushing any given participant with fixed behaviour lower down the slope.

This means that, over the longer term, evolution favours populations that retain the ability to climb hills, that is, to learn, rather than those that can perform optimally in any given circumstances. We begin to discern the nature of survivable organizations, and of sustainability itself..

This perspective on evolution shows us the error involved in the traditional "equilibrium" view that has been current. If each technology were sitting on a hilltop, then no advantage could be gained from exploration. Evolution would be "over" and there would be nowhere better to evolve to, and nothing to learn. Complicated equations in possibility space would be unnecessary, and rational analysis would be able to optimize a firm's behaviour without evolutionary adaptation. In short, life would be simple but boring.

Fortunately, or unfortunately, we need not worry about this possibility, because this would only be true if evolution were really over. In the real world, competitors, allies, clients, technologies, raw materials, costs, and skills all change. Any group or firm that fixed its behaviour would sooner or later be eliminated, having no adaptive or learning capacity with which to respond.

The landscape in possibility space reflects the advantage to be gained from any particular option, and depends on the techniques and behaviours that happen to be present at a particular moment. The peaks of the landscape represent the present performance goals of the firm or group in question, whose decisions and innovations will try to move up the slope. However, the other actors of the system will continue to modify the landscape as they also adapt and change in pursuit of their goals. The goals of each type of actor coevolve with those of the others present.

These experiments show that a mixture of exploratory diffusion paths in some behaviour or technology space, and their differential success, makes the difference between what is merely mechanical and what, on the contrary, contains the capacity for adaptation and creativity. It is the latter that might be called "organic." It is this vision of ecological structure as a temporary balance between exploration and constraint that is at the core of our new understanding.

Computer models have been developed that show explicitly how these adaptive landscapes are generated by the mutual interaction of behaviours or technologies. In the space of possibilities, closely similar products are mostly in competition with each other, but there is some "distance" in this space, some level of dissimilarity, at which two products or technologies do not compete with each other.

If we begin with a single type of product, then it will grow until it reaches the limits set by the competition either for underlying resources or for customers. At this point, the pay-off for explorers and entrepreneurs switches from negative to positive, as they can now escape somewhat from competition. We see that any successful behaviour eventually digs a hole in the landscape, until there is a hill to climb on either side and exploration is rewarded. Growth is restricted initially because of the "competitive shadow" of the original behaviour, but at a certain distance the products are sufficiently different from the original type; they begin to reach another market and require different resources.

In its turn, this new behaviour or product increases in volume until it too is limited by internal competition for the limiting resource; and once again there is a pay-off for innovation, particularly for those on the outside of the distribution, as they climb another slope towards new regions of possibility space. An evolutionary tree develops, branching as it grows. However, there are also moments when completely novel options emerge spontaneously during the simulation, and an ecology of interdependent behaviours emerges.

Fig. 3 The evolutionary landscape of untried options. Costly experimentation leads to better performance

The ecology that emerges is dynamic, since the identity of each behaviour is maintained by the balance between a continual diffusion of innovators outwards into the space of untried options and the competitive field that exists around it owing to the others. In fact, it would not be possible to anticipate the final range of technologies or products that will inhabit the system, because random events which occur during the "filling" process will affect the emerging pattern of new technologies or products. Instead of the system simply filling pre-existing market niches, the whole process is a creative one, which would be different if repeated from the same initial conditions.

This model offers us a non-reductionist, scientific basis for discussing the interaction of individuals and their collective structures. Such a system operates beyond the mechanical paradigm, because its response to external interventions can involve changes in structure and in the nature of the behaviours or technologies in the system. Suppressing particular components of such a system, perhaps as a result of changing market conditions or environmental regulations, will provoke a complex response from the system, as other behaviours adjust.

Although the "inventiveness" of the system is constantly present, as there is diffusion into the possibility space, it is fascinating to see that our research shows that only at certain moments in time does this lead to structural change. In other words, the system evolves in phases of apparent stability, separated by periods of instability and fairly rapid reorganizations, although the pressure of exploration and creativity is relatively constant.

Such a picture may eventually explain such phenomena as the cycles of growth and stagnation that seem to characterize our economic systems, a phenomenon that has been linked to "economic long waves" and the patterns of innovation and change.


