Updated: Sep 1, 2020
The Complex Dynamics of Collapse
Keywords: Collapse; Civilization; Population Growth; Environment; History; Complex System Dynamics
“Probably the greatest agreement among scholars, though, is that the failing civilizations suffered from growing hubris and overconfidence: the belief that their capacities after many earlier tests would always rise to the occasion and that growing signs of weakness could be ignored as pessimistic."
- Jeremy Grantham, 2013
“Systems often hold longer than we think, but they end up collapsing much faster than we can imagine.”
- Ken Rogoff, 2012
The historical collapse of ancient complex human societies has enjoyed a sustained scholarly and popular enchantment. Two world wars of abominable inhumanity fused with exponential population growth, the deterioration of the living environment, global climatic changes, and the growing scarcity of renewables and non-renewable resources - presaging hostility and bloodshed over their dominion – have all emphasized the uneasiness of societal collapse (Butzer & Endfield, 2012), and of global catastrophic risks (Kareiva & Carranza, 2018; Beard, Rowe & Fox, 2020; Kuhlemann, 2018) with the possibility of human extinction (Carpenter & Bishop, 2009).
Even though the subject of collapse has drawn substantial fascination over the ages, the field has persisted with substandard elucidation, with current theories of past events, starving for appropriate modeling (Greer, 2005). Moreover, there appears to be a slight tendency in the study of collapse, to search for distressing historical occurrences and extrapolate them into future events (Butzer & Endfield, 2012), although, and to not beat around the bush, we all have reasons to be alarmed, since societies with much lesser anthropic impact and entropic potential still succumbed to their mores (with environmental nudges). Given this, scientific publications that contemplate global catastrophes and an increasing probability of contemporaneous collapse are becoming more numerous and sustained by growing bodies of evidence (Servigne & Stevens, 2020).
To point out, in 2013, the Royal Society published Can a collapse of global civilization be avoided? (Ehrlich & Ehrlich, 2013) in which it presented the case that the consequences of anthropic global environmental changes had become so conspicuous and massive in scale that discussing the palpable possibility of collapse had become not just a legitimate but an insistent necessity. Equally important, the World3 model from Limits to Growth (Meadows et al. 1972, for a review, check Abegão, 2020a) has forewarned humanity, forty-eight years ago, to the likelihood of overshoot and collapse during the 2020-2030 decade if current trends are maintained. So far, World3 has anticipated reality with an astonishing degree of precision (Turner, 2014), which, undoubtedly, leads us to pose an admissible question, as the ecological economist William Rees has done (2017):
“Why are we not collectively terrified or at least alarmed? If our best science suggests we are en route to systems collapse, why are collapse — and collapse avoidance — not the primary subjects of international political discourse? Why is the world community not engaged in vigorous debate of available initiatives and trans-national institutional mechanisms that could help restore equilibrium to the relationship between humans and the rest of nature?”
In any event, this work does not intend to serve as an admonition to our collective inability to fend off collapse but to study and understand how the collapse in complex systems emerges. For that to happen, I have deemed it instructional to first delve into the collapse of past complex societies, in an attempt to discern the prime factors that engendered that outcome, thereupon examining the science sustaining it. To do that, I will first explore the roots of this field of reference.
The intricate subject of collapse was first coherently expressed by the Islamic historian Ibn Khaldun (1377 CE), who pinpointed the recurrent rise and fall of dynasties as macrostructures in the record of sedentary civilizations. In particular, Khaldun scrutinized the Roman Empire and past Islamic counterparts, crediting their dissolution to rural insurgencies or foreign invaders, which opposed a ruling hierarchy that had ceased to be supported by its people. Still, Khaldun didn't attribute the fall of those societies to global events, but instead to something which will be repeated throughout this work, the greed, selfishness, and inadequacy of ruling elites which end up disturbing the sociopolitical order (Butzer, 2012).
Succeeding Khaldun’s study, Western pursuit over the subject of collapse was catalyzed by Edward Gibbon during the eighteen century in his The History of the Decline and Fall of the Roman Empire (1776-1789). Following his predecessor’s footsteps, Gibbon traced the weakening and downfall of the Roman Empire to moral decay as well as barbarian assaults, although he diverged from Khaldun since he was still convinced that the integral Roman dimension could not be divorced from historical processes that had framed the dynamic context of the time. Moreover, Gibbon acknowledged the potential undoing of more complex, powerful, and civilized states (Butzer, 2012).
