Understanding the Complexity of Economic, Ecological, and Social Systems Author(s): C. S. Holling Source: Ecosystems, Vol. 4, No. 5, (Aug., 2001), pp. 390-405 Published by: Springer Science + Business Media Stable URL: http://www.jstor.org/stable/3658800 Accessed: 10/06/2008 15:13

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=ssbm.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We enable the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

http://www.jstor.org Ecosystems(2001) 4: 390-405 DOI:10.1007/s10021-00 -0101-5 ECOSYSTEMS! 0 2001 Springer-Verlag MINIREVIEWS

Understandingthe Complexity of Economic, Ecological, and Social Systems

C. S. Holling

Department of Zoology, University of Florida, Gainesville, Florida 32611, USA

ABSTRACT Hierarchiesand adaptive cycles comprise the basis continuity. An analysis of this processhelps to clar- of ecosystems and social-ecological systems across ify the meaning of "sustainabledevelopment." Sus- scales. Together they form a panarchy. The panar- tainabilityis the capacity to create, test, and main- chy describeshow a healthy system can invent and tain adaptivecapability. Development is the process experiment, benefiting from inventions that create of creating, testing, and maintaining opportunity. opportunity while being kept safe from those that The phrase that combines the two, "sustainablede- ex- destabilizebecause of their nature or excessive velopment," thus refers to the goal of fostering to at its own uberance.Each level is allowed operate adaptivecapabilities and creatingopportunities. It is from above slower, levels pace, protected by larger therefore not an oxymoron but a term that de- but from below faster, smaller invigorated by cycles scribes a logical partnership. of innovation. The whole panarchy is therefore both creative and conserving. The interactions be- Key words: hierarchy; adaptive cycles; multiple tween cycles in a panarchy combine learning with scales; resilience; sustainability.

INTRODUCTION and social/political/economicconflicts. Indicators of sustainabilityhave been identified for all the inter- The ecological status of nations and regions is a nal factors, while issues of concern have been sug- current item action on the for assessment and gested for the external ones. One unpublished re- of several In the United agenda organizations. port cited 76 specificsustainability indicators for the States, the National Academy of Sciences and the internal factors and a more diffuse set of attributes Heinz Center to sus- have issued guidelines identify for the external factors. indicators. the tainability Internationally, Species All of these indicators and all of the attributes Survival Commission of the World Conservation make sense. The is not that are Union has stated that either problem they (IUCN) sustainability, or that are not useful. if in a or of a on interactions wrong, they They are, region species, depends a com- among internal and external factors. The internal anything, incomplete. Rather, they suggest that can overwhelm even factors may be social, political, ecological, or eco- plexity understanding, in a subset of these nomic; the external factors include foreign debt, when, specific situations, only entities are relevant. There are two to structuralpoverty, global environmental problems, approaches complexity. One of them, which has been explored thor- This paper has been adapted from Gunderson and Holling oughly and incisively by Emory Roe (1998), views (2001), with permission of Island Press. as we do not understand,be- Received 7 March2001; accepted 16 March2001. complexity anything *e-mail:[email protected] cause there are apparentlya large number of inter-

390 UnderstandingComplex Systems 391 acting elements. The appropriateapproach, accord- tion of adequate integrative theory, rigorously de- ing to Roe, is to embrace the complexity and veloped. This theory is rooted in empirical reality resulting uncertainty and analyze different subsets and communicated with metaphor and example. of interactions,each of which seem relevant from a The first requirement is to begin to integrate the number of fundamentallydifferent operational and essence of ecological, economic, and social science philosophical perspectives. A recent article in Con- theory and to do so with the goal of being, in servationEcology offered a review of this thesis from Einstein'swords, "as simple as possible but no sim- four different disciplinary and policy perspectives pler." and a commentary on the reviews by the author The purpose of this paper is to summarizea the- (www.consecol.org/Journal/vol4/iss2/index.html). oretical framework and process for understanding An alternative view (Holling 2000; Gunderson complex systems. This concept has recently been and Holling 2001) suggests that the complexity of developed and expanded into a book-length thesis living systems of people and nature emerges not (Gundersonand Holling2001). In its expandedver- from a random association of a large number of sion, it provides a means of assessing information interacting factors rather from a smaller number of about the internal factors and external influences controlling processes. These systems are self-orga- that interact to determine systemic sustainability. nized, and a small set of critical processes create and To be useful, such a framework and process must maintain this self-organization. ("Self-organiza- satisfy the following criteria: tion" is a term that characterizes the development of complex adaptive systems, in which multiple * Be "assimple as possiblebut no simpler"than is outcomes typically are possible depending on acci- requiredfor understandingand communication. dents of history. Diversity and the individuality of * Be dynamic and prescriptive,not static and de- components, localized interactions among compo- scriptive. Monitoring of the present and past is nents, and an autonomous process that uses the static unless it connects to policies and actions outcomes of those local interactions to select a sub- and to the evaluation of different futures. set of those for enhancement are char- components ? Embraceuncertainty and unpredictability.Sur- acteristics of [Levin complex adaptive systems prise and structuralchange are inevitable in These establish a tem- sys- 1999]). processes persistent tems of people and nature. plate upon which a host of other variables exercise their influence. Such "subsidiary"variables or fac- tors can be interesting, relevant, and important,but AN INTEGRATIVETHEORY they exist at the whim of the critical controlling factors or variables. If sustainability means any- Background thing, it has to do with the small set of critical The theory was developed under the auspicesof the self-organized variables and the transformations "ResilienceProject", a 5-year collaborationamong that can occur in them during the evolutionary an international group of ecologists, economists, process of societal development. social scientists, and mathematicians. The project But these two views of complexity require alter- was initiatedto searchfor an integrativetheory and native perspectives and competing models and hy- integrative examples of practice. Its goal was to potheses. The goal of each approachis to mobilize develop and test the elements of an integrative evidence that can distinguishamong competing ex- theory that had the degree of simplicity necessary planations so that multiple lines of evidence begin for understandingbut also the complexity required to define what is known, what is uncertain, and to develop policy for sustainability.The results of what is unknown. We are always left with best that project are summarized in the final report to judgments, not certainties. the MacArthurFoundation found at http://www. The view presented here argues that there is a resalliance.org/reports. requisite level of simplicitybehind the complexity The heart of the work has now been amplifiedin that, if identified, can lead to an understandingthat Panarchy:Understanding Transformations in Human is rigorously developed but can be communicated and NaturalSystems (Gunderson and Holling.2001). lucidly. It holds that if you cannot explain or de- This book expands the theory and explores its im- scribe the issue of concern using at least a handful plications for ecological, political, institutional, and of causes, then your understandingis too simple. If management systems. It was intended to deepen you require many more than a handful of causes, our understanding of linked ecological/economic/ then your understandingis unnecessarilycomplex. decision systems through the use of a set of inter- That level of understandingis built upon a founda- active models, several analyses of institutions that 392 C. S. Holling

Table 1. Table of Contents for Panarchy: Understanding Transformations in Human and Natural Systems

