Human activities have dramatically altered the Earth. Analysts in a variety of academic fields have attempted to better understand the drivers and impacts of human pressures on the Earth’s life support systems, how these pressures have changed over time, and the ways in which societies have acted to both address and adapt to them. Reducing the human footprint on environmental processes while protecting and enhancing human well-being is at the heart of sustainability. To this end, some researchers are interested in better understanding the structure, function, and outcomes of complex adaptive systems that are relevant to sustainability. Others focus on defining critical aspects of sustainability or studying the dynamics of transition processes. Still others examine the design, implementation, and effects of governance arrangements for sustainability at and across various geographical scales. Insights of relevance to these different groups of researchers can be drawn from the analysis of the five mercury systems.
Humans have mobilized large amounts of mercury from the Earth’s crust, especially over the past five centuries. This has had large environmental and human health consequences, as we detailed in part II. Uses and discharges of mercury occur in parallel with other accelerating human pressures. These include increases in energy use, population, consumption, carbon dioxide emissions, nitrogen pollution, and biodiversity loss (Steffen et al. 2015a). Addressing these human influences on the Earth is an urgent matter. Researchers in different fields share common interests in how changes to the Earth’s life support systems occur, and how they threaten human well-being. Many scientists investigate the dynamics and consequences of altered physical and biological processes. Other scholars engage in efforts to better define sustainability, and to explain how transitions toward greater sustainability can be understood and encouraged. Those interested in governance issues examine the processes by which people come together to collectively address human activities that harm the environment and human well-being. All of these fields share a goal of understanding and theorizing about Earth systems that are under pressure from human activities in novel and changing ways.
The idea of human activities pressuring Earth systems relates to the message sent to negotiators of the Minamata Convention when Freddie Mercury and Queen’s “Under Pressure” echoed over loudspeakers to urge delegates to address specific aspects of the mercury issue. Insights from the analysis of the mercury systems provide fodder for further developing and testing theories (either mid- or smaller range) about the ways in which systems respond to human pressures, for charting the progress of systems toward sustainability, and for identifying the functions and dynamics of institutions that set out to govern them. In this chapter, we answer our fourth research question—What insights can be drawn from analyzing these systems?—by synthesizing across the three areas that we identified in part I and addressed in each chapter of part II: systems analysis for sustainability, sustainability definitions and transitions, and sustainability governance. We draw these insights with the aim of distilling actionable knowledge to support the formulation of collective, but often contentious, decisions on how to steer societies toward greater sustainability.
Multiple sets of literatures apply systems-oriented approaches to examine connections between identified components. Many systems relevant to sustainability are complex and adaptive, and their behavior can be difficult, if not impossible, to predict (Sterman 2011). Different scholarly fields propose varying methods and approaches for analyzing complex adaptive systems (e.g., de Weck et al. 2011; Binder et al. 2013). In this section, we draw three insights that are relevant to those who are interested in better understanding systems and their behavior. First, new systems-oriented analytical approaches could better account for dynamics of human, technical, environmental, institutional, and knowledge components in an integrated way. Second, adaptive capacity stands out as an important dynamic that explains system behavior over time, and it can have both positive and negative effects on human well-being. Finally, concepts that guide systems analysis in the Anthropocene could better capture system-wide variability and changes over temporal and spatial scales.
Understanding Complex Adaptive Systems
Much systems analysis is conducted within individual disciplinary fields. Nevertheless, as we discussed in chapter 8, gaining a deeper understanding of the dynamics of complex adaptive systems relevant for sustainability requires further integration of perspectives and expertise from multiple disciplines. The importance of multidisciplinary analysis is well known, yet few efforts have been able to achieve an integrated perspective that gives different types of system components and their interactions equal attention. Our analysis of the mercury systems illustrates why such an integrated perspective is methodologically and analytically necessary. It also suggests several ways forward for advancing such multidisciplinary analysis of interactions among people, technology, the environment, institutions, and knowledge in order to better understand interactions between different system components, and to explore and assess opportunities through which actors can intervene to promote greater human well-being.