The evolutionary models described above tell us that change is really the result of non-average behaviour exploring untried options, which, at certain moments, encounters mechanisms of positive feedback that lead to its amplification and self-reinforcement. Once a new option emerges it drives the system in the direction of its own amplification, irrespective of the objective or external value of these changes for the system as a whole. In other words, something doesn't have to be "good" for the overall system, or the environment, in order for it to happen; it only has to find self-amplifying mechanisms in its own immediate surroundings. Clearly, this exposes the root of the problem concerning the balance between collective and individual responsibilities and rights. The survival of the whole system may depend on the system's effective adaptation to external events, while the survival of the individuals of which it is comprised may require success in the internal adaptive processes.

As evolution proceeds, it gradually switches from introducing adaptations which deal with the external world to adaptations which succeed within the internal environment. The landscape of advantage ceases gradually to reflect the technology of the primary sector, dealing with the extraction and treatment of raw materials and energy, and gradually becomes increasingly concerned with techniques within techniques, with services to these and then to themselves. This is just the normal process of the development of ecological structure. And this principle applies not only to the balance of internal and external relations, but also to each organization within that structure.

Our evolutionary simulations begin to reveal the universality of myopic values in complex systems, whether they be natural or manmade. In natural ecosystems, each species too will be concerned with the solving of problems of its own, local survival and will not be thinking about planetary good. But if our situation is not intrinsically different from that of natural evolutions, why should we worry, and why should we change? The answer is that natural systems have taken a long time to evolve and explore the coexistence of different populations, and have suffered many local catastrophes and extinctions in the process. We should not forget that more than 90 per cent of species that have existed on the earth have become extinct. Throughout the millions of years of evolution, most of the forms that nature has tried have flowered briefly and then disappeared. We should, therefore, be most wary about assuming that nature's way is the best for us. We now have a conceptual framework of evolutionary systems which allows us a new basis of reflection, even upon the wisdom of nature itself!

The Industrial Revolution only started in earnest some 300 years ago, and was based on technologies that were clearly unsustainable. Despite our capacity to switch our appetite for fossil fuels and raw materials from one source to another, the growth in world population and in material expectations makes the present trajectory still quite clearly unsustainable. What mankind must do is to attempt to substitute reflection and anticipation for the actual experience of catastrophe, in order to learn about the obvious. This book is a step in this direction.

Quite naturally, then, self-amplifying mechanisms within industrial and postindustrial societies take the focus of evolution in human systems away from adaptations concerning the harmony of man and nature, and the sustainable use of the environment and natural resources, towards a multitude of ephemeral values and ends, whose consequences and effects for the collective system are largely unknown.

Many examples of these kinds of positive feedback "attractors" exist in the history of industrial development (Arthur, 1988), starting with the use of steam power to drive pumps in coal mines in order to provide coal for steam power. Similarly, the growth of industrial complexes required the development of railways and roads, while the development of the latter required industry. So, positive feedback loops led to the emergence of spatial concentrations of complementary activities, of pools of skilled labour. They also led to cities, where the different and multiple factors - from technology, through public infrastructure, to finance and investment - were available and working together in an emergent evolutionary "complex." Simulation models exploring the working of such structuring principles on settlement patterns, urban form, market systems, and learning processes have been developed over the past 15 years or so (Allen, 1985, 1988, 1990; Allen and Sanglier, 1978, 1979, 1981).

As this evolution progresses, additional technologies are developed to cope with the unintended consequences of existing technologies which "create" markets through environmental damage. This process of technological "fix" is a measure of the hidden inefficiency of the shortterm optimization of industry. A recent comparison of UK and German industrial development shows the Germans to be much more successful in industrial restructuring that begins to address sustainability, in part because of the longer time-scales that are acceptable in decisionmaking.

Instead of regarding human progress as following some steady path towards a better quality of life, meaning the gradual improvement of man's relationship with the natural world, we see the emergence of change driven by the values of an internal game. Imitation, economies of scale, learning by doing, perceived complementarities of behaviour, and the growth of interdependences all lead to the emergence of an artificial world, cut off from nature and yet of course embedded within it, and therefore potentially ripe for environmental catastrophe.

In human society, fashions, styles, and cultures may rise and fall without necessarily expressing any clear functional advantages with respect to the natural world in which they reside. Indeed, it may be informative to view culture not so much as being the best way of doing things somewhere, but more as resulting from the exclusion of knowledge concerning other ways of doing things. Ritual and shared ideology emerge and serve as the identity and focus of a social group, irrespective of the precise merits or truth of the ideology itself. So much of human attention is focused on playing a role in groups where values are generated internally, and the physical world outside is largely irrelevant.