Part I. Introduction Chapter 1. In Quest of a Theory of Adaptive Change. C.S. Holling, L.H. Gunderson,and D. Ludwig Part II. Theories of Change Chapter2. Resilience and Adaptive Cycles. C.S. Holling and L.H. Gunderson Chapter3. Sustainabilityand Panarchies.C.S. Holling, L.H. Gunderson,and G.D. Peterson Chapter4. Why Are Systems of People and Nature not just Ecologicalor Social Systems? F. Westley, S.R. Carpenter, W.A. Brock, C.S. Holling, and L.H. Gunderson Chapter 5. Back to the Future:Ecosystem Dynamics and Local Knowledge. F. Berkes and C. Folke Chapter6. The Dynamics of PoliticalDiscourse in Seeking Sustainability.L. PritchardJr. and S.E. Sanderson Part III. Myths, Models, and Metaphors Chapter7. Collapse,Learning, and Renewal. S.R. Carpenter,W.A. Brock, and D. Ludwig Chapter8. Dynamic Interactionof Societies and Ecosystems:Linking Theories from Ecology, Economy, and Sociology. M. Scheffer, F. Westley, W.A. Brock, and M. Holmgren Chapter9. A Future of Surprises.M. Janssen Chapter 10. Resilience and Sustainability:The Economic Analysis of Non-LinearDynamic Systems. W.A. Brock, K.G. Maler, and C. Perrings Part IV. LinkingTheory to Practice Chapter 11. Resilient Rangelands- Adaptationin Complex Systems. B. Walker and N. Abel Chapter 12. Surprisesand SustainabilityCycles of Renewal in the Everglades.L.H. Gunderson, C.S. Holling, and G.D. Peterson Chapter 13. The Devil in the Dynamics:Adaptive Management on the Front Lines. F. Westley Chapter 14. Planning for Resilience: Scenarios, Surprises,and Branch Points. G. C. Gallopin Part V. Summary and Synthesis Chapter 15. Discoveriesfor SustainableFutures. C. S. Holling, S. R. Carpenter,W. A. Brock, and L. H. Gunderson Chapter 16. Towardsan IntegrativeSynthesis. R. Yorque, B. Walker, C. S. Holling, L. H. Gunderson, C. Folke, S. R. Carpenter,and W. A. Brock

link people and nature, and an extensive explora- the points where it is vulnerable. It then becomes tion of two prototypical systems, the savannas and possible to use those leverage points to foster resil- grasslands of Australia and the Everglades of Flor- ience and sustainability within a system. ida. Table 1 summarizes the book's contents. The idea of panarchy combines the concept of "Panarchy" is the term we use to describe a con- space/time hierarchies with a concept of adaptive cept that explains the evolving nature of complex cycles. I will deal with each in turn and then show adaptive systems. Panarchy is the hierarchical the consequence of combining them in a synthesis. structure in which of nature (for systems example, Hierarchies forests, grasslands, lakes, rivers, and seas), and hu- mans (for example, structures of governance, set- Simon (1974) was one of the first to describe the tlements, and cultures), as well as combined hu- adaptive significance of hierarchical structures. He man-nature systems (for example, agencies that called them "hierarchies", but not in the sense of a control natural resource use) (Gunderson and oth- top-down sequence of authoritative control. ers 1995) and social-ecological systems (for in- Rather, semi-autonomous levels are formed from stance, co-evolved systems of management) (Folke the interactions among a set of variables that share and others 1998), are interlinked in never-ending similar speeds (and, we would add, geometric/spa- adaptive cycles of growth, accumulation, restruc- tial attributes). Each level communicates a small set turing, and renewal. These transformational cycles of information or quantity of material to the next take place in nested sets at scales ranging from a leaf higher (slower and coarser) level. Figure 1 shows to the biosphere over periods from days to geologic an example for a forested landscape, Figure 2 shows epochs, and from the scales of a family to a socio- a wetland system, and Figure 3 shows a social sys- political region over periods from years to centuries. tem. If we can understand these cycles and their scales, it As long as the transfer from one level to the other seems possible to evaluate their contribution to sus- is maintained, the interactions within the levels tainability and to identify the points at which a themselves can be transformed, or the variables system is capable of accepting positive change and changed, without the whole system losing its integ- Understanding Complex Systems 393

100 10 1000 ten hundred cm m m km km one hundredthousand million million I I

4 3 millenium

p 2- 2 century e

0- decade E E o -2 - 0 o0 0 year -4 -

-1 months -6 -

-2 -2 0 2 4 6 log space (meters) 0 2 4 6 8 10 1. Time and scales of the boreal forest Figure space log numberof people (Holling 1986) and the atmosphere (Clark 1985) and their relationshipto some of the processes that structure 3. Institutional of rule sets. In contrast the forest. meso-scale such as in- Figure hierarchy Contagious processes, to ecological hierarchies, this hierarchy is structured sect outbreaksand fire, mediate the interactionbetween dimensions of the number of involved in faster and slower along people atmosphericprocesses vegetation pro- rule sets and turnover times (Gunderson cesses. from Gunderson and 2001 approximate (Reprinted Holling and others 1995; and others 2001). with of Island Westley (Reprinted permission Press) from Gunderson and Holling 2001 with permission of Island 10 100 10 100 1000 Press) m m km km km km

4. understanding by shifting attention from the small- I 2. scale view that characterized much of biological ecology to a multiscale and landscape view that that biotic and abiotic could ._EE ?- recognized processes develop, mutually re-enforcing relationships over co0 -2 distinct ranges of scale. More recently, Levin (1999) has expanded that representation of cross-scale dy- -4 namics in a way that greatly deepens our under- standing of the self-organized features of terrestrial 0 1 2 3 4 5 6 ecosystems. Simon's key arguments are that each of the levels log space (meters) of a dynamic hierarchy serves two functions. One is to conserve and stabilize conditions for the faster Figure2. Time and space scales of levels of a hierarchyin and smaller levels; the other is to generate and test the from Gundersonand Everglades.(Reprinted Holling innovations by experiments occurring within a 2001 with of Island permission Press) level. It is this latter, dynamic function we call "an adaptive cycle" (Holling 1986). It is a heuristic model, a fundamental unit that contributes to the rity. As a consequence, this structure allows wide understanding of the dynamics of complex systems latitude for experimentation within levels, thereby from cells, to ecosystems, to societies, to cultures. greatly increasing the speed of evolution. Ecologists were inspired by Simon's seminal arti- The Adaptive Cycle cle to apply the term "hierarchy" to ecological sys- There are three properties that shape the adaptive tems and develop its significance for a variety of cycle and the future state of a system: ecological relationships and structures. In particu- lar, Allen and Starr (1982) and O'Neill and others *The inherent potential of a system that is avail- (1986) stimulated a major expansion of theoretical able for change, since that potential determines 394 C. S. Holling