Another timeless contribution to the discussion of collapse came to us from Oswald Spengler’s The Decline of the West (1918–1922). Spengler wrote about a dissolution of abstract thought, escorted by the ascendency of the rich, and the rise of malevolent and treacherous leaders. Spengler envisioned a society in a profound crisis, and his farsighted but pessimistic notions predicted the horrors of fascism and Stalinism. His vision remains admissible for modeling alternative pathways of political resilience in the wake of collapse (Butzer, 2012), because as we shall see, collapse is hardly ever unifactorial, with a plexus of interdependencies comprising the complex systems that lead to collapse.
Supplanting Spengler were the French scholars of the Annales School, who jolted the debate over collapse, by introducing an interdisciplinary method, in which demographic tendencies were constrained key interactive processes (Braudel, 1958; Whitmore, Turner, Johnson, Kates & Gottschang, 1990; Blintiff, 1991). The disconnections in the systems tended to occur due to a mix of struggling economic systems, long-distance networking, rivalry, and strife, or pandemics (Braudel, 1979) which gave way to the macro-scale and multidisciplinary approach to world history called world-systems theory (Frank & Gills, 1993; Chase-Dunn & Hall, 1997). Eventually, the annalistes adjusted their focus to more humanistic studies incorporating environmental fluctuations as an essential part of historical occurrences (Ladurie, 1978; 1976).
Provided that collapse and complex systems are what instructs this inquiry, it is indispensable to pore over the work of Joseph Tainter, in Collapse of Complex Societies (1988). In detail, Tainter has proposed a general theory of collapse, in which societies succumb when their level of complexity swells to the point that negative marginal returns take place, thus, a disintegration in sociopolitical and cultural complexity occurs, in other words, a simplification.(1)
Tainter wrote that:
“A civilization is the cultural system of a complex society. The features that popularly define a civilized society – such as great traditions of art and writing – are epiphenomena or covariables of social, political, and economic complexity. Complexity calls these traditions into being, for such art and literature serve social and economic purposes and classes that exist only in complex societies. Civilization emerges with complexity, exists because of it, and disappears when complexity does. Complexity is the base of civilization, and civilization, by definition here, can disappear only when complexity vanishes.” (Tainter, 1988)
Tainter has vastly enriched the dialectic surrounding civilizational collapse, by modeling different features that characterize it with a focal point on sociopolitical complexification. Besides, Tainter was also on the mark when he combined a thermodynamic factor, as in the case that as the complexity of sociopolitical institutions increases, a higher 'metabolic cost' develops – growing needs for matter, energy, and resultant low entropy. Under these circumstances, more elaborate civilizations are entangled on an entropic trap, which becomes hardly possible to escape (Servigne & Stevens, 2020). The political scientist William Ophuls asserts in Immoderate Greatness: Why Civilizations Fail (2012) that when:
“The available energy and resources can no longer maintain the existing level of complexity, the civilization begins to consume itself by borrowing from the future and feeding off the past, thereby preparing the way for an eventual implosion.”
At the same time, alternate and equally viable models of collapse based on human ecology and environmental degradation have been suggested (Hughes, 1975; Sanders, Parsons & Santley, 1979; Elvin, 1998), with the concept witnessing tremendous popularity with Jared Diamond’s Collapse – How Societies Choose to Fail or Survive (2005).
Diamond (2005) has identified five frequent and often synergistic factors, which were capable of precipitating collapse in past civilizations: environmental damage or depletion of resources; climatic changes; wars, the abrupt and unforeseen loss of trading partners and the inflexible reactions of society to mounting environmental problems. Nevertheless, Diamond (and others) affirm that the factor linking all the previous collapses is the last on his list: that of a decaying sociopolitical order, through institutional dysfunction, ideological myopia, escalating levels of inequality, and above all the inadequacy of society – particularly the elites – to proceed accordingly with the threat of catastrophic events (Servigne & Stevens, 2020).
Similarly, as John Greer proposes (2005), the relationships among resources, capital, waste, and production can lead to self-reinforcing cycles of decline, for instance, when production breaks down and maintenance of existing capital is no longer assured. When essential resources are depleted and crises mount, societies risk entering into a catabolic collapse, in which a self-reinforcing cycle of conversion of capital to waste (consisting of any material that is used or converted into a pollutant) occurs. Furthermore, and extremely relevant to this examination is the fact that systemic failures in one synergistic network have the capability to cascade and undermine other structures (Butzer, 2012). I will return to this later when examining how complex networks respond to disturbances.