In applying a systems perspective, researchers who study the environmental cycling of materials like mercury could benefit from focusing on the combined impacts of human and natural drivers. For example, biogeochemical cycling analyses aim to quantify the degree to which environmental concentrations of a variety of chemical elements have been increased by human activities; they sometimes do this by calculating global enrichment factors (Sen and Peucker-Ehrenbrink 2012; Schlesinger et al. 2017). Many mercury researchers focus on understanding to what extent global environmental concentrations of mercury have been enhanced due to human activities (Selin et al. 2008; Amos et al. 2013; Outridge et al. 2018). But a growing amount of historically emitted mercury, from both anthropogenic and natural sources, is remobilized from environmental storage by human-induced land-use changes and climate change, and as a result the distinction between natural and anthropogenic flows is becoming increasingly artificial. Thus, a focus on trying to specifically attribute the footprint of human influence may be less relevant to sustainability analysis than more comprehensive systems-oriented approaches.
Systems analysis that better accounts for two-way interactions between environmental processes and human activities is especially relevant to sustainability-focused studies. Uses and discharges of mercury both affect, and are affected by, environmental processes. The prevalence of ASGM and associated mercury use may increase as a result of ecological deterioration, pushing farmers who can no longer make a living growing crops and raising animals into the mining sector. The availability and relative price of coal and oil as feedstocks influenced the amount of mercury that was used in the chemicals industry, leading to discharges of mercury into the environment. Improved scientific understanding of the environmental and human health risks from mercury prompted societal concerns and the adoption of controls on mercury emissions and releases. Treating anthropogenic discharges of mercury as external to a system may be useful for addressing some questions, such as identifying the total amount of present-day emissions and releases. However, studies that do not account for feedbacks between the environment and human activities provide only a partial picture of the material flows that are important to sustainability analysis.
Efforts to quantify how toxic substances cross the globe should account for both societal and environmental transport. The long-range transport of mercury and other toxic substances is substantial regardless of whether it occurs via the atmosphere or international trade. Many quantitative models for mercury and other pollutants nevertheless calculate the impact of one region on another purely based on environmental factors such as long-range atmospheric transport (Fiore et al. 2009; Corbitt et al. 2011). Societal use and flows of substances are often overlooked in models of environmental transport, but they are also relevant in assessing the impact of one region on another. Mercury and many other hazardous substances that move through trade affect human well-being far from the site of their extraction or production. But reliable data on cross-border trade flows are lacking for many commercial substances, including mercury (UNEP 2017). This creates a need for the scientific community to expand efforts to gather further data on environmental cycling together with societal cycling of substances for the purpose of improving sustainability analysis and informing decision-making.
One way forward may be to better integrate biogeochemical cycling studies, environmental modeling, and methods from industrial ecology. A large body of literature in industrial ecology and related fields on material flows through society and the economy employs concepts of industrial, urban, and socio-economic metabolism to better understand the flow of pollutants (Ayres 1989; Fischer-Kowalski and Haberl 2007; Ferrão and Fernández 2013). Studies on heavy metals focus on whether such materials are recovered and recycled, or if they are released or “dissipated” in the environment (Müller et al. 2014). For example, mercury use in a pesticide is considered dissipative because it cannot be recovered, while mercury in a closed product such as a thermometer can, at least in principle, be recycled. Environmental modeling focuses on the results of this dissipation in the form of environmental discharges, transport, and transformation. Some work has been done to link material flow analysis with mercury transport and environmental behavior, for example in China (Hui et al. 2016; Wu et al. 2016). Further combining these perspectives would provide a more complete systems-level accounting of sustainability issues.
The development and use of quantitative models that capture dynamics of human, technological, and environmental interactions simultaneously helps researchers to better understand system behavior and develop related theories. Integrated modeling that accounts for both human and environmental dynamics has a long history (Dowlatabadi 1995). This is especially the case in climate change modeling, where integrated models simulate the economic drivers of fossil fuel emissions and the environmental and human health consequences of climate change. Some researchers have called for further improvements in representing human and environmental dynamics in models for studying climate change, as well as other issues such as forestry, agriculture, and marine ecosystems (Bonan and Doney 2018; Calvin and Bond-Lamberty 2018). Insights from the mercury systems reinforce that any system model intended to explore sustainability-relevant questions would likely neglect important information if it omitted whole categories of human, technical, or environmental components or related interactions. Effectively incorporating all three categories is a challenge, however, and requires model development. Capturing technology development, for example, has proved particularly difficult in integrated models that account for both human and environmental dynamics (Ackerman et al. 2009).