the range of future options possible. This prop- erty can be thought of, loosely, as the "wealth" of a system. * The internal controllabilityof a system; that is, the degree of connectedness between internal controlling variables and processes, a measure that reflectsthe degree of flexibility or rigidityof ?.:...... : . . :. such controls, such as their sensitivity or not to perturbation. * The adaptive capacity; that is, the resilience of the system, a measure of its vulnerability to or unpredictableshocks. This prop- unexpected connectedness - erty can be thought of as the opposite of the vulnerability of the system. Figure 4. A stylized representationof the four ecosystem of events them. These three controllability, functions (r, K, Qf,a) and the flow among properties-wealth, the flow in the and ones, whether The arrowsshow the speed of cycle. Short, adaptive capacity-are general arrows indicate a situa- at the scale of the cell or the the individ- closely spaced slowly changing biosphere, tion; arrows indicate a rapidly changing situation. ual or the culture. In case of long examples regional The cycle reflects changes in two properties:the y axis development and ecosystem management (Gun- (the potential that is inherent in the accumulated re- derson and others 1995), they are the properties sources of biomass and nutrients) and the x axis (the that shape the responses of ecosystems, agencies, degreeof connectednessamong controllingvariable). The and people to crisis. exit from the cycle indicated at the left of the figure Potential, or wealth, sets limits for what is possi- suggests,in a stylized way, the stage where the potential ble-it determines the number of alternative op- can leak away and where a flip into a less productiveand tions for the future. Connectedness,or controllabil- less organizedsystem is most likely (Holling 1986). (Re- from Gundersonand 2001 with ity, determines the to which a system can printed Holling permis- degree sion of Island control its own destiny, as distinct from being Press) caught by the whims of external variability.Resil- ience, as achieved by adaptive capacity, determines how vulnerable the system is to unexpected distur- and integrated during the progression from r to K. bances and surprisesthat can exceed or break that They also represent a potential that was developed control. and used in one setting but could be available in A stylized representation of an adaptive cycle is transformedones. shown in Figure 4 for two of these properties- As the progressionto the K phase proceeds in an potential and connectedness. The trajectory alter- ecosystem, for example, the accumulating nutrient nates between long periods of slow accumulation and biomass resources become more and more and transformationof resources (from exploitation tightly bound within existing vegetation, prevent- to conservation, or r to K), with shorterperiods that ing other competitors from utilizing them. The po- create opportunitiesfor innovation (from release to tential for other use is high, but it is expropriated reorganization,or f to a). That potential includes and controlledby the specific biota and processesof accumulated ecological, economic, social, and cul- the ecosystem in place. That is, the system's con- tural capital as well as unexpressed chance muta- nectedness increases, eventually becoming over- tions and inventions. During the slow sequence connected and increasingly rigid in its control. It from exploitation to conservation, connectedness becomes an accident waiting to happen. and stability increase and capital is accumulated. The actual change is triggered by agents of dis- Ecosystem capital, for example, includes nutrients, turbance, such as wind, fire, disease, insect out- biomass, and physical structure.Although this ac- break,and drought.The resources accumulatedand cumulated capital is sequestered for the growing, sequestered in vegetation and soil are then sud- maturing ecosystem, it also represents a gradual denly released and the tight organization is lost. increase in the potential for other kinds of ecosys- Human enterprisescan exhibit similarbehavior, as, tems and futures. For an economic or social system, for example, when corporations such as IBM, the accumulating potential could as well derive AT&T,or General Motors accumulate rigidities to from the skills, networks of human relationships, the point of crisis and then attempt to restructure and mutual trust that are developed incrementally (Hurstand Zimmerman 1994; Hurst 1995; Holling Understanding Complex Systems 395 and others 2001). The Soviet Union is a societal example of accumulated rigidities that precipitate a sudden collapse. The proximate agents of distur- bance in these cases can be stakeholder revolts, public-interest attacks through the legal system, or more extreme societal revolts. The phase from fl to oLis a period of rapid reor- ganization during which novel recombinations can unexpectedly seed experiments that lead to inno- rotation vations in the next cycle. The economist J. A. reveals Schumpeter (1950) appropriately called this phase resilience "creative destruction." Initially, the "front loop" of the trajectory, from r to K, becomes progressively more predictable as it develops. In contrast, the "back loop" of the adaptive cycle, from fl to a, is inherently unpredictable and highly uncertain. At that stage, the previously accumulated mutations, inventions, external invaders, and capital can be- come reassorted into novel combinations, some of which nucleate new opportunity. Figure 5. Resilienceis another dimension of the adaptive It is as if two separate objectives are functioning, cycle. A third dimension, resilience, is added to the two- but in sequence. The first maximizes production dimensional box of Figure 4 to show how resilience ex- and accumulation; the second maximizes invention pands and contracts throughout the cycle. Resilience and reassortment. The two objectives cannot be shrinks as the cycle moves towards K, where the system maximized simultaneously but only occur sequen- becomes more brittle.It expands as the cycle shifts rapidly into a back to accumulated resources for tially. And the success in achieving one inexorably loop reorganize a new initiation of the cycle. The appearanceof a figure 8 sets the stage for its opposite. The adaptive cycle in Figure4 is the consequence of viewing a three-dimen- therefore embraces two opposites: growth and sta- sional object in a two-dimensional plane. (Reprinted on the one hand, and on the bility change variety from Gunderson and Holling 2001 with permission of other. Island Press) Figure 5 adds the third dimension, resilience, to the adaptive cycle. The appearance of a figure 8 in the path of the adaptive cycle, as in Figure 4, is the consequence of the projection of a three-dimen- doxes of conservative nature vs creative nature; sional object onto a two-dimensional plane. We can sustainability vs creative change. view that three-dimensional object from different The cx phase is the stage that is least examined perspectives, emphasizing one property or another. and the least known. It is the beginning of a process Figure 5 rotates the object to expose the resilience of reorganization that provides the potential for axis. subsequent growth, resource accumulation, and This orientation of the figure shows that as the storage. At this stage, ecological resilience is high, as phases of the adaptive cycle proceed, a system's is potential. But connectedness is low and internal ecological resilience expands and contracts. The regulation is weak. There is a wide stability region, conditions that occasionally foster novelty and ex- with weak regulation around equilibria, low con- periment occur during periods in the back loop of nectivity among variables, and a substantial the cycle, when connectedness, or controllability, is amount of potential available for future options. low and resilience is high (that is, during the ao Because of those features, it is a fertile environment phase). The low connectedness, or weak control, for experiments, for the appearance and initial es- permits novel reassortments of elements that were tablishment of entities that would otherwise be out- previously tightly connected to others in isolated competed. As in good experiments, many will fail, sets of interactions. The high resilience allows tests but in the process, the survivors will accumulate the of those novel combinations because the system- fruits of change. It is a time of both crisis and wide costs of failure are low. The result is the con- opportunity. dition needed for creative experimentation. This In summary, there are four key features that recognition of resilience varying within a cycle adds characterize an adaptive cycle, with its properties of an element that can reconcile the delicious para- growth and accumulation on the one hand and of 396 C. S. Holling novelty and renewal on the other. All of them are the a and r phases, dominated by trophic dy- measurablein specific situations: namics (Walkerand Abel 2001). 1. Potential (that is, wealth as expressed in eco- Ecosystemsand human organizationswith pre- system structure,productivity, human relation- dictablebut variableinputs and some significant ships, mutations, and inventions) increases in- internal regulation of external variability over crementally in conjunction with increased certain scale ranges. For example, productive efficiency but also in conjunction with in- temperate forests and grasslands,large bureau- creased rigidity.This is the phase from r to K in cracies.These systems represent the full cycle of Figure 4. boom-and-bust dynamics shown in Figure 4 (Holling and Gunderson 2001). 2. As potential increases, slow changes gradually expose an increasing vulnerability (decreased * Biological entities with strong and effective ho- resilience) to such threats as fire, insect out- meostatic internal regulation of external vari- break, competitors, or opposition groups. The ability. Examples: cells and ionic regulation, system becomes an accidentwaiting to happen. "warm-blooded"organisms with endothermic of A break can trigger the release of accumulated control temperature.System variablesremain near an potential in what the economist Schumpeter equilibriumand the individual is freed to a wider called "creativedestruction" (1950). The trajec- exploit range of opportunitieswithin a or This is an tory then moves abruptlyinto a back loop from community ecosystem. example K to Q. of local control that can release external oppor- tunity and variability at a different scale-a 3. Innovation occurs in pulses or surges of inno- transfer of the full adaptive cycle to the larger vation when uncertainty is great, potential is arena of a higher level in the hierarchy. high, and controls are weak, so that novel re- * Human with and active combinations can form. This is the phase of systems foresight adap- tive methods that stabilize and reorganization represented in a (Figure 4) variability exploit busi- where low connectedness allows unexpected opportunity. Examples: entrepreneurial nesses, futures markets and resource combinations of previously isolated or con- scarcity, some traditionalcultures. The of strainedinnovations that can nucleate new op- high variability the can be transferredfrom the portunity. adaptive cycle society to an individual entrepreneur or, in a 4. Those innovations are then tested. Some fail, traditionalculture, to a "wise person" but others survive and in a (Westley adapt succeeding and others 2001; Berkes and Folke 2001). phase of growth from r to K.

