The choice was made at 30,000 feet, flying over Managua, Nicaragua on the way home from a workshop in Venezuela. The workshop was about comparing different disturbed regional systems, by exploring both theory and possible management actions. It was a workshop that brought together a Russian (a grand and wonderful mathematician who, sadly, later got into trouble with Soviet authorities), Canadians, Americans, Venezuelans, Argentineans, and Europeans from different countries. We had a great meeting but were still groping for an appropriate name for the planned book.

The news was full of the recent Managua earthquake and I had just read a review of Bob Kate’s work that emphasized how the poor formal facilities in the city provided little help to survivors of the quake. Instead, they drew upon their extended families, and thereby mobilized the key help needed for nurture and recovery. It became help at the scale of a neighborhood. And the image of that kind of crisis and that kind of recovery acted as a metaphor for the way ecosystems and regional systems function. Suddenly the word “adaptive” popped in my mind, and the name became the one that was beautifully suited for our work- both applied and theoretical.

We used the term adaptive management for the applied aspects of these new ideas, the adaptive cycle to describe the fundamental structure and dynamics of systems and adaptive assessment for the methods. And resilience, along with panarchy (though developed much later) were the words we used that eventually captured the theoretical foundations of the core ideas—non-linearity, surprise, alternative stable states and cross-scale dynamics in space and time.

Theory shaped the emergence of adaptive management. The 1973 ‘resilience’ paper (Holling 1973) really launched the adaptive management work we subsequently developed at the University of British Columbia. Resilience is the ability of a system to experience disturbances , to be changed thereby and then to re-organize and still retain the same basic structure and ways of functioning. It includes the ability to learn from disturbance. Flexibility and break points are at its heart. The precepts of adaptive management were developed as a response to defining an ecosystem in terms of resilience.

A resilient system is forgiving of external shocks. If resilience declines because of resource exploitation and loss of diversity, the magnitude of a shock from which it cannot recover gets smaller and smaller. Resilience shifts attention from growth and efficiency to recovery and flexibility. Growth and efficiency alone can often lead ecological systems, businesses and societies into fragile rigidities, exposing them to completely unexpected turbulent transformation. Resilience, in contrast, adds learning, recovery and flexibility as inherent properties of complex systems . It opens eyes to novelty and new worlds of opportunity.

Growth, as economists see it, is important, but equally so are the resilience forces in a healthy system that dominate during infrequent crises and collapses. And systems are healthy when they can grow for periods but can also generate creative collapses and can renew after collapses. During episodes when growth is halted or reversed, deep uncertainty appears, and alternative futures are unexpectedly perceived. Suddenly, the resulting unpredictability stifles informed action or triggers ignorant and fearful reaction, and there is a search for certainty.

That search for certainty smothers opportunity. Alternatives are suppressed and rigidity increases. Security is what is being sought, independent of evidence to the contrary, and often, when possible, such evidence is masked or hidden. In contrast, adaptive management seeks ways for the system itself to provide clues about opportunities and their consequences by setting up policies that in part provide products and in part are experiments that test causes of uncertainty and suggest solutions. For adaptive management the unknown is ever alive and present, with monitoring a constant need that can always be launched, but is difficult to sustain.

Once some of the relevant theory was developed, it led to more applied phases of investigation, where Carl Walters became a central partner. He was and is a truly brilliant, maverick scientist who walks a non-traditional path that creates new traditions. His work on adaptive management methods has been a classic contribution to the field (Walters 1986) , and more recently, he has advanced our understanding of ecosystem dynamics (Walters and Martell 2004).

The resilience research led a group, largely at British Columbia, to mobilize a series of studies of large-scale ecosystems subject to management, including terrestrial, fresh water and marine. Each study was coordinated with the key scientists involved in the ecosystem, and, in some cases, policy people who “owned” the systems and the data. Typically, several organizations were involved because of the different home bases of participants. The process encouraged two major advances.

One advance was that the set of deep studies allowed a comparative analysis of the theoretical foundations of ecosystem behavior and ecosystem management that was ecological, social and economic. That was the part that was particularly interesting, and it led to the book where the term ‘adaptive management’ was used for the first time (Holling 1978). The second advance was that in the course of conducting these ecosystem studies and comparisons, we developed a sequence of workshop techniques for working with experts in order to develop alternative explanatory models and suggestive policies. In the models, several scales were chosen, based on where we thought the causes lay, and we posed alternative hypotheses for the unknown relationships. Subsequent simulations then showed which, if any, of those alternatives were important in affecting behavior of the integrated system. If they were unimportant they were forgotten; if important they became a focus for further research. The models were then used in a second phase of the workshops to search for effective alternative policies. Three or four extreme policies with contrasting objectives were tested, and then a sustained policy was discovered that balanced economic, social and ecological objectives.

An immense amount was learned from our first experiments, which focused on the beautiful Gulf Islands, an archipelago off the coast of southern British Columbia. We chose to develop a simulation model of recreational property development. I knew little about land speculation, but we made up a marvelous scheme that used my earlier predation equations as the foundation of our modeling exercise—the land of various classes were the “prey”, speculators were the “predators” and a highest bidder auction cleared the market each year. The equations were modifications of the general predation equations (Holling 1988) . The predictions were astonishingly effective and persisted for at least two decades. As much as anything, it reinforced the earlier conclusion that these equations were powerful and general. But the important conclusion concerned the workshop process and the people.

The essence of those workshop methods were fun to present in a critical paper where the workshop processes were described and where key personalities were represented in delightful cartoons drawn by Roy Peterson, a cartoonist in Vancouver, and methods were expressed as a game (Holling and Chambers 1973). It was fun to reveal the truth about characters like Snively Whiplash, The Blunt Scot, The Utopians, the Compleat Amanuensis and The Peerless Leaders in this way. But a reviewer in Ecology turned the manuscript down by saying “no one wants to know about the games people in British Columbia play!” Bioscience reviewers were more enlightened so I happily published there.