It is important to recognize that adding more complexity to models may not always serve a useful analytical purpose for sustainability-related research. Including specific dynamics may only be relevant for sustainability analysis where they lead to different conclusions and suggest alternate intervention strategies. Analyses of different sustainability issues also require different model capabilities. Including and quantifying specific biological mechanisms for methylmercury uptake in animals and humans may be necessary for modeling a particular ecosystem with the goal of designing better interventions in the form of updated and more detailed dietary recommendations. But that kind of detailed analysis may be less relevant to efforts that seek to better understand and phase out mercury use in production processes with the aim of reducing worker exposure and environmental discharges. In such a case, a model of the use and flows of mercury in a specific type of manufacturing plant may be more analytically relevant. This variation in modeling needs across different aspects of a single issue such as mercury suggests the value of having access to a spectrum of different models, as we discussed in chapter 8.
Adaptation Dynamics and Human Well-Being
Much systems analysis focuses on better understanding how systems reconfigure as a result of disturbances or changes. Previous literature identifies adaptive capacity or adaptability as an important property of systems, including those relevant to sustainability (Walker et al. 2004; de Weck et al. 2012). The concept of adaptive capacity, as the ability of a system to adjust to maintain key functions when under stress, is also related to the idea of resilience, which can be defined as the capacity to maintain a desired set of functions in the face of disturbances (Biggs et al. 2012). Carl Folke (2016) argues that resilience “reflects the ability of people, communities, societies and cultures to live and develop with change, with ever-changing environments.” The resilience literature also addresses the important difference between maintaining system function as a positive characteristic, and the kind of negative rigidity in systems that may impede desirable transformations (Olsson et al. 2014). A high degree of adaptation is often seen as positive for maintaining system operations over time, but it is sometimes necessary to actively break down a system because it generates outcomes that undermine human well-being. In such instances, a resistance to change poses an obstacle to interventions that support greater sustainability.
Our analysis of the mercury systems further highlights the importance of understanding the adaptive behavior of complex systems when examining their dynamics over time. Technological innovation emerged as a type of adaptation that was important in the products and processes system, as mercury use over time both increased and decreased with the development of new techniques for producing goods. Innovation was also prominent in the atmospheric system, which reconfigured in response to regulatory pressure to reduce mercury emissions from industrial point sources. For example, the application of emission control technologies resulted in fewer mercury emissions, but the control technology altered the fraction of different forms of emitted mercury that travels short and long distances. Market forces and human social and economic behavior contributed to adaptation as well. Changes in food availability and dietary preferences as a result of mercury contamination affected people’s exposure to and impacts from methylmercury. Ecological degradation led to surges in the number of ASGM miners, and supply restrictions on mercury led to the illegal reopening of old mercury mines and trade.
Dynamics of adaptation can have varying and sometimes conflicting impacts on human well-being, leading to benefits as well as harms. Innovation, involving technical and human components, maintained the ability of several mercury systems to provide socially valuable goods over time. This was important for developing new mercury-free products and production processes and end-of-pipe control technologies for industrial point sources. The ability of producers and societies to innovate served to maintain benefits from the consumption of goods and energy while reducing mercury use and discharges. In contrast, adaptive dynamics sometimes proved negative for human well-being. Innovation resulted in new uses of mercury simultaneously with new mercury-free techniques. The supply of mercury to the ASGM sector continues, along with its harms to human health and the environment, despite domestic and international interventions to limit its flow into mining communities. Conversely, an absence of adaptation was also sometimes beneficial for human well-being: a resistance from international organizations, health authorities, and doctors to eliminating mercury use in vaccines has ensured the continuing provision of life-saving health services to many people.
The low ability of interactions involving environmental components to adapt to human-induced pressures related to toxic substances can have particularly negative consequences for human well-being. The sometimes very long timescales of interactions among environmental components can, at times, make it difficult for humans to modify these interactions, complicating both shorter-term and longer-term efforts to promote sustainability. Concentrations of mercury in the atmosphere, land, and oceans are relatively resistant to decreases due to the extended timescales in mercury’s global biogeochemical cycle. This makes it more difficult for human efforts to significantly reduce environmental concentrations of mercury by simply addressing current and future emissions and releases. The inability of some aquatic ecosystems to absorb mercury deposition without converting it into methylmercury is a further sign of the low adaptive capacity of environmental components that has negative consequences for the health of both animals and humans.
Interveners often try to enhance the resilience of a system. The fact that a high degree of adaptive capacity can lead to mixed outcomes in terms of human well-being, as we have observed in the mercury systems, makes efforts to enhance resilience a challenge. Encouraging positive adaptation while mitigating its negative elements requires careful and targeted intervention, and the recognition that some components are not easily changed. Where human activities have long-term consequences for environmental components, efforts to mitigate harms to human well-being through other means become important. These can include efforts to reduce human exposure through dietary advice or preventing the harvesting of seafood from certain bodies of water. Such measures are sometimes still partially ineffective—fish-eating birds do not follow dietary recommendations, no matter how sound they are. This suggests that although focusing on overall resilience can provide useful guidance, different aspects of systems may require very different types of interventions to promote sustainability.
