Global Environmental Change 16 (2006) 293–303

www.elsevier.com/locate/gloenvcha

Linkages between vulnerability, resilience, and adaptive capacity

×
Gilberto C. Gallopı´n
United Nations Economic Commission for Latin America and the Caribbean (ECLAC), Casilla 179 D, Avda. Dag Hammarskjold s/n, Santiago, Chile
Received 28 November 2005; received in revised form 27 February 2006; accepted 28 February 2006

Abstract

This article uses a systemic perspective to identify and analyze the conceptual relations among vulnerability, resilience, and adaptive
capacity within socio-ecological systems (SES). Since different intellectual traditions use the terms in different, sometimes incompatible,
ways, they emerge as strongly related but unclear in the precise nature of their relationships. A set of diagnostic questions is proposed
regarding the specification of the terms to develop a shared conceptual framework for the natural and social dimensions of global change.
Also, development of a general theory of change in SESs is suggested as an important agenda item for research on global change.
Ⓒ 2006 Elsevier Ltd. All rights reserved.

Keywords: Vulnerability; Resilience; Adaptive capacity; Systems analysis; Stability

1. Introduction dynamics necessarily involves the consideration of both the

social and biophysical components and their mutual
The terms vulnerability, resilience, and adaptive capa- interactions.
city, are relevant in the biophysical realm as well as in the This article represents an attempt to highlight the
social realm. In addition to being terms in colloquial fundamental attributes of the three concepts, and to
language, they are widely used by the life sciences and identify the conceptual linkages between them, through
social sciences, not only with different foci but often with the use of a generic systems approach that can be specified
different meanings. The reviews presented in this issue for different concrete system types (social, ecological, but
(Adger, 2006; Folke, 2006; Smit and Wandel, 2006) and particularly socio-ecological). Without attempting a re-
other sources, document the diversity of interpretations view, an effort has been made to incorporate the
and reformulations of these concepts across disciplines and contributions from the principal scientific communities
problem areas as varied as evolutionary biology, ecology, that have been investigating these concepts, particularly
cultural studies, and computer science, to cite just a few. those that have more direct bearing on the analysis of their
Sometimes, the concepts are used interchangeably or as mutual conceptual relations. As the concepts have been
polar opposites. used in many fields, it will be helpful to define the focus of
This plurality of definitions is possibly functional to the the present analysis—the socio-ecological system (SES)—
needs of the different disciplinary fields, as well as being a and offer a justification of why this is important for global
reflection of the different intellectual traditions (Adger, change research.
2006; Janssen et al., 2006), but sometimes it may also The concept of vulnerability of the SES and its basic
become a hindrance to understanding and communication components are discussed in their general sense in the
across disciplines. This also may be the case in interna- following section. The section on resilience introduces the
tional research on global change, where understanding the concept of domains of attraction, essential in the notion of
resilience, and discusses different levels of stability
× relevant for the study of SESs. These have implications for
Tel: +56 2 210 2329; fax: +56 2 208 0484.
E-mail address: [email protected]. the applicability of the concept of resilience in the social

0959-3780/$ - see front matter Ⓒ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.gloenvcha.2006.02.004
29 G.C. Gallopı´n / Global Environmental Change 16 (2006) 293–
4 303
sciences. The links between resilience and vulnerability are Young et al. (2006) discuss a number of synthetic
highlighted. A section on adaptive capacity in its broader properties of globalization that require integrated treat-
and more specific forms as interpreted by different ment, such as connectedness, speed, scale, and diversity. At
scientific traditions follows, with a discussion of the another scale, this non-decomposability of many core
relations between adaptive capacity and the notions of issues of sustainable development is nicely illustrated by
capacity of response and resilience. very simple mathematical models of lake-and-managers
The final part of the article includes the outcome of the systems (Carpenter et al., 1999). The analysis of the
overall comparison and identification of linkages among behavior of these coupled models provides various insights
the three concepts, with an indication of the major of strategic importance for sustainable management of
uncertainties involved. Without affecting the current use shallow lakes. For example, unwanted collapse can occur
of vulnerability, resilience, and adaptive capacity within even if the ecosystem dynamics are perfectly known and
disciplinary areas, reaching some kind of agreement managers have perfect knowledge and control of human
between social and natural scientists working together on actions. It is also clear that these insights could not have
global change on the way the concepts are used would been obtained by analyzing the lake dynamics and the
represent an important step forward. To this end, a number societal dynamics separately. The analysis of the concepts
of diagnostic questions are proposed to help guide the discussed in this article will be made in the context of
choices to be made. The results of the analysis presented in research on the dynamics of the global SES.
this article naturally lead to a suggestion for the research
agenda on global change. 3. Vulnerability