Not All Adaptive Cycles Are the Same THE PANARCHY:A SYNTHESIS Effortsto find exceptions that might invalidate the Because the word "hierarchy"is so burdenedby the preceding representation have identified different rigid, top-down nature of its common meaning, we classes of systems that representdistinct variants of, decidedto look for another term that would capture or departures from, that cycle. Examples of these the adaptive and evolutionary nature of adaptive exceptions include: cycles that are nested one within each other across and time scales. Our was to rationalize * where the lack of invention space goal Physical systems the between and be- and mutation limits the for evolution- interplay change persistence, potential tween the and the We tectonic predictable unpredictable. ary change. Examples: plate dynamics, thereforemelded the of the Greek Pan as and Per Bak's sand image god (1996) pile experiments the epitoma of with the no- from K to unpredictablechange demonstrating "organizedcriticality" tion of hierarchies across scales to invent a new Q2). term that could represent structures that sustain * Ecosystems and communities of plants and ani- experiment, test its results, and allow adaptiveevo- mals that are strongly influenced by uncontrol- lution. Hence, "panarchy". lable or unpredictable episodic external inputs A panarchyis a representationof a hierarchyas a and have little internal regulation and highly nested set of adaptive cycles. The functioning of adaptive responses to opportunity. Examples: those cycles and the communicationbetween them exploited arid rangelands,pelagic biotic commu- determines the sustainabilityof a system. That syn- nities. These systems tend to remain largely in thesis will be explored in this section. the lower left quadrantof the cycle, oscillatingin The adaptive cycle, as shown in Figures4 and 5, Understanding Complex Systems 397 transforms hierarchies from fixed static structures 100 10 1000 to dynamic, adaptive entities whose levels are sen- cm m m km km A I I I I ! sitive to small disturbances at the transition from 4 f ' * ! f 10,000 years growth to collapse (the Qfphase) and the transition from reorganization to rapid growth (the a phase). -century U)L- At other times, the processes are stable and robust, cncD constraining the lower levels and immune to the - - 10 years buzz of noise from small and faster processes. It is at 1-- '...... i~...... ~...... ~~-..- *.5E the two-phase transitions between gradual and 0 -year rapid change and vice versa that the large and slow months entities become sensitive to change from the small 0_ i and fast ones. However, the structural, top-down aspect of hi- erarchies has tended to dominate theory and appli- -2 0 2 4 6 cation, reinforced by the standard dictionary defi- log space (meters) nition of hierarchy as a system of vertical authority and control. Therefore, the dynamic and adaptive Figure 6. A stylized panarchy. A panarchy is a cross- nature of such nested structures has tended to be scale, nested set of adaptive cycles that indicates the lost. dynamic nature of structures depicted in the previous It is true that slower and levels set plots. (Reprintedfrom Gundersonand Holling 2001 with certainly larger of Island the conditions within which faster and smaller ones permission Press) function. Thus, a forest stand moderates the climate within the stand to narrow the range of tempera- tures experienced by its individuals constituents. periodically within each hierarchical level, in a way Similarly, cultures of different people establish that partially isolates the resulting experiments, re- norms that guide the actions of human individuals. ducing the risk to the integrity of the whole struc- But this representation has no way of accounting ture. for the dynamics of each level as symbolized in the The organization and functions that form biolog- four-phase cycle of birth, growth and maturation, ical, ecological, and human systems can therefore death, and renewal. be viewed as a nested set of four-phase adaptive This adaptive cycle captures in a heuristic fashion cycles. Within these cycles, there are opportunities the engine that periodically generates the variability for periodic reshuffling within levels, which main- and novelty upon which experimentation depends. tain adaptive opportunity, while simple interactions As a consequence of the periodic, but transient, across levels maintain integrity. One major differ- phases of creative destruction (Qf stage) and re- ence among biological, ecological, and human sys- newal (oxstage), each level of a system's structure tems is the way that inventions are accumulated and processes can be reorganized. This reshuffling and transferred over time. But more on that later. in the back loop of the cycle allows the possibility of There are two features that distinguish the pan- new system configurations and opportunities utiliz- archical representation from traditional hierarchical ing the exotic and entirely novel entrants that had ones. The first, as discussed earlier, is the impor- accumulated in earlier phases. The adaptive cycle tance of the adaptive cycle and, in particular, the cx opens transient windows of opportunity so that phase as the engine of variety and the generator of novel assortments can be generated. new experiments within each level. The various For organisms, those novel entrants are mutated levels of the panarchy can be seen as a nested set of genes or, for some bacteria, exotic genes that are adaptive cycles (Figure 6). transferred occasionally between species. For eco- The second feature is the connections between systems, the novel entrants are exotic, potentially levels. There are potentially multiple connections invasive species or species "in the wings" waiting between phases at one level and phases at another for more appropriate conditions. For economic sys- level. But two of these connections are particularly tems, these novel entrants are inventions, creative significant to our search for the meaning of sustain- ideas, and innovative people. The adaptive cycle ability. They are labeled as "revolt" and "remem- explicitly initiates a slow period of growth during ber" in Figure 7, where three levels of a panarchy which mutations, invasions, and inventions can ac- are represented. The revolt and remember connec- cumulate, followed by a briefer period when they tions become important at times of change in the undergo rearrangements. This process can occur adaptive cycles. 398 C. S. Holling