Challenges for Planetary-Scale Systems Analysis
Some believe that we live in the epoch of the Anthropocene (Crutzen 2006), although the usefulness of this concept and its potential start date as a new epoch are debated (Malm and Hornborg 2014; Lewis and Maslin 2015; Zalasiewicz 2015). Geologists define epochs by identifying a globally visible “golden spike” or environmental marker in the geological record to determine their beginning. Some analysts argue that the Anthropocene began in the 1900s, around the time that fossil fuel use and consumption, as well as population levels, accelerated (Malhi 2017). Others believe that a much earlier start date is more appropriate, positing that human activities began to have global-scale impact on environmental processes far back in history. Some archaeological evidence suggests that global human-induced land-use change began 3,000 years ago (Stephens et al. 2019). In 2019, an expert working group under the International Commission on Stratigraphy, a scientific body of the International Union of Geological Sciences, voted in favor of establishing the Anthropocene as a formal geological epoch with a mid-twentieth-century start date. If this proposal is accepted by the International Commission on Stratigraphy, it will be forwarded to the Executive Committee of the International Union of Geological Sciences, which will make the final decision (Subramanian 2019).
The human influence across regional and global scales involving both environmental and societal cycling of mercury goes back at least to the Spanish colonial silver and gold mining in Latin America starting in the 1500s. This is consistent with an argument that a New–Old World collision that began in 1492—which resulted in a large human population shift, the expansion of trading networks linking Europe, China, Africa, and the Americas, and a mixing of previously separate biotas—is a critical point in the history of human impact on the planet (Lewis and Maslin 2015). Mercury was a central element in this often environmentally and socially destructive process, due to its substantial trade and importance in gold and silver mining. The global transport of mercury escalated at that time, although the ultimate environmental fate of this mercury remains contested. Measurements of mercury in environmental archives, as discussed in chapter 3, do not all show a globally identifiable “golden spike” during that time. Even if the environmental distribution of mercury remained largely local, however, societal flows of mercury increased with the advent of transatlantic trade, as much mercury from Almadén and Idrija was shipped to the Americas.
Defining the Anthropocene based only on the geological record does not account for the full scope and influence of societal processes: environmental archives and geological records alone cannot identify all global-scale anthropogenic influences that are driven by interactions between humans and technology. Mercury has been traded across societies for millennia, going back at least to Roman times when cinnabar and mercury extracted from mines in Almadén was an important commodity. Evidence of this is found in written records of how much mercury different mines produced and in logs of mercury trades. Defining when dynamic global-scale interactions characteristic of the Anthropocene began, and identifying lessons for present-day systems-oriented sustainability analysis, thus requires consideration of societal flows in addition to environmental data. The concept of the Anthropocene is used not only by geologists but also by a broader group of researchers who may define and use the term differently from those who consider it only as a marker of global-scale influence measurable everywhere in environmental compartments. Scholars applying the Anthropocene idea to sustainability, especially those who define sustainability with a focus on human well-being, should therefore carefully define what they mean by the term and how it applies to their analysis.
Researchers have developed the concept of planetary boundaries to assess the ultimate limits of the Earth system based on the identification of global-scale environmental markers. They argue that going beyond these planetary-scale boundaries can have detrimental, or even catastrophic, consequences for humanity’s ability to continue to live in a relatively stable environment (Rockström 2009; Rockström et al. 2009). Scientists have also attempted to quantify the outer limits of specific planetary boundaries for critical Earth-system processes; for example, identifying a range of “safe” atmospheric carbon dioxide concentration levels that should not be exceeded to avoid catastrophic climate change. The idea of a planetary boundary is envisioned to ensure that society stays a safe distance from dangerous thresholds and tipping points in Earth systems. Some experts argue that some critical boundaries have already been reached or breached, in particular in the areas of biodiversity loss, climate change, and nitrogen cycling (Steffen et al. 2015b). Nevertheless, the concept of planetary boundaries is contested (Nordhaus et al. 2012).