Another key contribution comes to us from the historians Peter Turchin and Sergey Nefedov, who have generalized the phenomenon of simplification by characterizing and modeling recent history as a concatenation of stages of surplus and deficit in economic and energetic terms, which end up defining cycles of growth and decline (Servigne & Stevens, 2020). To explain, Turchin and Nefedov have identified several of these junctures, as in Medieval England (the Plantagenet cycle) and premodern England (the Tudor-Stuart Cycle), medieval France (the Capetian cycle) and ancient Rome (the republican cycle) have all undergone periods of expansion, stagflation, entanglements, and decline (Turchin & Nefedov, 2009).
Similar to Diamond’s five factors, Dmitry Orlov proposed a theoretical framework consisting of five stages that increase the seriousness of the collapse. As a result, the given system might be able to stall the progression of collapse at one of the stages, or cave in further through a spiral of collapse (correspondingly, the Cambridge Center for the Study of Existential Risk is currently scrutinizing the concept of ‘global systems death spiral’ (Dembicki, 2019)]. Orlov’s scale equips us with a more realistic sense of a gradation of collapse, with different severity and causes, consisting of a financial collapse first, followed by a commercial collapse, ensued by a political collapse, supplanted by a social collapse and finally a cultural collapse (Orlov, 2013a). Recently Orlov (2013b) added a sixth and last stage, that of ecological collapse (I will not delve further into this fascinating scale, but I invite the curious reader to acquire Orlov’s book as I just did).
At this point, the reader will be able to perceive how the analysis of collapse has witnessed various perspectives, ranging from ethical and social, to ideological and ethnocentric, to interdisciplinary and systemic, with the present concerns mainly focusing on the interface between environment and society, with the emergence of systems thinking, network theory, planetary boundaries, and tipping points (Butzer, 2012).
Correspondingly, picking off from ‘Diamond’s fifth factor’, the archaeologist and geographer Karl Butzer advanced a new classification. Butzer understood that the fifth factor was increasing the vulnerability (reducing the resilience of the system) and enhancing its sensitivity to stressors which it would normally be able to cope with, if not for the sociopolitical disorder (Servigne & Stevens, 2020). For this reason, Butzer dichotomized collapse by its ‘preconditions’ (which heighten societal vulnerability) and its 'triggers' (the forces which can throw society off the edge) or ‘concatenations’ (when several triggers act and cross thresholds leading to a downwards spiral of cascading feedbacks) (Butzer, 2012; Dembicki, 2019).
To demonstrate (figure 1), preconditions are often considered endogenous, such as elite corruption or incompetence, uncontrolled pastoralism [(G. Hardin’s Tragedy of the Commons (1968), illustrates how this can become a problem)], declining agro-productivity, anthropogenic degradation, poverty, among others. These preconditions reduce the resilience of society and are factors of decline (Servigne & Stevens, 2020).
On the other hand, triggers contemplate extreme weather events, foreign invasions, loss of trading partners or satellite regions, depletion of natural resources, economic crises, etc. If the resilience of the society is low, cascading feedbacks will ensue, an initial breakdown will take place and instability comes to dominate the system. In contrast, if resilience is higher, buffering feedbacks will occur and the system might stabilize and be reconstructed, but since this work is focused on the path to collapse, I’ll overlook the events of resilience and reconstitution (Butzer, 2012).
Before moving on further, there is something that I would like to add to this conceptual model by Butzer (2012). As my academic predilection is in human overpopulation, I couldn’t help shake the feeling that this model might have discounted the role of too many humans surpassing the carrying capacity of a place, elevating overpopulation to a precondition or trigger. Allow me to elaborate.
In his book Collapse, Jared Diamond (2005) illustrates how several societies navigated through the nadir of their downfall, and one prevalent factor is how a given period of favorable conditions (e.g., precipitation or more amenable temperatures which support greater agricultural yields), aids the population to grow above the limits that the territory would normally allow. In turn, when conditions change (e.g., drought, loss of economic networks or depletion of resources such as wood or hunting prey, etc.), the society finds itself with too many mouths to feed, which can then accrue into many of the preconditions and triggers exemplified by Butzer (2012).