crown of a tree, then to a patch in the forest, and then to a whole stand of trees. Each step in that cascade moves the transformation to a larger and slower level. A societal version occurs when local and slow activists succeed in their efforts to transform re- gional organizations and institutions, because the latter have become broadly vulnerable. Such a change occurred in New Brunswick, Canada when a few small groups opposed to spraying insecticide intermediate over the forest were able to transform this region's size and speed vulnerable forest management policies and prac- tices (Baskerville 1995). The arrow labeled "remember" in Figure 7 indi- cates a second type of cross-scale interaction that is important at times of change and renewal. Once a andfast catastrophe is triggered at one level, the opportuni- ties for, or constraints against, the renewal of the are influenced the K of the Figure 7. Panarchicalconnections. Three selected levels cycle strongly by phase of a panarchyare illustratedto show the two connections next slower and larger level. After a forest fire, for that are criticalin creating and sustaining adaptive capa- example, the processes and resources that have ac- bility. One is the "revolt"connection, which can cause a cumulated at a larger level slow the leakage of criticalchange in one cycle to cascade up to a vulnerable nutrients that have been mobilized and released stage in a larger and slower one. The other is the "re- into the soil. At the same time, the options for member"connection which facilitatesrenewal by draw- renewal include the seed bank, physical structures, on the that has been accumulated and ing potential and surviving species, which comprise biotic lega- stored in a larger, slower An of the se- cycle. example cies (Franklin and MacMahon 2000) that have ac- quence from small and fast through largerand slower and cumulated in the course of the forest's growth. thence to largestand slowest for a boreal forest ecosystem for its and renewal, a includes needles, tree crowns, and patches. For institu- Similarily, reorganization coral reef hit a storm draws on its own tions, those three speeds might be operationalrules, col- by legacies lective choice rules, and constitutional rules (Ostrom and the memory of the seascape of which it is a part 1992); for economies, they might be individual prefer- (Nystr6m and Folke 2001). It is as if this connection ences, markets, and social institutions (Whitaker 1987); draws on the accumulated wisdom and experiences for developing nations, they might be markets,infrastruc- of maturity; hence, the word "remember." ture, and governance (Barro 1997); for societies, they In a similar vein, Stewart Brand, in his marvelous be allocation might mechanisms, norms, and myths meditation on buildings (1994), described them as for be (Westley 1995); knowledge systems, they might adaptive, hierarchical entities. Buildings of endur- local knowledge, management practice, and world view ing character are a reflection of seasoned maturi- (Gadgiland others 1993; Berkes 1999; Hollingand others ty-the culmination of a series of idiosyncratic, 2001). (Reprinted from Gunderson and Holling 2001 wise, and in the with permission of Island Press) thought-provoking experiments form and content of a mature, evolved structure. In The Clock of the Long Now, Brand (1999) extends When a level in the panarchy enters its Q phase these ideas and generalizes the concept of fast and of creative destruction, the collapse can cascade to slow processes to society as a whole. His work res- the next larger and slower level by triggering a onates with features reminiscent of panarchy the- crisis. Such an event is most likely if the slower level ory. Similarly, Levin's Fragile Dominion (1999) is an is at its K phase, because at this point the resilience accessible and effective disquisition on self-organi- is low and the level is particularly vulnerable. The zation as it characterizes adaptive, complex ecolog- "revolt" arrow in Figure 7 suggests this effect, one ical systems. where fast and small events overwhelm slow and The panarchy is a representation of the ways in large ones. Once triggered, the effect can cascade to which a healthy social-ecological system can invent still higher, slower levels, particularly if those levels and experiment, benefiting from inventions that have also accumulated vulnerabilities and rigidities. create opportunity while it is kept safe from those An ecological version of this situation occurs that destabilize the system because of their nature when conditions in a forest allow a local ignition to or excessive exuberance. Each level is allowed to create a small ground fire that spreads first to the operate at its own pace, protected from above by UnderstandingComplex Systems 399 slower, largerlevels but invigoratedfrom below by Similarly,a long view of human history reveals faster, smallercycles of innovation. The whole pan- not regularchange but spasmodic,catastrophic dis- archy is thereforeboth creativeand conserving. The ruptions followed by long periods of reinvention interactionsbetween cycles in a panarchy combine and development. In contrast to the sudden col- learning with continuity. lapses of biologicalpanarchies, there are long peri- This process can serve to clarify the meaning of ods of ruinous reversal, followed by slow recovery "sustainabledevelopment". Sustainability is the ca- and the restoration of lost potential. Robert Ad- pacity to create, test, and maintain adaptive capa- ams's magnificentreconstruction of Mesopotamian bility. Development is the process of creating, test- societies (1966, 1978) and a later review of other ing, and maintaining opportunity.The phrase that archaeologicalsequences at regional or largerscales combines the two, "sustainable development", (R. M. Adamsunpublished) led him to identifytwo therefore refers to the goal of fostering adaptive trends in human society since the Pleistocene.The capabilitieswhile simultaneouslycreating opportu- .firstis an overall increase in the hierarchicaldiffer- nities. It is therefore not an oxymoron but a term entiation and complexity of societies. That is, levels that describesa logical partnership. in the panarchy are added over time. If enough potential accumulates at one level, it can pass a Collapsing Panarchies threshold and establish another, slower and larger Stochastic events external to a cycle can trigger level. The second trend is defined by the occurrence spasmodic collapses, particularlyif they encounter of rapid discontinuous shifts, interspersedby much vulnerabilitieswithin an adaptive cycle. Extremely longer periods of relative stability. A number of large events can overwhelm the sustaining proper- scholarshave focused on the study of such societal ties of panarchies, destroyinglevels, and triggering dynamics in more recent history. For example, destructivecascades down the successive levels of a Goldstone (1991) examined the wave of revolu- panarchy. The cataclysmicloss of biological diver- tions that occurredin Eurasiaafter a period of calm sity that occurred some 65 million years ago, de- in the 17th century. He hypothesized that political stroying about 70% of Earth's species; Jablonski breakdown occurs when there are simultaneous 1995), for example, is likely to have been caused by crises at several different organizational levels in the impact of an asteroid (Alvarezand others 1980). society. In other words, adaptive cycles at different That event, which may also be associatedwith mas- levels in a panarchy become aligned at the same sive volcanic eruptions that occurred around the phase of vulnerability.Thus, he explicitly posits a same time, unraveled the web of interactions cascading,panarchical collapse. within and between panarchicallevels over scales In TheGreat Wave, David Fischer (1996) presents from biomes to species. a somewhat similar model of political breakdown Since recovery from these events is so delayed, it that focuses less on social stratificationand revolu- is likely that mass extinction events eliminate not tionary dynamics than on empiricalprice data and only species but also ecological niches. For their inflation. Accordingto Fischer, at least three waves continued existence, species depend on an environ- of social unrest swept Eurasia, first in the 14th ment that is created by life. Because they destroy century and later in the 17th and late 18th centu- most species, mass extinction events concomitantly ries. He argues that currency mismanagementand eliminate many ecological niches. The recovery of the outbreak of diseases aggravatedthe destabiliz- biodiversity from such cataclysmicevents requires ing effects of an inflation that in turn was drivenby the reconstruction of these niches, as new species population growth. evolve to fill them. In effect, both of these models of societal change Notably, different families, orders, and species propose that slow dynamics inform social organiza- dominated the new assemblages after recovery; tion. Periods of success carry the seeds of subse- novel inventions and new ways of living emerged. quent downfall, because they allow stresses and The dinosaurs became extinct during the collapse rigiditiesto accumulate. Organizationsand institu- that occurred 65 million years ago; the mammals, tions often fail to cope with these slow changes inconspicuous before that time, exploded in a di- either because the changes are invisible to them, or versification that created new opportunity. The they are so complex and highly contested that no conservative nature of established panarchies cer- action can be agreed upon. tainly slows change, while at the same time accu- Modern democraticsocieties are clearlyvulnera- mulating potential that can be released periodically ble to the same process, but they have invented if the decks are clearedof constraininginfluences by ways to diffuse large episodes of creative destruc- large, extreme events. tion by creating smaller cycles of renewal and 400 C. S. Holling