Applying the concept of planetary-scale boundaries is challenging for toxic substances like mercury. This is grounded in the fact that pollution involves a large number of substances from multiple sources, in different places, with different toxicities, and in varying amounts (Diamond et al. 2015). Johan Rockström and colleagues (2009), however, argue that there are two possible approaches to identifying planetary boundaries for chemical substances. One is to focus on substances that have global environmental distribution, and the other is to focus on those that pose unacceptable long-term and large-scale effects on living organisms. The latter approach may require a combination of efforts: setting a range of sub-boundaries based on the effects of many individual substances, and identifying specific effects on sensitive organisms. When smaller-scale processes occur simultaneously in different places around the globe, their impacts may aggregate to a point that a common planetary threshold or tipping point is crossed. Matthew MacLeod and colleagues (2014) further argue that three conditions must be met for chemical pollution to pose a planetary boundary threat: (1) it has a disruptive effect on a vital Earth system process; (2) the disruptive effect is not discovered until it is a problem, or inevitably will become one, at a planetary scale; and (3) the disruptive effect is poorly reversible.
Defining a global boundary for mercury and related chemical pollutants is difficult, however. It may also turn out to be both impossible and undesirable in the context of sustainability. First, there is no clear way of empirically identifying a single planetary boundary for a substance such as mercury that is dispersed across the entire planet by both environmental and societal processes at highly varying concentrations and forms in air, oceans, and land. This variability also makes it difficult to determine from environmental data alone when a particular threshold has been crossed at global scale. Second, much harm to human health and the environment from bioaccumulative toxic substances like methylmercury can be highly locally concentrated, yet nevertheless is affected by global forces. Mercury discharges affect nearby as well as distant places, with major differences in human impacts that depend on local levels of mercury use and discharges, ecosystem processes, and seafood consumption patterns. In addition, individual variations in genetic characteristics are increasingly understood to affect the sensitivity of individuals to mercury and other toxic exposures. These factors combine to make the link between severe local harms from mercury and any global boundary tenuous at best.
Sustainability is both a contested concept and an urgent planetary goal. How sustainability is defined has important implications for analysis, as it influences which specific factors are seen as essential to sustainability, in both theory and practice. The choice of definition in turn affects analyses of important drivers and factors that hinder or promote transitions toward greater sustainability. We identify three overall insights concerning sustainability definitions and transitions. First, different values attributed to benefits and risks complicate analysis of human well-being across populations and over time, challenging efforts to define sustainability. Second, many transitions toward sustainability were characterized by incremental change, but some nevertheless had substantial benefits for human well-being. Finally, different patterns and modes of transitions that had interacting dynamics occurred simultaneously.
Sustainability and Human Well-Being
We elected in our sustainability-oriented analysis of the mercury systems to focus on human well-being, rather than another factor, such as the maintenance of particular natural resource stocks at certain levels over time—a perspective that would have been more aligned with a “stronger” sustainability definition. Our choice of an anthropocentric focus on sustainability shaped our analysis by centering it on factors and processes that improve or detract from human well-being. Applying a stronger sustainability criterion of natural resource maintenance to the mercury systems would have been conceptually challenging given mercury’s extensive use over history. Analysts have disagreed about the extent to which sustainability requires non-renewable resources to be maintained (and not depleted), and the degree to which different resources are able to substitute for one another (Ayres 2007). We allowed for the possibility that one resource can substitute for another in providing for human well-being, and we empirically examined where and when this was possible. The mercury systems provide several examples of how different people, in different places and at different times, have perceived these substitutions and associated trade-offs.
Insights from the mercury systems show that preventing extraction of a non-renewable resource like mercury does not always benefit contemporaneous and future human well-being. While mercury led to many harms, uses of mercury provided many tangible benefits to people all over the world for millennia. In more recent times, mercury’s unique properties facilitated innovation, including improving measurement precision and enabling the development of fundamental scientific knowledge that paved the way for new technology. The use of the mercury thermometer, mercury amalgam in dentistry, and very small doses of mercury as a preservative in life-saving vaccines made it possible to provide better health care for billions of people over many generations. Mercury use in chlor-alkali production enabled the manufacturing of chlorine to rapidly expand, and resulting societal benefits included disease prevention by chlorinating drinking water to make it safer for human consumption. Modern uses of mercury in compact fluorescent bulbs helped reduce energy demand from fossil fuels, and even reduced net mercury emissions in some places by decreasing the demand for electricity produced by coal.