It is therefore naive to believe that underneath the rich tapestry of life there is a rational scheme within which the complexities of the world would appear as being necessary and unavoidable. Instead we have what Margalef has called "the baroque of nature," but here we would include man. Evolution is creative beyond reason, and in that lies its resilience, since it is not framed to respond to any particular limited scheme.

In Europe, from initially agricultural societies, through the various accidents of history and the ebb and flow of ideas, religions, and authority, evolution moved away from its concerns with crops and live stock and developed new, internal and local values with which to define "successful" technologies and processes. Greatly increased scientific knowledge, together with urban growth and the gradual isolation of the bulk of the population from the realities of their relationship with natural resources, meant that the external consequences (ecological and social) of progress gradually dropped out of the consciousness of society.

Indeed, it became something of a new dogma that if something could be produced and sold at a sufficient profit, then it should only be suppressed if it were contrary to the law. The "burden of proof" has rested with the public or public authorities, and it was rather glibly assumed that the "law" would somehow know what it was wise to inhibit. But this is obviously not the case. Once we begin to understand how complex systems evolve, we see that we can only have an imperfect understanding of the consequences of our actions.

This brings us to the "precautionary principle." If it is difficult to prove that some product or technology is causing harm, then it is probably equally difficult to prove the contrary. But both positions reflect a view based on the concepts of traditional science, that such propositions have definite answers. Despite the attraction of a clear and definite stance, we must try instead to recognize the real complexity of such matters, and move to a new position, involving a mixture of both caution and risk. However, it appears clear that we should at least examine the size of the risk involved, and shift the debate from one in which "scientific proof" is the central focus, to one in which the gains and risk to society are assessed. To destroy mankind for the sake of some new ice-cream or cosmetic would seem to most people to be too high a price to pay, even for the upholding of the principle of individual freedom.

To achieve sustainability we need to understand what the implications of the proposed ecological restructuring of industry might be. The first point concerns the short- and the long-term view. As we have shown above, the qualities which are good in the short term for optimal efficiency are not those which lead to long-term survival. Indeed, they are the opposite. It is equally true, however, that failure in the short term will preclude the chance to show how good something would have been in the long term. In any case, "optimality" can only be defined if measured in certain terms; the measurement system of efficiency is in reality quite complicated, and judgement and wisdom often feature more than economists would suggest. Actual decisionmaking processes in industry are a curious mixture of formal economic accountability to the external capital system and what might be called common sense, much of which reflects human centred, longterm views and sensitivity to environmental issues. The role of government in determining the shape of the external environment is therefore absolutely crucial to sustainable industrial structures.

In general, then, sustainability will result from a balanced strategy corresponding both to fairly efficient average behaviour and to a permanent exploration of possibility space. The adaptability of the whole system results from the adaptability of its component parts, and instead of simply trying to outcompete others in a given domain, creativity and originality can allow a company to move into new areas with less competition. And the power to do this resides in the microscopic diversity of the system, which now can be seen to be vital for two reasons. First, it is the motor of creativity and adaptation, and, second, cultural and technological diversity tend to encourage a wide range of different activities and requirements and hence spread the stress on natural resources and the environment.

However, sustainability also requires that we affect local internal value systems and decisions so as to make them take into account their externalities and their collective effects. Mathematical models capable of providing real information about the evolution of the system as a whole will therefore be valuable in bringing the unconsidered consequences of a decision into view, and making them part of an evaluation. They can also provide vital information concerning the kinds of attributes that are related to survival and sustainability.

Another important point concerns the other properties of ecological systems that we observe. In mature ecosystems, for example, recycling is a major phenomenon, so that a carbon atom, in entering the ecosystem of figure 2, is actually recycled about 15 times as it goes through the system. This clearly is related to sustainability, as the evolutionary principle has a tendency to englobe and then recycle the factors that it requires as inputs. In other words, it evolves in such a way as to expand the organic, living, organized parts into the physical boundaries of the system, and then to retain and re-use materials in the system. At every stage, outputs are being tested as potential inputs for new processes.

This brings us to an important underlying principle in our whole thinking about industrial metabolism, the ecological restructuring of industry, and the problems of waste and pollution. Our evolutionary model shows us that when some successful innovation occurs in the system, some new source of positive feedback has been discovered. Now, in natural ecosystems, this would result in the 'success" of some population for a time, during which its prey would decline in numbers and its wastes would build up. However, in the natural example, after some time, some variant of another population would discover that it could "use" this newly successful population and its accumulations of waste. This is because any spatial concentrations of matter having high free energy are potential sources of food for other populations. After a further period, the initial innovative population would have been reincorporated into the ecosystem, and the challenge that it offered initially would have been met from the internal diversity of the populations of the ecosystem.