or due to an external force, an impoverished state can result, with low connectedness, low potential, t and low resilience, thus creating a poverty trap. This condition can then propagate downward through levels of the panarchy, collapsing levels as it goes. An ecological example is the productive savanna that, through human overuse and misuse, flips into an irreversible, eroding state, beginning with sparse vegetation. Thereafter, subsequent drought precipitates further erosion, and economic disincentives maintain sheep production. The same persistent collapse might also occur in a society traumatized by social disruption or conflict, so that its cultural cohesion and adaptive abilities are lost. In such a situation, the individual members of the society would be able to depend only on themselves and perhaps their immediate family members. Some such societies might continue to exist in this degraded state of bare subsistence, barely able to persist as a group, but unable to accumulate enough potential to form the larger structures and Figure 8. Maladaptive systems. A poverty trap and a of a complete panarchy. Oth- rigiditytrap are illustratedas departuresfrom an adaptive sustaining properties ers into Berkes cycle. If an adaptivecycle collapsesbecause the potential might simply collapse anarchy. and diversity have been eradicateddue to misuse or an (1999) and Folke and others (1998) tried to deter- external force, an impoverishedstate can result, with low mine how far such erosion must progress before connectedness, low potential, and low resilience, thus recovery becomes impossible. When recovery is creating a poverty trap. A system with high potential, possible, it would be useful to know what critical connectedness, and resilience is representedby the rigid- attributes need to be reinvented and reestablished ity trap. It is suggestive of the maladaptive conditions from the residual memory stored in slowly fading in such as bureaucracies present hierocracies, large traditions and myths to recreate a new, sustaining and others from Gunderson (Holling 2001). (Reprinted panarchy. and Holling 2001 with permission of Island Press) Figure 8 also suggests that it is possible to have a sustainable but maladaptive system. Imagine a sit- uation of great wealth and control, where potential change through periodic political elections. So long is high, connectedness great and-in contrast to the as there is a literate and attentive citizenry, the phase where those conditions exist in an adaptive painful lessons learned from the episodic collapses cycle-resilience is high; that is, a wealthy, tightly of whole societal panarchies can be transferred to regulated, and resilient system. The high resilience faster learning at smaller scales. Various designs in would mean that the system had a great ability to business, from the creation of "skunk works" to the resist external disturbances and persist, even be- introduction of total quality management, serve the yond the point where it is adaptive and creative. It same purpose. would have a kind of perverse resilience, preserving a The potential would be and maladaptive system. high Poverty Traps Rigidity Traps measured in accumulated wealth or abundant nat- Collapsing panarchies begin to decline within spe- ural capital. The high connectedness would be cre- cific adaptive cycles that have become maladaptive. ated by efficient methods of social control, in which Earlier, I described the path of an adaptive cycle as any novelty is either smothered or its inventor oscillating between conditions of low connected- ejected. It would represent a rigidity trap. ness, low potential, and high resilience to their We see signs of such sustained but maladaptive opposites. Could there be systems with other com- conditions in great "hierocracies," such as societies binations of those three attributes in which variabil- that operate under rigid and apparently immutable ity is sharply constrained and opportunity is lim- caste systems. Other examples occur in regions of ited? We suggest two such possibilities in Figure 8. the developing world that have abundant natural If an adaptive cycle collapses because the potential resources but are subject to the rigid control of and diversity have been eradicated through misuse corrupt political regimes. But all such systems are UnderstandingComplex Systems 401 likely to have the seeds of their own destruction dency for large organizationsto develop rigidities, built in, as was the case with the totalitarianbu- thus precipitatingmajor crises that initiate restruc- reaucracyof the now defunct Soviet Union (Levin turing in a larger social ecological, economic set- and others 1998). ting? Or, the many examples of long-term, ruinous reversals in the of These What Human development societies? Distinguishes Systems? collapses seem to be more extreme and require Human systems exhibit at least three features that much longer recovery than the internallygenerated are unique-features that change the characterand cycles of ecosystem panarchies. location of variabilitywithin the panarchy and that Certainly, in management agences, the exercise can dramaticallyenhance the potential of the pan- of foresight and intentionality is often brilliantly archies themselves. Those three features are fore- directed to protect the positions of individuals sight, communication, and technology. ratherthan to furtherlarger societal goals. The fore- Foresightand intetionality. Human foresight and sight that maintains creativity and change when intentionality can dramatically reduce or even connected to an appropriateeconomic market can eliminate the boom and bust character of some lead to rigid organizationsthat are maintainedeven cycles. Predictionsof looming economic crises and when that particularmarket no longer exists. The collapses caused by resource scarcity, for example, marketin these cases is a marketfor politicalpower are an importantissue in debates about sustainabil- of the few, not a free market for the many (Prit- ity. The economist R. Solow (1973) provided a chard and Sanderson 2001). Foresight and inten- withering critique of such doomsday scenarios, tionality can therefore precipitateruinous reversals pointing out that they ignore the forward-looking if they are not connected to a market with essential behaviors of people. These behaviors play a role in liberal and equitable properties. transmitting future scarcities into current prices, Communication,Organisms transfer, test, and thereby inducing conservationbehaviors in the real store experience in a changing world genetically. economic world. This forward-looking process Ecosystems transfer, test, and store experience by functions through futures marketsand the strategic forming self-organized patterns that repeat them- purchaseand holding of commodities.They provide selves. These patterns are formed and refined by a very largeincentives for some people to forecastthe set of interacting variablesthat function over spe- coming scarcitybetter than the rest of the market cific scale ranges and form a mutually reinforcing and to take a position to profit from it. But what core of relationships.In fact, an ecosystem is devel- one marketparticipant can do, all can do; thus, this oped out of a few such sets that establish a repro- process transmits information to the market as a ducing, discontinuous template to provide niches whole. for species diversificationand the adaptationof in- But there limits to this process, as describedby dividual organisms. Carpenterand others (1999, 2001). These limits are In human systems, the same self-organizedpat- illustratedin specific examples of models that com- terns are strongly developed, but humans uniquely bine ecosystem simulationswith economic optimi- add the ability to communicate ideas and experi- zation and decision processes.These models suggest ence. As they are tested, these ideas can become that even when knowledge is total, a minimally incorporated into slower parts of the panarchy, complex ecosystem model together with stochastic such as cultural myths, legal constitutions, and events, can thwart the forward-looking economic laws. Many sources of information,including tele- and decision-making capacity to eliminate booms vision, movies, and the Interet, are global in their and busts. These minimal requirementsfor the sys- connectednessand influence. These media are con- tem are the same ones that characterizethe ecosys- tributingto a transformationof culture, beliefs, and tem panarchy-that is, at least three speeds of vari- politics at global scales. ables, separation among those speeds, and Technology. The scale of the influence exerted nonlinear, multistablebehavior. by every animal other than humans is highly re- That analysis is the source of our conclusion that stricted. But technology amplifies the actions of ecosystems have a minimal complexity we call the humans so that they affect an astonishing range of "Ruleof Hand"whose featuresmake linear polices scales from the submicroscopic to global and- more likely to produce temporary solutions and a however modestly at the moment-even extend greater number of escalating problems. Only an beyond Earthitself. actively adaptive approachcan minimize the con- As human technology has evolved over the last sequences. hundred thousand years, it has progressivelyaccel- Finally, how can we explain the common ten- erated, changing the rules and context of the pan- 402 C. S. Holling

archies in the process. The specializedtools, habita- cycles of innovation. The whole panarchy is there- tion, and weapons of hunter-gatherers, for fore both creative and conserving. The interactions example, together with the domestication of ca- between cycles in a panarchy combines learning nines for use as hunting companions, created op- with continuity. portunities over wide scales. The use of fire by early The four R's, then, represent the criticalprocesses humans made them part of the ecological structur- that manage the balance and tension between ing process. In temperate North America and Aus- change and sustainability. tralia, for example, they became capable of trans- It is often useful to begin the analysis of a specific forming mosaics of grasslands and woods into problem with a historical reconstruction of the extensive regions of contiguous grasslands or for- events that have occurred,focusing on the surprises ests (Flannery 1994). and crises that have arisen as a result of both ex- Progressively,the horse, train, automobile, and ternal influences and internal instabilities. In es- aircrafthave extended the ambit for human choices sence, a sequence of adaptive cycles can be de- from local to regional and thence to planetary scribed, for the so-called natural system, the scales, but the time allotted for each of these choices economy, management agencies, users, and poli- has changed little, or even decreased.Trips between tics. We think it is necessary to consider three scale home and work, for example, have always been ranges for each system, although the particular largely limited to less than an hour or so, although scales might be different for different subsystems. the spatial scale has expanded from a maximum of One of the principalaims is to define where in their a few kilometers by foot to potentially a few hun- respectiveadaptive cycles each of the subsystemsis dred kilometers by commuter aircraft.The slope of now. Actions that would be appropriate at one the decision panarchy for humans, if plotted in the phase of the cycle might not be appropriateat other same space as in Figures 1-3, now angles sharply phases. Knowing where you are helps you to define upward, intersecting and dominating other panar- what action needs to be taken. chies of nature. In many instances, the motive for an assessment is a crisis or transformation that has already oc- Assessing Sustainability curred or is anticipated. In these situations, the The current state of our understanding of panar- conditions of the back loop of the adaptive cycle chies is summarizedin Table 2. The theory is suffi- (Figure4) dominate. However, it is these times of ciently new that its practicalapplication to regional greatest threat that offer the greatest opportunity, questions or the analysis of specific problems has because many constraints have been removed. In just begun. Panarchytheory focuses on the critical an insightful analysis of local communities as seen features that affect or trigger reorganization and from this perspective, Berkes and Folke (2001) transformationin a system. First, the back-loop of showed that local societies often develop reserves the cycles is the phase where resilience and oppor- that are necessary during back-loop restructuring. tunity is maintained or created, via "release"and In the same book, Westley (2001) presented an "reorganization"(Figures 4 and 5). Second, the equally incisive analysis of a sequence of decisions connections between levels of the panarchy are and actions taken in specific examples of problem where persistence (via "remembrance")and evolv- solving by a resourcemanager. Figure 9 providesan ability (via "revolt")(Figure 7) are maintained. example of the kind of analysis that is possible. These four phases or processes make up the four Such transformations across scales are qualita- R's of sustainabilityand development: release, re- tively different from the incremental changes that organization, remembrance, and revolt. They pro- occur during the growth phase of the adaptive cy- vide new categoriesthat can be used to organize the cle. They are also qualitatively different from the more specific indicators and attributesdiscussed in potentially more extreme changes and frozen acci- documents aimed at finding ways to evaluate sus- dents that can occur during the more revolutionary tainability and development. shift from creative destruction (1l) to renewal (a). To summarize: The panarchy describes how a These transformationscascade and transform the healthy socioecological system can invent and ex- whole panarchy along with its constituent adaptive periment, benefiting from inventions that create cycles. opportunity while it is kept safe from those that Because a unique combination of separatedevel- destabilizethe system due to their nature or exces- opments has to conspire to occur simultaneously, sive exuberance. Each level is allowed to operate at extreme events are rare. Some developments its own pace, protectedfrom above by slower, larger emerge within adaptive cycles during the back loop levels but invigoratedfrom below by faster, smaller of the cycle, when recombinations and external Understanding Complex Systems 403