The concept of a circular economy, which focuses on the recovery and recycling of materials, is gaining increased attention from analysts and policy-makers (Ghisellini et al. 2016). The establishment of a more circular economy can help better manage flows of substances, but implications of these flows on sustainability can be mixed and complex. Some analysts consider a circular economy a necessary condition for sustainability, others assess it as beneficial but not required, and still others argue that it has both pros and cons from a sustainability perspective (Geissdoerfer et al. 2017). Among those analysts who have critiqued the circular economy perspective, Julian M. Allwood (2014) points out that implementing a circular economy for materials can increase energy demand for recovery and recycling, potentially resulting in decreased overall sustainability depending on the impacts of the sources of this energy. This underscores the importance of not just studying societal and environmental flows of a particular substance such as mercury in isolation, but also considering how those flows affect, and are affected by, the flows of other substances.
Examples from the mercury systems highlight the complexity of assessing the impact of material flows. Ensuring sustainable use of materials can defy simple solutions (Olivetti and Cullen 2018). A growing number of countries are adopting supply restrictions on mercury to curb its use and trade, including the adoption of the US and EU export bans on elemental mercury in the 2000s. However, this has resulted in an increase in illegal mercury mining together with the smuggling of mercury into local ASGM communities in many countries. This mercury is sold on the black market and subsequently discharged into the environment from amalgamation processes and accidental spills. Yet not all uses and environmental discharges of mercury may be net negative for human well-being. Mercury in a vaccine enters the human body and is then excreted into the environment. Most analysts would argue that the benefits of safe and effective vaccines for human well-being far exceed the harms of this miniscule “loss” of mercury dissipated from human bodies into the environment.
A better understanding of the long-term challenges stemming from early and more recent uses of mercury could nevertheless have prevented some past, present, and future harms to human health. Assessing the impacts of mercury use on human well-being requires understanding the complexity of how it affects humans over time and space. The perceived benefits and harms of mercury use changed dramatically over time as a result of new scientific and medical knowledge and societal norms. Mercury as a commodity, once a valued asset, is now increasingly seen as a liability. Even today, though, mercury remains a valuable commercial product, not least because of its extensive use in ASGM. At the same time, public authorities and firms have spent billions of US dollars on cleanup where mercury has contaminated local environments, and more money will go toward addressing contaminated sites in the future. Much is also spent on health care costs related to mercury exposure as well as public information campaigns about the dangers of mercury in food and the formulation of dietary recommendations. Estimates of the monetized harms of mercury pollution to health can far exceed expected costs of mitigating emissions (Giang and Selin 2016; Sunderland et al. 2016).
Accounting for the value of assets that might become liabilities in the future is a major challenge for efforts to define sustainability. Following the literature on inclusive wealth, an analyst from the not-so-distant past might have valued mercury positively as a depletable asset and regarded underlying stocks of mercury as contributing to human well-being. Mercury, despite its contribution to environmental damage, still maintains a positive commodity price and is traded both legally and illegally. But through most of the history of its human mobilization, the commercial stock of mercury represented a yet unknown future claim on human well-being. This is not just a matter for history: a recent version of the World Bank’s Changing Wealth of Nations report, detailing benchmarks on comprehensive wealth, includes the value of mineral reserves for lead, another toxic heavy metal (Lange et al. 2018). A similar argument can be made today for including reserves of coal and oil in such accounting. Of currently identified reserves, a third of oil, half of gas, and more than 80 percent of coal need to remain unused if global average temperature increases are to stay within a 2-degree Celsius target (McGlade and Ekins 2015).
The existence of trade-offs in benefits and costs of mercury uses also reinforces the importance of considering equity and power when defining and analyzing sustainability. The way benefits and harms are valued is contingent on what, and who, matters, and to whom. Many Arctic indigenous communities where people consume large quantities of seafood are shouldering the human costs of being affected by methylmercury that originates from distant economic activities—and are receiving few of the related benefits. Responses to contamination incidents such as those in Minamata, Grassy Narrows, and Kodaikanal show that mercury exists in the context of other, competing economic values, such as the worth of an industrial manufacturing plant to a local community and factory owners. The lack of value that Spanish colonial rulers in South America gave to the human lives of indigenous peoples allowed for mining, and its health and environmental hazards, to persist for centuries. In contemporary ASGM, the use of mercury adds to other challenges present in poor communities where many miners and other community members operate in the informal sector, vulnerable to exploitation.