2. The socio-ecological system Vulnerability is a concept that has been used in different
research traditions (Adger, 2006; Smit and Wandel, 2006)
Gallopı´n et al. (2001) have argued that the but there is no consensus on its meaning. Depending on the
natural analytical unit for sustainable development research research area, it has been applied exclusively to the societal
is the socio-ecological system or SES. An SES is defined subsystem, to the ecological, natural, or biophysical
as a system that includes societal (human) and ecological subsystem, or to the coupled SES, variously referred also
(biophysical) subsystems in mutual interaction (Gallopı as target system, unit exposed, or system of reference.
´n, 1991).1 The SES can be specified for any scale from Adger (2006) examines the evolution of approaches to
the local community and its surrounding environment to vulnerability originated in the social and the natural
the global system constituted by the whole of humankind sciences. He concludes that vulnerability is most often
(the ‘‘anthroposphere’’) and the ecosphere. Schellnhuber conceptualized as being constituted by components that
(1998) was the first to label the SES at the global scale as include exposure to perturbations or external stresses,
the ‘‘Earth System’’, and the term was later adopted by sensitivity to perturbation, and the capacity to adapt.
major international global environmental change research Vulnerability, like resilience, is generally viewed as
pro- grams, represented in the Earth System Science being specific to perturbations that impinge on the system;
Partnership (www.essp.org). in other words, a system can be vulnerable to certain
The need to investigate the whole SES arises from the disturbances and not to others. Two other widely accepted
increasingly recognized evidence that understanding and points are (1) the multiscale nature of the perturbations
anticipating the behavior of the social and ecological and their effects upon the system and (2) the fact that most
components of the SES in many cases requires simulta- SESs are usually exposed to multiple, interacting perturba-
neously taking into account both components; in other tions (van der Leeuw, 2001; Turner et al., 2003). Vulner-
words, SESs are non-decomposable systems. Of course it is ability is also thought of as a susceptibility to harm, a
always possible to single out certain social or ecological potential for a change or transformation of the system
components for study, and this strategy has provided when confronted with a perturbation, rather than as the
important understanding of the components, as has been outcome of this confrontation. However, diverse views
traditionally done with great success by social and natural regarding the precise meaning of vulnerability are also
scientists. However, there are important traits related to evident. Some of the differences are important for the task
the behavior and future trajectory of the Earth System, for of identifying the relationships between vulnerability,
example, that cannot be understood through this analytical resilience, and adaptive capacity, and will be discussed in
approach alone, because they emerge from the dynamic what follows.
interplay between the social and ecological components.
Many of the issues related to vulnerability, resilience, and 3.1. Perturbation, stress, hazard, or shock
adaptive capacity fall in this category (Walker et al., 2004;
Turner et al., 2003). For Turner et al. (2003), hazards are threats to a
system, comprised of perturbations and stress (and
1
Also called social–ecological system (Berkes and Folke, 1998) and stressors, the sources of stress). Perturbations are major
coupled human-environmental system (Turner et al., 2003).
spikes in pressure (e.g., a tidal wave or hurricane)
beyond the
 G.C. Gallopı´n / Global Environmental Change 16 (2006) 293– 2
normal range of variability in which the system operates, has been noted in relation to resilience, in the sense that
and commonly originate beyond the system or location ‘‘resilience is not always a good thing’’ (Walker et al.,
in question. Stress is a continuous or slowly increasing 2004).
pressure (e.g., soil degradation), commonly within the Another aspect of the notion of change or transforma-
range of normal variability. Stress often originates tion is its degree or depth. A system would not be called
within the system, and stressors often reside within it. vulnerable if the effect of the perturbation is limited to the
For simplicity, the term perturbation will be used in this generation of trivial and ephemeral changes; those changes
article to denote the external or internal processes would hardly qualify as a ‘‘transformation’’ of or
interacting with the system and with the potentiality of ‘‘damage’’ to the system. However, it is not always clear
inducing a significant transformation in the system, be it what is meant by the term. Changes in the system could
slow or sudden. range from variations in the behavior of some variables of
Perturbations are usually assumed to come from outside the system up to radical changes in the structure of the
the system. But this may be an unduly restrictive definition. system (as discussed later under resilience).
Both societal and ecological systems survive, thanks to a This is an important point, because, for some, the
constant exchange of matter, energy, and information with fundamental distinction between vulnerability and resi-
their external environment. Those processes can give rise lience lies in that vulnerability refers to the capacity to
to modifications in the functioning or structure of the preserve the structure of the system while resilience refers
system triggered by changes in the system’s environment to its capacity to recover from non-structural changes in
(e.g., the effects of an earthquake on a population, dynamics (van der Leeuw, 2001). Sometimes, the concept
dissemination of infectious germs in the environment of a of vulnerability is restricted to situations when the
person), by internal alterations (e.g., the impact of civil war system suffers structural change (Gallopı´n et al.,
on a country, the manifestation of a genetic disease in a 1989; van der Leeuw, 2001; Young et al., 2006).
person), or by the interaction among external and internal Unfortunately, in practice, it is not always easy to decide
processes (e.g., the effects of a prolonged drought in a when an observed change is behavioral or structural
country with internal conflicts). Young (2005), discussing without investigating the system thoroughly. The previous
institutional dynamics in environmental and resource discussion highlights the importance of specifying what is
regimes, illus- trates how the sources of stress may be meant by ‘‘transforma- tion’’, impact, or ‘‘harm’’ when
internal, external, or both. Turner et al. (2003) explicitly discussing (or defining) vulnerability in the natural and
state that the hazards acting on the system arise from social realms of the SES.
influences outside and inside the system, and Kasperson et
al. (2005) allocate a section to endogenous perturbations. 3.3. Sensitivity
In another field, Nicolis and Prigogine (1977)
demonstrated for dissipative (open, far from equilibrium) The concept of sensitivity varies across authors; for
systems the phenomenon of self-amplification of internal instance, Adger (2006) defines it as ‘‘the extent to which a
fluctuations (perturbations) and their ultimate human or natural system can absorb impacts without
breakthrough at the system level. suffering long-term harm or other significant state
On the other hand, whether the disturbance is described change’’; Smit and Wandel (2006) talk about exposure-
as external or internal depends also on the scale at which sensitivity, arguing that sensitivity is not separable from
the system has been defined. Earthquakes, hurricanes, or exposure. Luers (2005) also combines sensitivity and
global economic crises are clearly internal phenomena for exposure, and defines sensitivity as the degree to which a
the global SES, but they are obviously external events for system will respond to an external disturbance, and also
systems such as a Central American village. includes in the concept the ability to resist change and the
ability to return to a previous condition after the stress has
3.2. Change or transformation of the system been removed—properties that are usually seen as asso-
ciated with resilience or with coping capacity. In discus-
In general, transformation is taken to mean harm or sions of climate (IPCC, 2001), sensitivity is the degree to
damage to a system (human, natural, SES). However, a which a system is affected, either adversely or beneficially,
different interpretation has been proposed on the basis of a by climate-related stimuli. The effect may be direct (e.g., a
systemic analysis of the concept of vulnerability (Gallopı change in crop yield in response to a change in the mean,
´n, 2003). According to this conception, vulnerability range, or variability of temperature) or indirect (e.g.,
is not always a negative property. It is possible to speak damages caused by an increase in the frequency of coastal
of positive vulnerability in cases where change leads to a flooding due to sea-level rise).
beneficial transformation such as the emergence of a given For Gallopı´n (2003), in its general sense,
social group from chronic poverty or the collapse of an sensitivity is the degree to which the system is modified
oppressive regime. Young (2005) discusses situations or affected by an internal or external disturbance or set of
where institutional crises become windows of distur- bances. Conceptually, it can be measured as the
opportunity for improvement even though they are amount of transformation of the system per unit of
dangerous events that can produce destructive outcomes. change in the disturbance, i.e., @ transformation=@
A similar ambivalence perturbation
29 G.C. Gallopı´n / Global Environmental Change 16 (2006) 293–
6 303
(Tomovic, 1963),2 but in the simplest case it only specifies
whether or not the system is sensitive to a given factor
(Gallopı´n, 2003). In this view, sensitivity is an
inherent property of an SES, distinguished from its
capacity of response (the actual transformation may be
smaller, depending on the capacity of response of the
system). It is an attribute of the system, existing prior to
the perturbation, and separate from exposure.

