How do people make decisions in complicated situations? Many psychology texts argue that human reasoning is deductive. Deduction is reasoning from the general to the particular. A perfectly logical deduction yields a conclusion that must be true provided that its premises are true. Thus it involves specifying a set of axioms and proving consequences that can be derived from those premises. In reality, deduction is handy for solving a host of theoretical problems and a handful of simple practical problems, but it is less helpful for tackling complicated practical problems like those associated with the management of human ecosystems. Moreover, it is definitely of no help when the managers involved are likely to behave and react differently over time.

The truth of the matter is that each of us is a unique product of our own brain and our uniquely individual experiences. Our personal knowledge is honed by the concepts, notions and models which we choose to use to represent it. All of this has to be created, put together and revised over time by us as well as by others in society as a whole. Learning has both an individual and a collective dimension, which builds adaptively on the inductive and intuitive skills of a heterogeneous collection of minds. Erwin Schrödinger summed it up well:

The world is a construct of our sensations, perceptions, memories. It is convenient to regard it as existing objectively on its own. But it certainly does not become manifest by its mere existence. Its becoming manifest is conditional on very special goings-on in very special parts of this very world, namely on certain events that happen in a brain.
(Schrödinger 1967: 94)

As a matter of fact, human agents do more than make rational choices by way of simple deduction. In most cases, no one agent knows what all the others are doing, being forced to rely on a limited amount of shared information and unique experiences of its own. Hence, they make decisions on the basis of whims, hunches, heuristics or mental models and are willing to modify their preferred mental models (to a greater or lesser extent) and come up with new ones where necessary. In other words, when facing open-ended situations, human agents reason inductively.

In chapter 3, Pascal Perez draws upon philosophy of mind and semiotics to suggest, along with an increasing number of colleagues, that these contextual cognitive processes have deceived attempts to develop a predictive modelling formalism so far. He argues that the Artificial Intelligence paradigm needs to be replaced with more modest, but more robust approaches. Anne Dray and colleagues (chapter 12) provide a meaningful example of such innovative approaches through their description of a companion modelling experience in the South Pacific.

Inductive reasoning places different demands on we thinking individuals than the deductive metaphor. It involves pattern formation and pattern recognition, aided by intuition and creativity. Clearly some people are more intuitive or creative than others. They are better at seeking and discovering novel solutions to problems, being willing to experiment, adapt and instigate change. Others merely follow existing patterns, often resisting change under almost any circumstances. Like the spectrum of light, cognitive equipment consists of a mixture of cognitive skills of varying intensities.

We may classify agents in terms of two extreme forms of behaviour: those who actively search for new possibilities are explorers, and those who prefer to remain with the status quo are sheep. The above mentioned spectrum of cognitive skills implies that we all possess sheep and explorer qualities, albeit in different doses. Pure explorers tend to be imaginative, creative, highly-strung individuals who constantly search for better solutions to the problems they face. They are more inclined to reason inductively, to learn quickly and to adapt willingly to changing circumstances. Sheep are more placid, patient and resigned than explorers. Preferring to reason deductively, they are prone to choosing a well established pattern. They mostly cling to particular beliefs because they have worked well in the past. Sheep are slow learners who must accumulate a record of failure before discarding their favourite beliefs.

In chapter 4, David Batten explores sheep and explorer strategies among fishing fleets searching for profitable fishing zones. Non-equilibrium systems scientists such as Peter Allen, having studied the behaviour of fishing fleets, call the sheep Cartesians and the explorers Stochasts. As Allen notes, the first group makes good use of information, but the second generates it.

Conventional economic wisdom claims that agents have only one reasoning skill: the ability to process the information available to them in a purely logical, deductive manner to arrive at the best decision in a given situation. But this is useless in a self-referential situation like a stock market, where the best thing to do depends on what everyone else is doing. Self-referential situations often arise in human ecosystems, but are rarely recognised as such. A self-referential situation is one in which the forecasts made by the agents involved serve to create the world they are trying to forecast (Batten 2000). In many human decision-making situations, there is no optimal predictor. The best thing each agent can do is to apply the predictor that has worked best so far, then to be willing to re-evaluate the effectiveness of his set of predictors, and to adopt better ones as new information becomes available.

In chapter 4, David Batten develops the mechanisms of such ‘perpetual experiments’ and their consequences in terms of the collective behaviour of what he calls self-defeating systems. Drawing upon the powerful metaphor proposed by Brian Arthur (1994), the El Farol bar problem, the author suggests that many human ecosystems fall into this category and offer agent-based modelling an unchallenged field of exploration. To a certain extent, Ryan McAllister and colleagues (chapter 14) illustrate the ill-fated consequences of such self-defeating systems in the case of privately-owned pastoral regimes in Australian rangelands.