In the case of the human population, why therefore are our wastes not recycled as part of the ecosystem? Why are we worrying so much about protecting the environment?

The first reason may simply be one of time. There simply is not sufficient time for an ecological response, particularly as we keep changing what we are dumping on the environment. The second reason, however, is that we are tending to use dispersion as our method of getting rid of wastes. So, instead of accumulations building up and becoming a potential source of raw material for some unknown future process, we are dispersing our wastes into the soils, the oceans, and the atmosphere. Clearly, part of the solution to such problems would be for each new technology to be obliged to provide the "antidote" to itself, that is, the mechanisms necessary to break down the new wastes that have been created. The stimulation both of clean technologies and of recycling should therefore be considered along with such ideas as a carbon or raw materials tax.

This does raise a very general point. Obviously, there are two choices in dealing with wastes: one is dispersion, and the other is concentration. While concentration can appear to be very dangerous in the case of toxic wastes and should require special costly permits, this approach does offer the future possibility that new processes or organisms may be found which can use waste materials, so that they will then become part of an ecological recycling process. In the longer term dispersion may be far more dangerous, since it gradually leads to a shift in the basic parameters of the biosphere, and to a potentially irreversible global change.

The new vision of an industrial metabolism requires not only that we control dispersion as a means of waste treatment, but that mechanisms should be foreseen for making concentrations of wastes part of an alternative ecology. Instead of disrupting the ecological structure and producing zones of accumulation with no known outlets, we must try only to allow technologies that can demonstrate how they can fit within an ecological pattern of flows.

Such changes, however, require vision. And the existence of a strong need for a radical shift in the functioning of the industrial system does not guarantee at all that industry itself will respond. Indeed, the internal games played within the various large companies, often multinationals, that dominate most markets today may well mean that once again concern for the external world will always in fact be a secondary matter.

New taxes on environmental damage, carbon, or raw materials, for example, may well lead to shifts towards more sustainable behaviour, but it may also require quite a sophisticated study for the real effects to be anticipated. In reality, we do not know the environmental damage that we are doing when we destroy some local ecosystem, since we do not understand precisely the source of its resilience and adaptability. Merely setting a price on such actions may therefore simply allow the wealthy to continue to do whatever they like.

There is, however, a place for governmental and international action. It is up to the political process to set the framework within which industry should operate. When Voltaire visited England in the eighteenth century, he remarked that political considerations had been subordinated to commercial ones in England, contrary to the case elsewhere in Europe, and that this was the reason for English prosperity. Perhaps it is time now to re-examine this idea, and to subordinate commercial interests at least to some restrictions concerning the natural environment.

We must begin to face up to our responsibilities, and consider the whole metabolism of modern society and its industrial motor. We must set about the difficult task of finding policies and regulations that will lead to the evolution of "earth-friendly" technologies and industry. One important strength in this is that the nations and companies that stimulate this kind of technological evolution will become the new industrial leaders, since environmental concern and regulation can only grow over the next decades as increasing cognizance is given to the external effects of internal processes.

The question is how we can change industrial culture, and again we may see what evolutionary theory and ecology have to suggest. This transition requires that we step outside any narrow disciplines and try to understand the overall evolution of the physical, socio-economic, demographic, technological, and cultural systems that are the objects of concern. It is not sufficient to ask "experts" what to do, since in some ways they are already part of the problem. It is the narrow values of experts that we are attempting to escape from. We need to construct a systemic view of our society, using as a basis the kind of evolutionary models referred to in this chapter, which can be used quantitatively to integrate the information of experts in the different fields.

This chapter is really about a research agenda concerning the implications for mankind of the conceptual framework of evolutionary processes and ecological structure. The identification of attributes related to survivability and adaptiveness is clearly key, and the links to cultural and social diversity require examination. The development of holistic methodologies and respective tools for policy evaluation and decision support seem also to be urgent. Similarly, methods with which to assess the size and scale of risks involved in the growth or continuation of technological processes seem equally pressing, together with drafting of some precautionary principle that will limit the possibility of trivial aims putting large parts of humanity at risk. In short, our research agenda is one which tries to reintegrate creativity and adaptability into scientific thinking, and man back into nature, providing thereby the understanding necessary to mankind for the development of sustainable strategies.

Contents - Previous - Next