3.4. Capacity of response

The system’s coping capacity (Turner et al., 2003), or

capacity of response (Gallopı´n, 2003), is also
called adaptive capacity by Adger (2006) and Smit and
Wandel (2006) and the IPCC (2001). Turner et al. (2003)
distinguish capacity to cope or respond from adaptive
capacity, and consider both as components of the
resilience of a system. They refer to adaptations as the
system’s restructuring after the responses. As noted by
Smit and Wandel (2006), some authors apply ‘‘coping
ability’’ to shorter-term capacity or the ability to just
survive, and employ ‘‘adaptive capacity’’ for longer-term
or more sustainable adjustments. In view of this lack of
agreement, the term used here for this component of
vulnerability will be ‘‘capacity of response’’ (of the
system to the perturbations).
Fig. 1. Generic relations between vulnerability, threat, exposure, and
In general, capacity of response is the system’s ability to impact or transformation of a system (social, natural, or SES)
adjust to a disturbance, moderate potential damage, take according to Gallopı´n (2003). Time moves from top to bottom of
advantage of opportunities, and cope with the conse- the figure. The target system is represented by an oval; its exchanges
quences of a transformation that occurs. Capacity of with its external environment are represented by arrows in both
response is clearly an attribute of the system that exists directions, and the normal operation of its internal processes is
symbolized by a regular spiral. The components of the vulnerability of
prior to the perturbation. the system (its sensitivity and capacity of response) are highlighted with
boxes. The considered (external) process or perturbation is represented
3.5. Exposure by the looped shape at the right, with its relevant attributes in boxes.
The exposure of the system to the perturbation is represented by the
overlap between the two elements, and the transformed or impacted
The other central concept related to vulnerability is system is represented by the wobbly oval at the bottom.
exposure, meaning in general the degree, duration, and/or
extent in which the system is in contact with, or subject to,
the perturbation (Adger, 2006; Kasperson et al., 2005). From this perspective, a system (i.e., a city, a human
Exposure in most formulations is seen as one of the community, an ecosystem) may be very vulnerable to a
elements constituting vulnerability. However, Bohle certain perturbation, but persist without problems insofar
(2001), building on proposals by Robert Chambers, as it is not exposed to it. A person with low immunological
recognizes a qualitative difference between exposure defenses would be called vulnerable to infectious diseases,
(defined as the external side of vulnerability) and coping whether or not he or she is exposed to the infectious agent;
(the internal side). Since, unlike sensitivity and capacity software vulnerabilities existed before the Internet exposed
of response, exposure seems to be an attribute of the every computer in the world to every hacker on the planet.
relationship between the system and the perturbation, From the perspective that includes exposure as a compo-
rather than of the system itself, Gallopı´n (2003) did not nent of vulnerability, a system that is not exposed to a
consider exposure as a component of vulnerability. perturbation would be defined as non-vulnerable. The
Rather, vulnerability is a function of the system’s hypothetical person with low defenses would not be called
sensitivity and capacity of response, and the vulnerable to infectious diseases if confined to a sterile
transformation suffered by the system is a function of its environment.
vulnerability, the properties of the perturbation, and the The choice of including or not including exposure as a
exposure of the system to the perturbation (Fig. 1). component of vulnerability has consequences. In the first
2
case, vulnerability becomes a property of the relationship
Of course, in many cases, the transformation is discontinuous, and
between the system and its environment (specifically
therefore the expression above (strictly applicable for continuous change)
should be replaced by a more appropriate one. between the system and the perturbation), rather than a
 G.C. Gallopı´n / Global Environmental Change 16 (2006) 293– 2
property of the target system. The concrete characteriza- The concept of resilience has a rich history (see Folke,
tion of vulnerability (i.e., indices, maps, etc.) needs to take 2006), sometimes with a considerable stretch from its
into account the full set of possible combinations of original meaning. There is also a body of thought around
situations, and must be changed if the distribution of resilience and the ‘‘adaptive cycle’’—a metaphor for the
exposure changes (for instance, when alternative climate dynamics of ecosystems later extended to the social aspects
scenarios are examined). and SESs—referring to the concept of ‘‘panarchy’’ or
On the other hand, if exposure is externalized from cross-scale dynamics and interplay between nested
vulnerability, exposure is a relational property, and adaptive cycles (Gunderson and Holling, 2002; see also
vulnerability is a property of the system, becoming Folke, 2006). These ideas, interesting as they are, fall
expressed/revealed when the system is exposed to the outside the scope of this article.
perturbation. Vulnerability in this view is a system The concept of domains or basins of attraction is central
attribute existing prior to the disturbance, although it is to the notion of resilience. It essentially means the
often related to the history of disturbances to which the portion of the state space of a dynamic system that
system was exposed in the past (hence the importance of contains one ‘‘attractor’’ toward which the state of the
the system’s history). If the vulnerability of different target system tends to go, and is therefore one region of the
systems could be characterized on the basis of their state space where the system would tend to remain in the
sensitivity and capacity of response, their exposure to a absence of strong perturbations. ‘‘State of a system’’ means
particular perturbation could be independently deter- ‘‘any well-defined condition that can be recognized if it
mined, and the harm or outcome of the resulting occurs again’’ (Ashby, 1956) or, in more precise terms,
transformation of the system could be estimated from the the set of values adopted by all the variables of the
composition of the two factors. Using a climate change system at a given time.
example, only one map of vulnerability would be needed, If the state of the system changes in time, then the
that could be overlaid with different maps of exposure succession of states through time can be interpreted as
resulting from diverse models or scenarios. Therefore, the defining a trajectory (also called an orbit) of the system,
difference between the two perspectives is not trivial, going from some initial state to the current state to some
reflecting on the possibilities for generalization and also the future state and (possibly, but not necessarily) to a final
design of policies to reduce vulnerability. state. This ‘‘trajectory’’ unfolds in an abstract state space
The differences between sensitivity, response capacity, defined by the number of variables of the system. Each
and exposure can be illustrated with two very simplified state of the system is represented by a point in the state
examples. The first refers to the effects of a flood on a space. Fig. 2 illustrates a trajectory for a system with only
community. The most precarious homes are hit harder by a two state variables, X1 and X2, that define the state space.
flood than the solid ones (sensitivity). Oftentimes, the The trajectory tends to move in time toward, or remain
poorest homes are located in the places most susceptible to in, an attractor of the system. This could be a point or
flooding (exposure). The families with the greatest steady state, a stable closed orbit (a limit cycle), a
resources have a greater availability of means to repair ‘‘strange’’ (chaotic) attractor, an open-ended trajectory
water damage (response capacity). The magnitude of the that never reaches a steady state, or more complicated
final impact will also depend on the intensity, magnitude, geometries when the system has many state variables. In
and permanence of the flood (attributes of the perturba- general terms, an attractor characterizes what the behavior
tion). The second example refers to the impacts of the of a system settles down to. If the attractor is a fixed point,
spread of an infectious disease in the population of a the system will tend to reach that state (steady state
region. The population segment constituted by children or dynamic equilibrium) where it will tend to remain
and the elderly is likely to exhibit more serious symptoms
if infected than the rest of the population (sensitivity). The
high-income sector often has better access to medical care
and medicines (capacity of response). In this particular
example, the likelihood of entering in contact with the
germs (exposure) might be evenly distributed across the current state