Table 2. Summary Findings from the Assessment of Resilience in Ecosystems, Economies, and Institutions

Statement Brief Explanation

Multistablestates are common in many systems. Abrupt shifts among a multiplicityof very different stable domains are plausiblein regional ecosystems, some economic systems, and some political systems. The adaptive cycle is a fundamentalunit of dynamic An adaptive cycle that aggregatesresources and that change. periodicallyrestructures to create opportunitiesfor innovation is a fundamentalunit for understandingcomplex systems, from cells to ecosystems to societies to cultures. Not all adaptive cycles are the same and some are Variantsto the adaptivecycle are present in differentsystems. maladaptive. These include physical systems (because of the absence of mutations of elements), ecosystems strongly influenced by external pulses, and human systems with foresight and adaptive methods to stabilizevariability. Some systems are maladaptiveand triggerpoverty and rigiditytraps. Sustainabilityrequires both change and persistence. We propose that sustainabilityis maintained by relationships that can be interpretedas a nested set of adaptive cycles arrangedas a dynamichierarchy in space and time-the panarchy. Self-organizationshapes long-term change. Self-organizationof ecological systems establishesthe arena for evolutionarychange. Self-organizationof human institutionalpatterns establishes the arena for future sustainableopportunity. There are three types of learning. Panarchiesidentify three types of change, each of which can generate a differentkind of learning: (a) incremental (r to K, Figure 4), (b) lurching, (Qfto t, Figure 4), and (c) transforming. The world is lumpy. Attributesof biologicaland human entities form clumped patterns that reflect panarchicalorganization, create diversity,and contributeto resilience and sustainability. Functional diversitybuilds resilience. Functionalgroups acrosssize classes of organismsmaintain ecosystem resilience. Tractabilitycomes from a "Ruleof Hand." The minimal complexityneeded to understanda panarchy and its adaptivecycles requires at least three to five key interactingcomponents, three qualitativelydifferent speeds, nonlinear causation.Vulnerability and resilience change with the slow variables;spatial contagion and biotic legacies generate self-organizedpatterns over scales in space and time. Emergentbehavior emerges from integrated systems. Linked ecological,economic, and social systems can behave differentlyfrom their parts. Integratedsystems exhibit emergent behavior if they have strong connectivity between the human and ecologicalcomponents and if they have key characteristicsof nonlinearityand complexity as suggested in the "Ruleof Hand." Managementmust take surpriseand unpredictability Managing complex systems requires confrontingmultiple into consideration. uncertainties.These can arise from technical considerations, such as models or analytic frameworks.The examples suggest that as much complexity exists in the social dimensions as in the ecological ones and that managers must juggle shifting objectives. Is adaptive management an answer? For linked ecological/social/economicsystems, slow variables, multistablebehaviors, and stochasticitycause active adaptive management to outperformoptimization approaches that seek stable targets.

Reprintedfrom Gunderson and Holling2001 withpermission of IslandPress .

404 C. S. Holling

* Encourage new foundations for renewal that build and sustain the capacity of people, econo- p political mies, and nature to deal with change. * Encourage programs to expand an understand- ing of change and communicate it to citizens, og intei rorga;,organizational businesses, and people at different levels of ad- ministration and governance, engaging them in the process of change. agerss ', actios interorganizational A principal conclusion from the Resilience Project is that the era of ecosystem management via incre- mental increases in efficiency is over. We are now individual in an era of transformation, in which ecosystem must build and maintain beginningof project end of project management ecological resilience as well as the social flexibility needed to Figure 9. Separate adaptive cycles are used to depict cope, innovate, and adapt. phases of issues as interpretedin four systems-political, organizational,interorganizational, and individual.Man- ACKNOWLEDGMENTS agers' actions and solutions must account for these dy- This paper draws extensively on Gunderson and namics of these from systems (Westley 2001). (Reprinted Holling, (2001), particularlythe three chapters I wrote Gunderson and Holling 2001 with permission of Island with my colleagues Buz Brock, Steve Carpenter, Press) Lance Gunderson, and Garry Peterson. Along with Carl Folke and Brian Walker, they served as the co- influences can generate unexpected new seeds of organizers of the Resilience Project, which provided opportunity that can nucleate and modify the sub- the crucible for these ideas, models, and examples. I sequent phase of growth. So long as connections thank these marvelous friends and collaborators for are maintained with other levels, those innovations their contributions, their inspiration, and their imag- are contained and do not propagate to other levels. ination. But all of the authors who contributed to the But if these recombinations and inventions accu- book were equally important to the development of mulate independently in a number of adjacent lev- these ideas. They comprised an international group of els, a time will come when the phases of several ecologists, economists, social scientists, and mathema- neighboring cycles become coincident, and each ticians whose depth of knowledge in their respective becomes poised as an accident waiting to happen in areas helped to produce a very real synthesis. In ad- a shift from fl to ca. Windows open that can then dition to those just mentioned above, they include allow those independent inventions and adapta- Nick Abel, Fikret Berkes, Pille Bunnell, Gilberto Gal- tions to interact, producing a cascade of novel self- lopin, Milena Holmgren, Marco Janssen, Don Lud- organized patterns across a panarchy and creating wig, Karl-Goran Maler, Charles Perrings, Rusty Prit- fundamental new opportunity. There is an "align- chard, Steve Sanderson, Marten Scheffer, Frances ment of the stars." Such a coincidence in phases of Westley, and Ralf Yorque. Over the 5 years of the vulnerability at multiple scales is quite rare. That is, project, we became the best of friends and collabora- true revolutionary transformations are rare, tors. Finally, I thank the MacArthur Foundation for whether in systems of people or systems in nature. the support of a grant and Dan Martin of the Under conditions of crisis in a region, the ele- MacArthur Foundation for his sustained advice and ments of a prescription for facilitating constructive help. are as follows: change REFERENCES * Identify and reduce destructive constraints and in- Adams RM. 1966. The evolution of urban Meso- hibitions on such as subsidies. society: early change, perverse potamia and PrehispanicMexico. Chicago:Aldine. * Protect and preserve the accumulated experi- Adams RM. 1978. Strategiesof maximization,stability and re- ence on which change will be based. silience in Mesopotamiansociety, agricultureand settlement. Proc Am Philos Soc 122:329-35. * Stimulate innovation and communicate the re- Allen TFH,Starr TB. 1982. for sults in a of fail-safe de- Hierarchy:perspectives ecological variety experiments complexity. Chicago:University of ChicagoPress. to directions in a that signed probe possible way AlvarezLW, Alvarez W, Asaro F, Michel HV. 1980. Extraterres- is low in costs in terms of human careers and trialcause for the Cretaceous-Tertiaryextinction. Science 208: organizational budgets. 1095-108. UnderstandingComplex Systems 405