I learned that the key design feature for these workshops was to start with two goals. The first goal was long term: to identify large, unattainable goals that can be approached, but never achieved, that relate to fundamental values of freedom, equity, tolerance and education. And then, for the second goal, to add a tough design for the first step, in a way that highlights or creates options to design, later, a second step—and then a third and so on. We found that the results were steps that rapidly covered more ground than could ever be designed at the start. At the heart, that is adaptive design, where the unknown is great, learning is continual and actions evolve. But it is tough for staff of a granting agency, who when they ask what specifics we expected, blanch at being told “wait and see”.

My work always shifted between fundamental theory and applied research. Surprises occurring at one stage became explored in the second, so the old categorization into basic or applied research had no meaning where my work was concerned. Each was intertwined with the other, and each benefited thereby. I found so much of existing theory seriously irrelevant at that time in the 1970’s. It was too simplistic, too static, too uniform in scale , too linear and perceived by the originators as too certain. Traditional ecological practice based on such ideas was therefore grossly ineffective. As an example, it was no wonder that cod collapsed on the east coast of Canada, and since 1992 still shows only a weak sign of recovery even with fishing banned.

We had no difficulty in facing and discussing these issues when people were in workshops. The majority of participants grasped the non-linearity, the thresholds separating different quasi-stable states, the varied spatial patterns at different scales, the inherent uncertainty, the unknown and the necessary complexity of social-ecological systems . When these concepts are understood, the fixed world of standard environmental protection is recognized as being rigid and wrong. Those who got it became the subset of folks in ecology, economics, social science, political science, etc., the ones who could work together to design different solutions as acts of mutual discovery.

That rationale of mutual discovery, developed over 30 years in workshops and in theoretical studies from fish to forests, led us to form an internet organization in 1999 called the Resilience Alliance (http://​www.​resalliance.​org), that combined several groups around the world in collaborative research and collaborative publishing. Fundamentally it was meant to sustain the international cooperation that had emerged in several of these earlier projects, and assist in the continued search for a deeper understanding and ever-broader examples of complexity in nature.

However, when such groups attempt to encourage implementation of adaptive management , the success is often only partial. Carl Walters described the failures well in a paper (Walters 1997) . It is a failure of implementation, not of analysis, evaluation, understanding or policy. I found that two projects I got deeply involved with provided beautiful examples of successes and failures. Both moved to a phase of implementation with variable success, but prior to that, the models , understandings and alternative policies coming from the workshops were central to the initial suspension of regional conflicts in each case.

The first project concerned the Florida Everglades, with Lance Gunderson contributing deep insight and personal experience in the process. The Everglades project was undergoing one of its crises of transformation. That project succeeded in developing an understanding of a system functioning at three different spatial and temporal scales, from sawgrass and tree islands, to slough structures and sugar plantations, to topography. But it was also the example that failed on implementation, because the adaptive experiments were lost and the system became locked into an enormously expensive effort of ecosystem restoration . There was no respected, responsible leader who could survive the political games among the four jurisdictions involved—municipal, regional district, State, and Nation. No one, therefore, could continue with the responsibility to manage a transition. There were just committees of local, state and national government, combined with a good NGO, which became politically active and politically rigid.

The other project concerned forest growth, forest crises and harvesting in the eastern Boreal Forest in the face of spruce budworm outbreaks. This project was housed at the International Institute of Applied Systems Analysis (IIASA), with links to a team in New Brunswick led by Gordon Baskerville. Bill Clark, Dixon Jones, Mike Fiering and I led the effort at IIASA—a wonderful group with a remarkable ability to blend different experiences. Over the centuries, spruce budworm outbreaks periodically swept from Manitoba, through Ontario and Quebec, into New Brunswick and Nova Scotia and still further east to Newfoundland and the eastern U.S. We focused on New Brunswick, Canada, an area just big enough to contain the essential scales and to connect to still larger areas. The project depended on the deep understanding of ecological dynamics that had earlier come from Frank Morris’ classic study of Budworm outbreak dynamics and forest growth (Morris 1963). It succeeded in the sense that non-linearities and cross-scale dynamics were discovered, and we also developed a three-scale simulation model (Holling et al. 1977). Simplified versions of the model (Ludwig et al. 1978) were developed, adaptive experiments were identified and eventually flourished. Three scales of management evolved, monitoring and forest inventory was transformed, and the partner on our project, Gordon Baskerville, became continuing leader of the implementation (Clark et al. 1979).

Some of the leaders in those fields were part of IIASA’s wonderfully innovative early days and found that the small budworm and its up-scale effects presented a rich set of data at a large scale. Howard Raiffa, George Dantzig and Tschalling Koopmans became our partners in this evaluation of the usefulness of those methods for multi-scale ecological/economic systems. The conclusions are reviewed in Bell (1972, 1977).

This leads us to a third revealing example of implementing adaptive management from Carl Folke, now Director of the Beijer Institute of Ecological Economics, and his colleagues. This is the example of a cycle of the continuing transformation of Kristianstads , a small city in southern Sweden, and its wetland landscapes (Olsson et al. 2004). After a major flooding crisis the traditional response of dike building and land draining was rejected and replaced by a vision of land and waterscapes integrated with a variety of peoples’ activities. Solutions, when discovered, often “stayed in the back pocket” until public understanding emerged—a very real process of bottom-up design and implementation. Experiments and education were deeply involved in the transformation, which became a continuing effort with deep and extensive public involvement. A senior leader, Karl Magnusson, orchestrated the process from its beginnings. Vision, practicality, and leadership came together.