population.

4. Resilience
trajectory
Resilience, a concept originated within ecology, is also initial state
applicable in the realm of social systems and SESs. For
instance, Adger (2000) defines social resilience as the
ability of groups or communities to cope with external x2
stresses and disturbances as a result of social, political, and
Fig. 2. A trajectory of a two-variable, dynamic system depicted in the
environmental change. state space (here, a plane defined by all possible values of state variables
X1 and X2).
29 G.C. Gallopı´n / Global Environmental Change 16 (2006) 293–
8 303
thereafter. Given that all real systems are permanently
exposed to perturbations, a state often will be pushed away
from the steady state, but will tend to return to it. When one large attractor one

the attractor is a trajectory, the state of the system will tend large attractor
to move toward it when pushed away by perturbations, but bifurcation
two attractors

paramete
will never reach constancy; it will change periodically (if
bifurcation
the attractor is a limit cycle) or along an open path in the
state space; in these cases, nearby states will tend to go
toward the trajectory, but not to a fixed point. one attractor

If a system has only one attractor, in due course its state

will end there; such a system is called globally stable.
VIEW FROM ABOVE:
However, systems containing non-linear relations between
their variables (as is the case for all SESs) usually possess
more than one attractor, and therefore surprises can occur
as the state of the system shifts from the domain of one large attractor
influence of one attractor (known as the domain or basin of
attraction) to that of another. Fig. 3 shows a system with
three basins, one containing a steady state (attractor A),