As an illustration, Diamond (2005) examines the Anasazi society which inhabited the territory now known as Chaco Canyon in northwest New Mexico (they had satellite regions, but I'll focus on Chaco Canyon which was known as an entropic black hole because it had no autonomy and depended on these other regions). Between 700 and 1130 it has been shown that their population grew, with a concomitant increase in complexity and interdependency, which left the Anasazi without any self-sufficiency during some rough years. However, these rough years were particularly harsh for them because the Anasazi had been growing in size beyond their means, setting themselves for disaster when the continuum of beneficial conditions was disrupted. For instance:
When their irrigation structures stored more water from especially rainy seasons increasing agricultural yields, although when drought was installed the people at Chaco Canyon became completely dependent on their satellite regions seeing that the groundwater level diminished to the point of being unworkable;
When they settled in some areas for a couple of decades until the soil and hunting were exhausted and then migration was required, although, when their populations grew, population density increased and more of these territories were simultaneously occupied, eliminating the possibility that the territories could regain vegetation and soil nutrients;
When the population increased and more materials were needed to build structures and provide energy, the Pinus and Cedrus trees were ultimately deforested (to the point where the ecosystem hasn’t hosted any of these trees for roughly 1000 years and is now a desert). This left the Anasazi without the seed of the Pinus trees which was a part of their diet, no timber for construction and animals to be hunted (since they had no habitat left);
As the population grew and conditions for their maintenance were not found the initial breakdown (figure 1) had commenced, with the people becoming more resented with the elites and religious authorities, who were not fulfilling their end of the bargain of bringing forth more rain for food and materials, while at the same time inequality increased as hoarding of resources was initiated (loss of authority, political simplification, and polarization of elites). Loss of faith in the hierarchy led the satellite regions to stop providing Chaco Canyon with food, ceramics, and other crucial materials, dissolving an essential economic network and the sociopolitical order (subsistence crises, depopulation). Facing hunger, a conflict between the regions broke out (Advanced Breakdown), probably for more than a century, as it is testified by the sudden rise in constructions of defensive walls, towers, and pits (recall Greer’s catabolic collapse, and how a society with already sparse resources is forced to convert more capital into waste as it spirals out of control). A fracture of social order ensued and cannibalism followed. Finally, Chaco Canyon was abandoned sometime between 1150 and 1200 and mass migration took over, for those that hadn’t perish in civil conflict or from famine. The place was to be re-occupied only by the Navajo circa 1800.
The Anasazi provide what I recognize to be a series of crises precipitated by a population that had grown beyond its carrying capacity, at the same time that anthropogenic degradation was expedited by too many end-users. To this effect, I would add population growth to the preconditions and overpopulation to the triggers recited by Butzer (2012). Accordingly, if we follow the bottom arrows of the conceptual diagram (figure 1), we can identify some of the traces of Initial Breakdown/Instability and eventual Advanced Breakdown/Collapse in the Anasazi society. As Jared Diamond argues (2005), the risks of growing your population beyond their means of subsistence brings forth interdependencies and complexity, which wouldn’t, on one hand, inflict so much damage on natural systems, while on the other hand, they would become more self-sufficient by not overshooting the limits of their territories. Furthermore, as Diamond argues throughout his book, overpopulation precedes sociopolitical disintegration (as in the Maia, the Anasazi, and the Greenland Norse, for example).
As the reader can discern from the Anasazi, even though increases in population played a huge part in the downfall of their society, Butzer (2012) accentuates the need to rise above simple assumptions or truisms and monocausal descriptions, so we can become aware of the interplay between an amalgamation of socio-ecological variables and the feedbacks that might ensue. Although this may be true, according to the archeologist Shelley Hartman in her work Entropy, Individualism and the Collapse of Empires (2014), the primary causal factor of collapse is the expansion of the population and the resource depletion that ensues. Hartman’s assertion complements that of Diamond connecting population growth and sociopolitical decline, acting both as the antecedent of civilization and inducer of collapse:
“When population size increases, there is a requirement for increased organizational complexity, which leads to hierarchies, constitutions, and bureaucracies. Without a relative growth in complexity with size, the cultural or living entity will not survive. Civilization is the managerial, organizational response to population pressure […] but unchecked, unlimited growth of population is not advisable […] when there are too many people, major “trigger” disasters, events that disrupt the poised state of an empire – flood, famine, climate change, war, earthquakes, plagues - produce situations from which there is no recovery.”
Hartman supplements the discussion around population growth with another essential argument, that of a population growing beyond not just its ecological means but also beyond its social cohesiveness. For the archeologist, when populations expand, the functioning of hierarchies is threatened by exceeding the functional limits of what came to be known as the ‘Dunbar Number’ (2010), the maximum number of relationships that people can maintain (roughly 150). Hartman links this multiplier of 150 and the fact that when populations continue to grow a major crisis arises from people being disconnected from those at the top of the hierarchies, leading to a disseverance from collectivism and the emergence of individualism. For Hartman, “individualism is the early sign of system collapse”, and this is illustrated in thermodynamics (Hutchinson, 1962 quoted in Hartman, 2014):
"If atoms of this material are all moving in unison in the same direction, minimal energy is lost. However, the energy dissipates rapidly if the units move in uncoordinated random directions."