Human ecosystems constitute a subset of complex adaptive systems. They correspond to real life systems characterised by very strong and long-term interactions between human communities and their environment. According to John Holland (1995), these systems display the following characteristics:

According to Stepp et al. (2003), human ecosystems not only process matter and energy flows, but—more specifically—information flows as well. Therefore, they display very specific characteristics. As stated in the previous section, human ecosystems are inherently complex and adaptive, due to the ability of human beings to switch from rational deductive reasoning to inductive pattern recognition. Besides, our ability to communicate and learn from others creates the conditions for co-evolutionary processes in which positive feedback loops follow negative ones, punctuation dispels equilibrium, chaos threatens order, and chance gives a hand to necessity. In chapter 2, Bradbury argues that, until recently, human beings had been able to adapt to changes and to cope with co-evolution through rather simple heuristics. But human activities have gradually strengthened the links—let’s call it globalisation—between loosely connected environments and societies. The author suggests that:

the balance of [our] effective adaptive strategies has shifted decisively and forever from heuristics to what we might call symbioses—the sorts of strategies that evolution favors in closely connected systems, the ones we see today inside cells and organisms, and between symbiots.
(chapter 2, p. 23)

It is indeed remarkable to observe the analogies between what is known as human body functional integrity and what is nowadays called environmental sustainability. More information, more interactions, and shorter communication paths tend to create what David Newth refers to as small worlds in his network-centric analysis of social interactions (chapter 5). The symbiotic dimension of our current understanding of sustainable development is illustrated in two contrasted chapters of this book. In chapter 7, Katherine Daniell and colleagues present a tentative framework to assess sustainable urban development. Taking Christie Walk housing development in Adelaide (Australia) as a case study, the authors insist on the necessity of taking into account water distribution, gas emission, ecosystem health, waste management, economic viability, and social aspiration in order to derive relevant and interrelated sustainability indicators. In chapter 12, Anne Dray and colleagues describe the same necessity of taking into account complex social and spatial interactions in order to help local stakeholders to self-design sustainable water management on the atoll of Tarawa (Republic of Kiribati).

By admitting that many human ecosystems are complex and adaptive, we accept their inherent uncertainty. Indeed, if the system is sufficiently complex, it may not be practical or perhaps even possible to know the details of each local interaction. Obviously, the understanding of system-level patterns is often purchased at a cost:

the observer must usually give up the hope of understanding the workings of causation at the level of individual elements.
(Lansing 2003: 185)

As a matter of fact, uncertainty in human ecosystems can result from two different causes: unpredictable non-linear interactions or ill-defined predicates. The latter—more frequently encountered than usually admitted—relies on our limited ability to infer robust causality links among given sets of elementary processes. For example, Durkheim (1979: 58), in his famous study of suicide, concluded that no matter how much a researcher knows about a collection of individuals:

It is impossible to predict which of them are likely to kill themselves. Yet the number of Parisians who commit suicide each year is even more stable than the general mortality rate.

A process that seems to be governed by chance when viewed at the level of the individual turns out to be strikingly predictable at the level of society as a whole. One would argue that statistics prevail in this case, others would invoke Richard Dawkin’s memes (Dawkins 1976), but we could also admit that we don’t know enough yet about the intimate social dynamics that control such a deviant behaviour. In chapter 10, Pascal Perez and colleagues describe a first attempt to simulate illicit drug use and local markets in Australia. Authors admit that blending together law enforcement, harm reduction, and treatment strategies already represents a daunting challenge; but, trying to infer users' or dealers' behavioural patterns—elusive and secretive by nature—needs to be dealt with through trans-disciplinary and consensual approaches.

On the other hand, unpredictable non-linear interactions are the raison d’etre of complex adaptive systems. The self-referential problem proposed by Arthur (1994) can drive a system to an equivalent situation to the one described by Durkheim. But in the El Farol case, individual behaviours are perfectly deterministic while totally unpredictable for an external observer. Conversely, some perfectly predictable individual behaviours interacting together can lead to unpredictable global behaviour of the whole system. In chapter 11, David Batten and George Grozev provide a clear illustration of the impact of such non-linear interactions in the case of the Australian National Electricity Market (NEM). They describe the NEM as an evolving system of complex interactions between human behaviour in markets, technical infrastructures and the natural environment and propose to explore plausible sustainable futures through agent-based simulations.

In chapter 3, Pascal Perez insists on the inherent uncertainty attached to human ecosystems. He argues that scientists studying these complex adaptive systems fall into two broad categories based on their epistemological postures: Legalists believe in a positivist and experimental science, while Revolutionaries claim that human ecosystems can only be apprehended through a post-normal approach to science. Drawing upon recent theoretical and methodological advances in agent-based modelling, he argues that an uncertainty-compliant science needs to overcome this dichotomy.

In their 2002 book entitled Panarchy: Understanding Transformations in Human and Natural Systems, Gundarson and Holling proposed a holistic and history contingent view of human ecosystems. The panarchy concept describes how a healthy system can invent, experiment, and survive through hierarchies and adaptive cycles that represent ecosystems and socio-ecosystems across scales. Each level of the panarchy operates at its own pace, protected from above by slower, larger levels but invigorated from below by faster, smaller cycles of innovation. The whole panarchy is therefore both creative and conserving. The interactions between adaptive cycles in a panarchy combine learning with continuity (Holling 2001). An obvious strength of the approach is to allow us to view resource management in a more structured way.

Indisputably, this theoretical framework helps clarify the meaning of sustainable development. But critics of the approach point out the direct filiation of panarchy from ecological modelling, in particular dynamical system modelling (DSM). While biological cycles are easy to define, much remains to be said about cycles within human societies. Stepp et al. (2003) provide a compelling list of human characteristics hardly tractable at the level of societies or human ecosystems. The way forward lies in our ability to transcend boundaries between systemic and atomistic approaches. In chapter 13, David McDonald and colleagues demonstrate that it is technically possible to blend system dynamics and individual (or group) behaviours of complex marine environments. In an entirely different context, Ang Yang and colleagues (chapter 8) analyse interconnections between several communication, observation and command networks on a battlefield in order to infer relevant conclusions for modern warfare, including fallible human judgement.