paramete
the other a stable cycle (attractor B), and the third
containing a stable trajectory (attractor C). The complete two attractors
figure depicts the ‘‘stability landscape’’ of the system,
represented by the configuration of all basins of attraction,
including the boundaries separating them. The stability
landscape is part of the structure of the system, depending
on the values of the parameters (fixed or very slowly one attractor
varying factors) of the system. In a dynamical system with
multiple attractors, a continuous variation in some critical
parameter can result in discontinuous changes in the Fig. 4. Qualitative changes in attractors. Continuous variation of a
stability landscape of the system (Fig. 4). These disconti- parameter can cause attractors to shrink, split, or disappear. The upper
nuities are called bifurcations in the mathematical theory part of the figure is a three-dimensional representation; the lower part is a
view from above, at three sections of the solid shape.
of dynamical systems (Tu, 1994; Butenin, 1965) and
catastrophes in catastrophe theory (Thom, 1972).
Holling (1973) introduced a new, non-equilibrium vision
in ecology with the concept of ecological resilience, arising
from the analysis of different empirical studies, mathema- them are multistable; that is, they have two or more
tical models, and experience with managed ecosystems. He domains of attraction where the system variables tend to
showed that even natural, undisturbed ecological systems remain. Within each domain, the system’s state may
are often in transient states and demonstrated that many of fluctuate widely (i.e., may be highly unstable) but if it
tends to stay within the boundaries of the domain, the
system is resilient. Resilience was originally defined by
Holling (1973) as ‘‘a measure of the persistence of systems
and of their ability to absorb change and disturbance and
still maintain the same relationships between populations
or state variables’’ (p. 14) and, more recently, by Walker
et al. (2004) as ‘‘the capacity of a system to absorb
disturbance and reorganize while undergoing change so as
to still retain essentially the same function, structure,
identity, and feedbacks—in other words, stay in the same
basin of attraction’’. Folke (2006) reviews the concept in
detail.
Note that resilience, in terms of the stability
landscape, implies the ability of a multistable system to
keep the values of its state variables within a given
domain of attraction in the face of perturbations, and is not
concerned with the stability or constancy of the state within
Fig. 3. State space of a two-variable system with three attractors, the basin. This concept is called ecological resilience; in
indicating the respective basis of attraction with dotted lines. principle, ecological resilience can be measured by the
magnitude of
 G.C. Gallopı´n / Global Environmental Change 16 (2006) 293– 2
the perturbation that can be absorbed before the state of possibility of a true transformation of the original system
the system falls outside its domain of attraction.3 into a different one.4
By contrast, the concept of stability as commonly Among the various theoretical approaches to real-life
utilized focuses on the behavior of the system near an complex systems, the one derived from the theory of
equilibrium point or trajectory, and can be measured by dissipative structures (developed essentially by Ilya Prigo-
the speed at which the system returns to the stable point or gine and collaborators) seems particularly suitable for
trajectory following a perturbation. This is essentially the investigating the dynamics of structural change and
concept of resilience defined by Pimm (1984) and called persistence in SESs. The theory of dissipative structures
engineering resilience by Holling (1996). It is equivalent to shows that open, self-organizing systems maintain their
the well-known mathematical concept of local stability structural order by keeping their internal state far from
(Hahn, 1967; Tu, 1994; Nicolis and Prigogine, 1989). thermodynamic equilibrium, through active exchanges
In any dynamical system, the property of multistability with their environment. Those dissipative structures are
implies that the behavior of the system may change in principle stable as long as the exchanges with the
qualitatively in a way that is surprising to an external environment are maintained and as long as the continu-
observer; for instance, a cumulative series of small ously occurring fluctuations (or perturbations) are ab-
perturbations of the state of the system, each individually sorbed within the framework of the given dynamic regime.
reversible, may finally move the state of the system over However, any structure of a non-equilibrium system may
the boundary of the current basin of attraction, thus flipping be driven beyond a threshold into a new regime when the
it into another (possibly undesirable) domain where it will fluctuations exceed a critical size. At that point, after going
tend to remain. This is precisely the behavior exhibited by through phases of instability and high entropy, the system
many ecosystems under management (Holling, 1986; may evolve to a different stable regime with a new
Gunderson and Holling, 2002). characteristic structure (Nicolis and Prigogine, 1977,
Those sudden shifts in behavior occur in the absence of 1989; Prigogine and Stengers, 1979).
structural change in the system. Holling (1986) also As noted by Adger (2000), the concept of resilience
showed that in a number of cases, the size and shape (and cannot be transferred uncritically from the ecological
the genesis or disappearance) of the domains of attraction sciences to social systems. But using the concept for social
can change because of the unperceived evolution of systems (and for SESs) does not really imply that there are
parameters of the system (implicitly assumed constant), no essential differences in behavior and structure between
which are often affected by long-term management or are social and ecological systems. The legitimate use of the
internally determined by processes that link variables. concept only requires the assumption that the state space of
Thus, the stability domains themselves may expand, the system considered contains more than one basin of
contract, and disappear in response to changes in slow attraction. This is a natural assumption for all kinds
variables, resulting in the loss of resilience of the system of non-linear dynamical systems (although the applicability
(similar to the changes depicted in Fig. 4). Resilience can of the concept of dynamical systems to social systems
operate at different scales, and it has been noted that there might not be acceptable to some social scientists).
can be losses of resilience at some scales thereby When the concept of resilience is unlinked from the
increasing it at other, higher scales (Walker et al., 2004). notion of multistability, it becomes very difficult to
In this context, three types or levels of stability may be distinguish it from structural stability, or even from local
distinguished. The first, local stability or engineering stability or adaptive capacity. This is one risk with some of
resilience, refers to the behavior of the trajectories of the the recent reformulations of resilience, such as making
system in the neighborhood of an attractor, within a given adaptive capacity and self-organization properties of
domain of attraction. The second level refers to changes in resilience (Carpenter et al., 2001). It is sometimes said that
the state of the system between the different domains of vulnerability is the flip side, or the antonym, of resilience
attraction, within the stability landscape of the system. The (Folke et al., 2002). However, this is by no means clear;
capacity of the system to remain within the same domain of obviously a resilient system is less vulnerable than a non-
attraction is called ecological resilience. The third level resilient one, but this relation does not necessarily imply
includes changes in the stability landscape itself. This is the symmetry.
domain of structural stability of dynamical systems, the Resilience is clearly related to the capacity of response
capacity of the system to preserve the topology of its component of vulnerability, and thus it would be less than
trajectories (the qualitative features of its stability land- the flip side of vulnerability. A more fundamental
scape) under perturbations of the dynamic equations of the difference is that resilience, as discussed earlier, applies to
system (Tu, 1994). Structural instability represents the the preservation of the behavior of the system as expressed