Incremental versus Fundamental Transitions
Scholars disagree about the importance of incremental relative to fundamental transitions toward sustainability, with some analysts putting more faith in incremental steps than others. The development and stepwise application of more environmentally friendly technology is central to moving toward greater sustainability in the literature on ecological modernization (Mol 2003). However, critics argue that the concept of ecological modernization relies too heavily on the future promise of technological salvation while not fully recognizing the necessary scope and depth of behavioral change (Gibbs 2006). Some research on sustainability transformations instead emphasizes that radical changes are necessary: Derk Loorbach and colleagues (2017) note that major transitions are not merely technological, but are accompanied by socio-cultural change that has a deep effect on institutions, routines, and beliefs. Technological and behavioral changes are, of course, not mutually exclusive, but can take place in tandem. Yet, perspectives that emphasize incremental solutions often focus more on technological improvements, whereas those who argue that fundamental change is needed often stress the need for behavioral changes.
Transitions in the mercury systems show that incremental changes can have substantial benefits for the well-being of both present and future generations. The gradual implementation of end-of-pipe control technology on large industrial point sources, such as on waste incinerators and coal-fired power plants in Europe and North America, mitigated local harms by cutting mercury emissions, and relatively quickly reduced exposure levels to nearby populations. These controls also prevented the mobilization of mercury that would otherwise have contributed to future mercury cycling in the environment for decades to centuries. Similarly, the application of retorts in ASGM that reduce mercury emissions and exposure has large benefits for the health of miners, people working in gold processing shops, and other community members as well as for nearby and faraway consumers of seafood. The introduction of increasingly stringent worker protection measures in other mercury-using sectors also prevented much health damage. In addition, piecemeal efforts that reduced levels of mercury use in industrial processes lessened mercury releases to the environment and their short-term and long-term damage.
The introduction of new technology had multifaceted implications for sustainability transitions in the mercury systems, with both positive and negative elements. Compact fluorescent bulbs are an improvement over the incandescent bulb with respect to energy efficiency and, in many cases, associated mercury emissions from power generation. However, they still contain mercury. Increased technological efficiency of production processes, such as in the European chlor-alkali plants that still relied on the mercury-based production method, reduced mercury use and discharges but delayed progress toward fully eliminating mercury use by extending the lifetime of the old production method. The expanded use of retorts on local mining sites has clear benefits to human health and the environment, but lessens the urgency to phase out mercury use in ASGM. The development of more efficient pollution control technology, capturing mercury that would otherwise be emitted to air from industrial point sources, significantly reduces the amount of mercury that enters the environment. This technology, however, allows for the building of additional coal-fired power plants as long as any mandated pollution controls are implemented, contributing to climate change.
The mercury systems provide further evidence that a substitution for a known hazardous substance can lessen one kind of damage, but may create new problems. For example, the environmental and human health implications of discharges from the production and use of membranes made of PFAS, toxic and long-lasting substances that substituted for mercury-based technology in chlor-alkali plants, are unclear. This was also the case in the past when early substitutes for DDT were less persistent but more acutely toxic (Walker et al. 2003). Furthermore, some substitutes for ozone depleting substances are potent greenhouse gases (Wallington et al. 1994). The toxicity of different heavy metals that are needed to make alternative sources to fossil fuel energy (such as photovoltaics and batteries) may (or may not) be less than that of mercury, and their environmental and human health effects locally and globally are at least partly unknown. The use of such heavy metals in energy technologies also changes where and for how long they are present in the atmosphere, land, and water. This suggests that analyses of benefits and costs to human well-being should be approached with humility given the limitations of scientific knowledge.
Our analysis of the mercury systems also shows that incremental changes can add up to support more fundamental change. The gradual phaseout of mercury in medicine over the last 100 years profoundly altered how mercury is used in medical treatments; the main remaining uses are in dentistry and vaccines. The stepwise controls of mercury in products facilitated a global ban of the production and sale of most currently existing mercury-added products under the Minamata Convention. The strengthening over time of mandates for air pollution control technology on large stationary sources in some instances both reduced mercury emissions and impacted fossil fuel production. The US Mercury and Air Toxics Standards and the Canadian mercury regulations likely contributed to the early closure of old coal-fired power plants, where it did not make economic sense for the owners to install new mercury-control technology. This interplay between stepwise and fundamental change suggests the need for more critical analysis of positive and negative consequences of incremental versus comprehensive approaches to transitions. In this context, some analysts have called for “incremental change with a transformative agenda,” or “radical incrementalism” (Najam and Selin 2011, 453).