All in all, Hartman (2014) upholds that:
“Large complex civilizations are without doubt impressive, but they all have the same weakness: the people themselves. Humans have limiting factors – exceptional fecundity, top predator status, habitual over-population of environments, short-term thinking […] little self-restraint, and, perhaps most dangerous of all, hubris. The result: every empire has built within it already, the elements of its inevitable collapse.”
Having now maneuvered through this preliminary diagnosis of the collapse of past societies, I would like to move this examination into the study of the dynamics of complex systems and networks, with the intent of unveiling that a model of systemic collapse might be the most relevant mode of behavior of a highly complex system such as that affecting our industrial civilization.
As already mentioned, the problem of complexification appears to be at the root of collapse. Joseph Tainter argued in The Collapse of Complex Societies that the inexorable tendency of societies to self-perpetually seek greater levels of complexity, specialization, and sociopolitical control become the origin of their collapse.
On a highly industrialized and globalized society such as the one that came to dominate most of the planet in the 21st century, the requirement to accumulate natural resources which become progressively harder to encounter and obtain, leads to increasingly more expensive costs [and diminishing returns, which Tainter (1988) considers the cause of collapse, although Hartman (2014) argues that the law of diminishing returns is driven by overpopulation)], while at the same time, the complexification leads to the breaching of boundaries (Rockström et al. 2009; Steffen et al. 2015) and limits (Meadows et al. 2004), consequently growing more vulnerable to risks and disruptions. In this second part, I intend to examine how complex systems behave when they leave the zone of stability and surpass the thresholds that hold them in place.
For a long time, the chief perception was that nature reacted to disruptions steadily and consistently. However, the reality is that ecosystems which suffer disruptions (hunting, fishing, pollution, deforestation, droughts, etc.), do not show signs of impairment instantaneously, but progressively – and undetectably – are deprived of their capacity to rebound (by losing their resilience) until arriving at a tipping point (Steffen et al. 2018), an inconspicuous threshold beyond which the ecosystem breaks down in a brisk and unstable way (Sheffer et al. 2001).
For example, in a biome of semi-arid forests (just as the one which used to dominate the landscape where the Anasazi inhabited), once enough vegetation cover is removed, the soil dries out under increased temperatures and reduced moisture retention, precipitating a severe transition into a shrub-predominated biome (Kéfi et al. 2007). Such tipping points couldn’t be more relevant than those that that have been identified to have the potential to dramatically and catastrophically accelerate climate change (Lenton et al. 2008; Schellnhuber, 2009), as well as those affecting global ecosystems (Kareiva & Carranza, 2018; Barnosky et al. 2012).
The existence of these tipping points is inherent to the connectivity and homogeneity of systems, arising due to a mixture of feedback loops and domino effects. To be sure, a complex living system (ecosystems, societies, markets, etc.) is expressed by innumerable interlacing feedback loops that maintain the system stable and with considerable resilience. However, when the system is about to tip over to another state, one small disruption is enough to trigger cascading loops which can lead to a violent collapse with a low probability of rebound or the transformation into another state of equilibrium, categorically more stable and resilient, but at the expense of whatever depended on its previous state (Servigne & Stevens, 2020).
Admittedly, as can be observed in figure 2, complex networks are susceptible to the factors of heterogeneity and connectivity between their integral elements (Scheffer et al. 2012). A heterogeneous and modular network (autonomous parts with insubstantial links) will tolerate shocks by adaptation. The losses suffered are local with the system becoming progressively more damaged. On the contrary, a homogeneous and exceptionally connected network withstands change at first because local losses are absorbed through the connectedness of the elements in the network. However, if turmoil persists, a domino effect will take over, with catastrophic consequences. The seeming resilience of homogeneous and highly linked systems camouflages an unnoticeable fragility (Servigne & Stevens, 2020), with the pertinent version of Liebig’s Law expressing that any complex system contingent on several pivotal inputs can be degraded by a single factor in the least supply (Rees, 2017). As Hartman (2014) affirms:
“In a mature complex society, all sections are so interdependent that perturbation in the smallest part of the system ripples outward in all directions.”