4
Note that the new system may result from adding new components and
3
Recently, Walker et al. (2004) proposed latitude, resistance, precar- relations, but also from losing or modifying components or relations. In
iousness, and panarchy as essential attributes of resilience; other resilience this sense, structural instability is somewhat more general than the concept
measures could be derived from these. of transformability as recently defined by Walker et al (2004).
30 G.C. Gallopı´n / Global Environmental Change 16 (2006) 293–
0 303
by its state remaining within the considered domain of latter. This is more so with human systems, capable of
attraction, while vulnerability refers to transformations learning and technological progress. Of course, in the
that may go beyond a single domain. The flip side of human realm, and thus also in the SES, the criterion
vulnerability would be a concept that denotes capacity to for adaptness goes far beyond ‘‘being able to live and
maintain the structure of the system against reproduce’’; it includes the viability of social and eco-
perturbations, even if its resilience is overcome; nomic activities, and the quality of human life. Adapt-
robustness is a good candidate.5 ability or adaptive capacity of human systems also can be
Moreover, resilience seems to be a proper subset of defined as the capacity of any human system from the
capacity of response, at least for the social component of individual to humankind to increase (or at least maintain)
the SES. Capacity of response includes, for most the quality of life of its individual members in a given
authors, not only the resilience of the system (maintenance environment or range of environments (Gallopı´n et
within a basin) but also coping with the impacts al., 1989). As noted by Smithers and Smit (1997), while the
produced and taking advantage of opportunities. The responses of biological systems to perturbations are purely
relation of resi- lience to the sensitivity component of reactive, the responses of human systems are both reactive
vulnerability is also unclear. A sensitive system may or and proactive.
may not be resilient. An insensitive system (i.e., an From these considerations, a generic concept of adaptive
‘‘armored system’’) may exhibit low vulnerability and capacity of an SES would seem to involve two different
low resilience (it is the exposure to perturbation that components, namely (1) the capacity of the SES to cope
builds resilience in natural systems). Sensitivity may with environmental contingencies (to be able to maintain
open a system to threats, but an insensitive system may be or even improve its condition in the face of changes in its
unable to adapt and seize opportunity. The concept of environment(s)) and (2) the capacity to improve its
resilience does not include exposure (similar to condition in relation to its environment(s), even if the
vulnerability as adopted here) but refers to the reaction latter does not change, or to extend the range of
of the system when exposed to perturbations. On the environments to which it is adapted.
other hand, a history of past exposures may be important In more recent usage in the field of climate change,
to build resilience (Holling, 1973, 1985, 1986). adaptive capacity is defined as ‘‘the ability of a system to
adjust to climate change (including climate variability and
5. Adaptive capacity extremes) to moderate potential damages, to take advan-
tage of opportunities, or to cope with the consequences’’
The concept of adaptive capacity has been reviewed by and adaptation is defined as an ‘‘adjustment in natural or
Smit and Wandel (2006). Here, only the fundamental traits human systems in response to actual or expected climatic
of the concept will be explored. Adaptability (or adaptive stimuli or their effects, which moderates, harms, or exploits
capacity)6 was originally defined in biology to mean an beneficial opportunities,’’ including anticipatory and reac-
ability to become adapted (i.e., to be able to live and to tive, autonomous or spontaneous and planned, and public
reproduce) to a certain range of environmental contingen- and private (IPCC, 2001). However, Kasperson et al.
cies. Adaptness is the status of being adapted, and an (2005) distinguish between adjustments and adaptations.
adaptive trait or an ‘‘adaptation’’ is a feature of structure, For them, adjustments are system responses to perturba-
function, or behavior of the organism that is instrumental tions or stress that do not fundamentally alter the system
in securing the adaptness (Dobzhansky, 1968). Adaptness itself, they are commonly (but not necessarily) short-term
is not a generic property, but it refers to a certain and involve relatively minor system modifications. Adap-
environment or range of environments, and different tations are system responses to perturbations or stress that
organisms, different populations, or different species are are sufficiently fundamental to alter the system itself,
adapted to different environments. Dobzhansky noted that sometimes shifting the system to a new state.
high adaptness is not the same as high adaptability; a The IPCC further distinguishes between adaptation as
species may be highly adapted to a special and constant actions that operate upon the system itself and mitigation,
environment but have little capacity to adapt to others or or actions that operate upon the origin and attributes of
to changes in its environment. the perturbation (i.e., reducing greenhouse gases emis-
In general, a species, population, or individual may sions). This concept of adaptive capacity is clearly limited
also become better adapted by improving its condition to coping with changes in the environment of the system
in its environment, even in the absence of changes in the (actually to climate change) and seems to exclude the
element of increasing adaptness when the environment
5
The concept of robustness has not been analyzed here, as it is still much does not change. One question that arises is whether the
debated. Robustness here is used simply as the converse of vulnerability
concept of adaptive capacity of the SES should include this
and basically to denote the ability of the system to preserve its structure in
the face of perturbations (Tu (1994, p. 160) uses robust as synonymous of second element. It would seem that, at least with reference
structurally stable). Many other definitions of robustness have been to the human component of the coupled system, it might be
proposed (see Anderies et al., 2004; Jen, 2003). of interest not to reduce a priori the general concept of
6
Adaptability and adaptive capacity are usually treated as synonymous adaptive capacity.
(i.e., Smit and Wandel, 2006; IPCC, 2001).
 G.C. Gallopı´n / Global Environmental Change 16 (2006) 293– 3
Another issue is the relationship between adaptive or capacity of response. But again it is unclear whether
capacity of an SES and capacity of response, as used here resilience includes capacity of response, or is an element of
for the component of vulnerability. The IPCC treats both the latter. Given that capacity of response, as an element of
concepts as synonymous, but, as mentioned in the vulnerability, is supposed to refer to the response of the
preceding paragraph, it uses a definition of adaptive system to structural changes, it would appear that
capacity that may be too restrictive for use in broader resilience should be considered as a subset, or a compo-
problems of the dynamics of the global SES. In general nent, of capacity of response.
terms, adaptive capacity would seem to be broader than The conceptual links between adaptive capacity as an
capacity of response; specific adaptations may include attribute of an SES and capacity of response as a
modifying the sensitivity of the system to perturbations, component of its vulnerability are not clear beyond the
increasing its resilience (indeed, Walker et al. (2004) define confirmation of the existence of the relationship. If
adaptability as the capacity of humans to manage resilience adaptive capacity in general is considered to include also
in the SES), and reducing the exposure of the system to improvements in the adjustments of the system to its
perturbations. However, this will depend on the concrete environment even in the absence of changes in the latter,
definitions for adaptive capacity and capacity of response then it is clearly more general than capacity of response.
adopted in the investigation of a coupled SES. Another point is that adaptive capacity can include
The relations between the concepts of adaptive capacity reactions of the system that modify its sensitivity to
and resilience are more unclear, because of the diversity of perturbations, and its exposure to them. As described in
views. As mentioned by Smit and Wandel (2006), some the vulnerability section, capacity of response also has been
authors equate adaptive capacity with resilience and social distinguished from adaptive capacity using the criteria of
resilience. Gunderson (2000) defines adaptive capacity as short- or long-term adjustments, or of their timing, but in
system robustness to changes in resilience, Carpenter et al. this case both terms have been viewed as belonging to
(2001) use adaptive capacity as a component of resilience vulnerability.
that reflects the learning aspect of system behavior in Fig. 5 summarizes the conclusions on the major
response to disturbance, and Walker et al. (2004) define conceptual relations among the three concepts. The
adaptability as the collective capacity of the human actors comparative analysis of the concepts of vulnerability,
in an SES to manage resilience, including making desirable resilience, and adaptive capacity puts in evidence important
basins of attraction wider and/or deeper, and shrinking similarities and differences, and in some cases contra-
undesirable basins; creating new desirable basins, or dictions, between the concepts as specified, or utilized, in
eliminating undesirable ones; and changing the current different fields of inquiry. The comparison also shows that
state of the system so as to move either deeper into a there is no generally accepted meaning for these concepts.
desirable basin, or closer to the edge of an undesirable one. The lack of general agreement on the concepts when
considered one by one becomes more visible when they are
6. Conclusions taken together.