Bak P. 1996. How nature works: the science of self-organized transformationsin human and natural systems. Washington criticality.New York:Copernicus Press, Springer-Verlag. (DC):Island Press. in press. BarroR. 1997. Determinantsof economic growth: a cross coun- Holling CS, GundersonLH, Peterson GD. 2001. Sustainability try empiricalcomparison. Cambridge (MA): MIT Press. and panarchies.In: GundersonL, Holling CS, editors. Panar- BaskervilleG. 1995. The forestryproblem: adaptive lurches of chy: understandingtransformations in human and natural renewal. In: GundersonLH, Holing CS, Light SS, editors. systems. Washington(DC): Island Press, in press Barriersand bridgesto the renewal of ecosystems and institu- Hurst DK. 1995. Crisis and renewal. Boston (MA): Harvard tions. New York:Columbia University Press. Business School Press. BerkesF. 1999. Sacredecology: traditional ecological knowledge Hurst DK, ZimmermanBJ. 1994. From life cycle to ecocycle. and managementsystems. Philadelphia and London:Taylor & Manage Inquiry 3:339-54. Frances. JablonskiD. 1995. Extinctionin the fossil record.In: LawtonJH, Berkes F, Folke C. 2001. Backto the future:ecosystem dynamics May RM, editors.Extinction rates. Oxford (UK):Oxford Uni- and local knowledge. In: GundersonL, Holling CS, editors. versity Press.pp. 25-44. Panarchy:understanding transformations in human and nat- Levin S. 1999. Fragiledominion. Reading (MA):Perseus Books. ural in systems. Washington(DC): Island Press. press Levin SA, Barrett S, Aniyar S, Baumol W, Bliss C, Bolin B, BrandS. 1999. The clockof the long now. New York:Basic Books. DasguptaP, EhrlichP, Folke C, Gren IM, and others. 1998. BrandS. 1994. How buildingslearn. New York:Viking Penguin. Resiliencein naturaland socioeconomicsystems. Environ Dev Econ 3:222-35. Brock W. 1972. On models of expectationsthat arise from max- imizing behaviorof economic agents over time. J Eco Theory Nystr6m M, Folke C. 2001. Spatial resilience of coral reefs. 5:348-76. Ecosystems.4:406-417. CarpenterSR, BrockWA, HansonP. 1999. Ecologicaland social O'Neill RV, DeAngelisDL, WaideJB, Allen TFH. 1986. A hier- dynamics in simple models of ecosystem management. Con- archicalconcept of ecosystems.Princeton (NJ): Princeton Uni- serv Ecol 3:4. [online] URL:http://www.consecol.org/vol3/ versity Press. iss2/art4 OstromE. 1992. Craftinginstitutions for self-governingirrigation CarpenterSR, Brock WA, Ludwig D. 2001. Collapse, learning systems.San Francisco:Institute for ContemporaryStudies Press. and renewal. In: GundersonL, HollingCS, editors. Panarchy: PritchardL Jr, Sanderson SE. 2001. The dynamics of political understandingtransformations in human and naturalsystems. discourse in seeking sustainability.In: GundersonL, Holling Washington(DC): Island Press. in press CS, editors. Panarchy:understanding transformations in hu- ClarkWC. 1985. Scalesof climateimpacts. Clim Change7:5-27. man and naturalsystems. Washington(DC): Island Press. Roe E. FischerDH. 1996. Greatwave: price revolutionsand the rhythm 1998. Takingcomplexity seriously: policy analysis,trian- and of history. New York:Oxford University Press. gulation sustainabledevelopment. Boston (MA):Kluwer AcademicPublishers. FlanneryT. 1994. The future eaters:an ecological history of the JA. 1950. Australasianlands and people. New York:Braziller. Schumpeter Capitalism,socialism and democracy. New York:Harper & Row. Folke C, BerkesF, ColdingJ. 1998. Ecologicalpractices and social Simon HA. 1974. The of HH mechanisms for building resilience and sustainability. In: organization complex systems. In: Pattee, editor. the of Berkes F, Folke C, editors. Cambridge(UK): CambridgeUni- Hierarchytheory: challenge complex New York:Braziller. versity Press.in press systems. pp. 3-27. Solow R. 1973. Is the end of the world at hand? Franklin JF, MacMahonJA. 2000. Ecology: messages from a Challenge mountain. Science 288:1183-5. March/April:39-50. Walker B, Abel N. 2001. Resilient in GadgilM, Berkes F, Folke C. 1993. Indigenous knowledge for rangelands:adaptation In: Gunderson editors.Pan- biodiversityconservation. Ambio 22:151-6. complex systems. L, Holling CS, archy: transformationsin human and natural GoldstoneJ. 1991. Revolutionand rebellionin the modern understanding early systems. Washington(DC): Island Press. in press world. Berkeley (CA):University of CaliforniaPress. Westley F. 2001. The devil in the dynamics:adaptive manage- GundersonL, CS, editors.2001. understand- Holling Panarchy: ment on the front lines. In: GundersonL, HollingCS, editors. ing transformationsin human and natural systems. Washing- Panarchy:understanding transformations in human and nat- ton (DC):Island Press. ural systems. Washington(DC): Island Press. in press GundersonLH, CS, SS. 1995. Barriersand Holling Light bridges Westley F. 1995. Governingdesign: the management of social to the renewal of and institutions. New York: ecosystems systems and ecosystems management. In: Gunderson LH, Columbia Press. University Holling CS, Light SS, editors. Barriersand bridgesto the re- Holling CS. 1986. Resilience of ecosystems; local surprise and newal of ecosystems and institutions. New York:Columbia global change. In: ClarkWC, Munn RE, editors. Sustainable UniversityPress. in press of the development biosphere. Cambridge(UK): Cambridge WestleyF, CarpenterSR, BrockWA, HollingCS, GundersonLH. Press. 292-317. in: Cleveland C, University Reprinted 2001. Why are systems of people and nature not just ecolog- Costanza R, editors. 1997. The of PerringsC, development ical or social systems? In: GundersonL, Holling CS, editors. economics. Brookfields:Edward ecological ElgarPublishing. Panarchy:understanding transformations in human and nat- Holling CS. 2000. Theoriesfor sustainablefutures. ConservEcol ural systems. Washington(DC): Island Press. in press 4:7. [online] URL:http://www.consecol.org/vol4/iss2/art7 WhitakerJK. 1987. Alfred Marshall.In: Eatwell J, Milgate M, HollingCS, GundersonLH. 2001. Resilienceand adaptivecycles. Newman P, editors. The new Palgrave:a dictionaryof eco- In: GundersonL, HollingCS, editors. Panarchy: understanding nomics. London:MacMillan Press.