The analysis of the concepts of vulnerability, resilience,

and adaptive capacity from a systemic perspective in the
context of research on the dynamics of the global SES
shows that these concepts are related in non-trivial ways. If
care is not used, the field of human dimensions research
can become epistemologically very messy.
For instance, it seems natural to view vulnerability and
resilience as related properties of an SES. But the specific
nature of the relation is not obvious. The views expressed
in the literature range from considering vulnerability as the
flip side of resilience to have resilience as one of the
components of vulnerability. However, vulnerability does
not appear to be the opposite of resilience, because the
latter is defined in terms of state shifts between domains of
attraction, while vulnerability refers to (or at least also
refers to) structural changes in the system, implying
changes in its stability landscape. Moreover, resilience is
an internal property of the system, not including exposure
to perturbations. Fig. 5. A diagrammatic summary of the conceptual relations among
Resilience would appear to be more obviously related to vulnerability, resilience, and adaptive capacity as described in this article.
The signs represent relationships between sets: C ¼ ‘‘subset of’’;
one of the components of vulnerability, the same that is
g ¼ ‘‘not a subset of’’; R, V, AC, and CR stand for resilience,
variously called adaptive capacity, coping capacity, coping, vulnerability, adaptive capacity, and capacity of response, respectively.
30 G.C. Gallopı´n / Global Environmental Change 16 (2006) 293–
2 303
Interdisciplinary research on the Earth System and SESs and it could be an appropriate item for a research agenda
at other scales would clearly benefit from having a general, on global change.
self-consistent set of these basic concepts that could be
applied across disciplines. Therefore, there is a need to
Acknowledgments
develop clear (and, hopefully, shared) specifications of the
concepts in the abstract, ecological, and social senses, that
are mutually compatible; this can be critical for the I am grateful for the comments received from three
interactions between social and natural sciences in the anonymous reviewers, and to Bill Turner for providing
study of the Earth System and coupled SESs at other specific information. I also thank Elinor Ostrom and
scales. Marco Janssen, and the scientific committee of the
Some of the questions that ought to be considered in the IHDP, for their invitation to write this article, and the
process have been identified as: participants of the various IHDP workshops for useful
comments. The writing of this article was greatly
facilitated by the generous access to the in-depth reviews
●What is meant by ‘‘harm’’ or ‘‘transformation’’ (struc- prepared for this issue of the Journal by Adger, Folke, and
tural change, shifting domains of attraction, moving Smit and Wandel. The manuscript was improved by the
away from equilibrium states or trajectories)? editorial suggestions provided by Joanna Broderick.
●Is positive vulnerability a useful notion?
●Does vulnerability apply to internal perturbations? References
●Is vulnerability a property of the system or of the
relationship between the system and the perturbation? Adger, W.N., 2000. Social and ecological resilience: are they related?
●Is negative (perverse) resilience a suitable concept? Progress in Human Geography 24 (3), 347–364.
●Does adaptation include improvements of the system in Adger, W.N., 2006. Vulnerability. Global Environmental Change 16 (3),
a non-changing environment? 268–281.
● Is adaptive capacity the same as capacity of response? Anderies, J.M., Janssen, M.A., Ostrom, E., 2004. A framework to analyze
the robustness of social-ecological systems from an institutional
● Is resilience the same as adaptive capacity? perspective. Ecology and Society 9 (1) art. 18 [online], URL: http://
www.ecologyandsociety.org/vol9/iss1/art18.
In this article, answers have been proposed for each of Ashby, W.R., 1956. An Introduction to Cybernetics. Chapman and
Hall, London.
the questions, but what is essential for good research on the Berkes, F., Folke, C. (Eds.), 1998. Linking Social and Ecological
global SES is that these definitions are not only scientifi- Systems: Management Practices and Social Mechanisms for Building
cally and epistemologically valid, but also that they are Resilience. Cambridge University Press, Cambridge.
shared by the research communities in the social and Bohle, H.-G., 2001. Vulnerability and criticality: perspectives from social
natural sciences cooperating on the study of the dynamics geography. IHDP Update 2/01, art. 1 [online]. URL: http://www.ihdp.
uni-bonn.de/html/publications/update/IHDPUpdate01_02.html.
of global change. It is hoped that the present analysis Butenin, N.V., 1965. Elements of the Theory of Nonlinear Oscillations.
contributes to that. Blaisdell, New York.
Ultimately, vulnerability, resilience, and adaptive capa- Carpenter, S., Brock, W., Hanson, P., 1999. Ecological and social
city (and robustness) are different manifestations of more dynamics in simple models of ecosystem management. Conservation
general processes of response to changes in the relationship Ecology 3 (2) art. 4 [online], URL: http://www.consecol.org/vol3/iss2/
art4.
between open dynamical systems and their external Carpenter, S.R., Walker, B.H., Anderies, J.M., Abel, N., 2001. From
environment. This suggests that an interesting and useful metaphor to measurement: resilience of what to what? Ecosystems 4,
line of research could be represented by the investigation of 765–781.
the general dynamics of change in SES.7 Dobzhansky, T., 1968. Adaptness and fitness. In: Lewontin, R.C. (Ed.),
A general theory of change and transformation of SESs Population Biology and Evolution. Syracuse Univ. Press, Syracuse,
New York, pp. 109–121.
would involve the relevant internal dynamics of the SES,
Folke, C., 2006. Resilience: the emergence of a perspective for social-
including aspects such as local stability, resilience, struc- ecological systems analyses. Global Environmental Change 16 (3),
tural stability, and self-organization sensu Prigogine 253–267.
(Nicolis and Prigogine, 1977, 1989), the various forms of Folke, C., Carpenter, S., Elmqvist, T., Gunderson, L., Holling, C.S.,
interaction of the system with its environment (including Walker, B., Bengtsson, J., Berkes, F., Colding, J., Danell, K.,
Falkenmark, M., Gordon, L., Kaspersson, R., Kautsky, N., Kinzig,
both threats and opportunities), and the kind of resulting
A., Levin, S.A., Ma¨ler, K.-G., Moberg, F., Ohlsson, L.,
deleterious or beneficial transformation of the system. The Olsson, P., Ostrom, E., Reid, W., Rockstro¨m, J., Savenije, H.,
investigation of change and transformation of the SES Svedin, U., 2002. Resilience and sustainable development: building
would require the collaboration of social and natural adaptive capacity in a world of transformations. Report for the
scientists, as well as system theorists and mathematicians, Swedish Environmental Advisory Council 2002:1. Ministry of the
Environment, Stockholm, Sweden.
7
An interesting step in this direction is the development of the adaptive Gallopı´n, G.C., 1991. Human dimensions of global change: linking
and renewal cycle by the Resilience Alliance (see Folke, 2006) although the global and the local processes. International Social Science
up to now the adaptive cycle is more of a metaphor or a model than a Journal 130, 707–718.
general theory of change in SES. Gallopı´n, G.C., 2003. Box 1. A systemic synthesis of the relations
between vulnerability, hazard, exposure and impact, aimed at policy
identification.
 G.C. Gallopı´n / Global Environmental Change 16 (2006) 293– 3
In: Economic Commission for Latin American and the Caribbean Luers, A.L., 2005. The surface of vulnerability: an analytical framework
(ECLAC). Handbook for Estimating the Socio-Economic and for examining environmental change. Global Environmental Change
Environ- mental Effects of Disasters. ECLAC, LC/MEX/G.S., 15, 214–223.
Mexico, D.F., pp. 2–5. Nicolis, G., Prigogine, I., 1977. Self-Organization in Non-Equilibrium
Gallopı´n, G.C., Gutman, P., Maletta, H., 1989. Global Systems: From Dissipative Structures to Order through Fluctuation.
impoverish- ment, sustainable development and the environment. Wiley, New York.
A con- ceptual approach. International Social Science Journal 121, Nicolis, G., Prigogine, I., 1989. Exploring Complexity. Freeman, New York.
375–397. Pimm, S.L., 1984. The complexity and stability of ecosystems. Nature 307
Gallopı´n, G.C., Funtowicz, S., O’Connor, M., Ravetz, J., 2001. (26), 321–326.
Science for the 21st century: from social contract to the scientific Prigogine, I., Stengers, I., 1979. La Nouve` lle Alliance. Me
core. International Social Science Journal 168, 219–229. ´tamorphose de la Science. Gallimard, Paris.
Gunderson, L.H., 2000. Resilience in theory and practice. Annual Review Schellnhuber, H.J., 1998. Earth system analysis—the scope of the
of Ecology and Systematics 31, 425–439. challenge. In: Schellnhuber, H.J., Wenzel, V. (Eds.), Earth System
Gunderson, L.H., Holling, C.S. (Eds.), 2002. Panarchy. Island Press, Analysis: Integrating Science for Sustainability. Springer, Heidelberg.
Washington, DC. Smit, B., Wandel, J., 2006. Adaptation, adaptive capacity and vulner-
Hahn, W., 1967. Stability of Motion. Springer, Berlin. ability. Global Environmental Change 16 (3), 282–292.
Holling, C.S., 1973. Resilience and stability of ecological systems. Smithers, J., Smit, B., 1997. Human adaptation to climatic variability and
Annual Review of Ecology and Systematics 4, 1–23. change. Global Environmental Change 7 (2), 129–146.
Holling, C.S., 1985. Perceiving and managing the complexity of Thom, R., 1972. Stabilite´Structurelle et Morphoge´ne` se. W.A.
ecological systems. In: Aida, S., et al. (Eds.), The Science and Praxis Benjamin, Reading, MA.
of Complexity: Contributions to the Symposium Held at Montpellier, Tomovic, R., 1963. Sensitivity Analysis of Dynamic Systems. McGraw-
France, May 9–11, 1984. GLDB-2/UNUP-560. United Nations Hill, New York.
University Press, Tokyo, pp. 217–227. Tu, P.N.V., 1994. Dynamical Systems: An Introduction with Applications
Holling, C.S., 1986. The resilience of terrestrial ecosystems: local surprise in Economics and Biology, second ed. Springer, Berlin.
and global change. In: Clark, W.C., Munn, R.E. (Eds.), Sustainable Turner II., B.L., Kasperson, R.E., Matson, P.A., McCarthy, J.J., Corell,
Development of the Biosphere. IIASA/Cambridge University Press, R.W., Christensen, L., Eckley, N., Kasperson, J.X., Luers, A.,
Cambridge, UK, pp. 292–317. Martello, M.L., Polsky, C., Pulsipher, A., Schiller, A., 2003. A
Holling, C.S., 1996. Engineering resilience versus ecological resilience. framework for vulnerability analysis in sustainability science. Proceed-
In: Schulze, P.C. (Ed.), Engineering within Ecological Constraints. ings of the National Academy of Sciences of the United States of
National Academy Press, Washington, DC, pp. 31–43. America 100 (14), 8074–8079.
IPCC (Intergovernmental Panel on Climate Change), 2001. Technical van der Leeuw, S.E., 2001. ‘Vulnerability’ and the integrated study of
summary: climate change 2001: impacts, adaptation, and vulnerability. socio-natural phenomena. IHDP Update 2/01, art. 2 [online]. URL:
A Report of Working Group II of the Intergovernmental Panel on http://www.ihdp.uni-bonn.de/html/publications/update/IHDPUpdate01_
Climate Change. URL: http://www.grida.no/climate/ipcc_tar/wg2/ 02.html.
pdf/wg2TARtechsum.pdf. Walker, B., Holling, C.S., Carpenter, S.R., Kinzig, A., 2004. Resilience,
Janssen, M.A., Schoon, M.I., Ke, W., Bo¨rner, K., 2006. adaptability and transformability in social-ecological systems. Ecology
Scholarly networks on resilience, vulnerability and adaptation within and Society 9 (2) art. 5 [online], URL: http://www.ecologyandsociety.
the human dimensions of global environmental change. Global org/vol9/iss2/art5.
Environmental Change 16. Young, O.R., 2005. Institutional Dynamics: Resilience and Vulnerability
Jen, E., 2003. Stable or robust? What’s the difference? Complexity 8 (3), in Environmental and Resource Regimes. September 2005 Draft; cited
12–18. with permission from the author.
Kasperson, J.X., Kasperson, R.E., Turner II., B.L., Schiller, A., Hsiel, W.- Young, O.R., Berkhout, F., Gallopı´n, G., Janssen, M.A., Ostrom, E.,
H., 2005. Vulnerability to global environmental change. In: Kasper- van der Leeuw, S., 2006. The Globalization of Socio-Ecological
son, J.X., Kasperson, R.E. (Eds.), Social Contours of Risk, vol. II. Systems: An Agenda for Scientific Research. Global Environmental
Earthscan, London, pp. 245–285. Change 16 (3), 304–316.