The fact that fundamental change can sometimes come from cumulative incremental steps makes it difficult to empirically distinguish between different degrees of transitions. Several kinds of transitions that had aspects of both incremental change and fundamental disruptions occurred in the mercury systems. Because technological change involving many mercury-added products and mercury-based production processes were not obviously associated with large-scale societal changes, some analysts may classify these as incremental changes. But a closer look reveals that there were substantive institutional developments in markets and domestic and international regulations as part of these changes. Consumers and governments put pressure on companies to reduce mercury use in products, and regional and global agreements codified maximum allowable thresholds for mercury concentrations in these products. There were also major changes in scientific knowledge and societal perceptions about the environmental and human health risks from mercury. Analysts must thus look at a combination of technological, economic, political, legal, social, and knowledge factors when examining the scope and depth of a change process.
Drivers of Transition Dynamics
Analysts are discussing the uniqueness of the drivers and dynamics of sustainability transitions compared to other types of societal transitions. Frank W. Geels (2011) posits that sustainability transitions differ from many other types of previous socio-technical transitions. However, Björn-Ola Linnér and Victoria Wibeck (2019) argue from a historical and comparative perspective that contemporary efforts to move toward sustainability can learn from other major changes that societies have undergone in the past, including the Industrial Revolution and the abolition of slavery. Transitions that are relevant to sustainability are deeply social and political processes that often are characterized not only by cooperation, but also by contestation and confrontation (Avelino et al. 2016). Systems can change both slowly and abruptly, but many system transitions related to sustainability are characterized by strong path-dependencies and lock-ins, as established practices are intertwined with individual choices and lifestyles, cultural traditions, business models and economic systems, technology, regulations and other policies, and organizational and political structures (Markard et al. 2012).
Transitions in the mercury systems reflected both economic and technical drivers (e.g., the development of cheaper and better ways to make large mirrors) and scientific advances (e.g., penicillin replacing previous remedies). The transition literature on niches, regimes, and landscapes helps analyze the multi-scale nature of these processes (Geels 2002). Many technical innovations that allowed for the phaseout of mercury use in products and processes occurred in niches that supported experimentation. For example, mercury-free technologies for chlor-alkali production emerged in different places in different times, sometimes from an interplay between environmental factors (such as the availability of brine wells) and national regulatory regimes. Later mercury phaseouts were driven by interactions between innovation processes and government-led regulatory measures focused on protecting the environment and human health. The landscape of the Global Mercury Partnership and the Minamata Convention accelerated some of these phaseouts. The prevalence of economic and technical drivers suggests that increased attention to socio-technical transition dynamics is useful in sustainability analysis.
Distributions of agency and power, sometimes embedded in institutions, can play a substantive role in driving transitions. Many employers had a much greater influence on the use of mercury than did their employees, in both mining and manufacturing. Similarly, doctors who prescribed mercury-based treatments were in a position of power over their patients. This gave employers and doctors, and later also governments, a dominating role in driving transitions away from these mercury uses. There are also instances where private sector actors successfully fought to delay or stop government actions, including for setting maximum allowable mercury concentrations in commercial fish, banning phenylmercury compounds in paints, and the introduction of pollution controls on stationary sources. All ASGM is locally concentrated, but many choices by potential and active miners (including those related to the use of mercury) are influenced by domestic laws on mining rights and by international markets for gold and mercury, over which individual miners have very little influence. These examples highlight the importance of assessing power relations among different actors in transition analyses.
Much research on sustainability is motivated by the goal of identifying opportunities for more effective interventions that benefit human well-being. The challenge of governance, involving institutions at a variety of spatial scales, is in itself a complex system problem. We identify three main ways in which the many efforts to govern mercury-related issues over centuries illuminate how governance for sustainability could become more effective. First, ensuring that institutions fit the physical problems that they are designed to address involves paying close attention to material system components and their interactions. Second, governance strategies can look to address multiple sustainability issues simultaneously through institutional design. Third, evaluating institutional effectiveness requires simultaneously considering environmental and societal factors that shape outcomes.
Problem Structure and Institutional Fit
Many scholars argue that polycentric governance structures are often a better institutional fit with complex environmental and sustainability issues than more monocentric and traditional top-down governance structures (Victor and Raustiala 2004; Ostrom 2010; Keohane and Victor 2011; Abbott 2012). Polycentric governance structures involve a large number of actors and policy instruments that are not organized in a strictly hierarchical way, and governance happens in a large number of forums at multiple geographical scales simultaneously. This polycentricism can create dynamic opportunities to address different aspects of large and complex issues in separate places at the same time, which can add up to better overall governance than attempting to accomplish all goals in the same venue. A polycentric governance approach also allows for greater policy experimentation, learning, and diffusion across venues and jurisdictions (Hoffman 2011). It is, however, important that governance efforts in different forums and across different geographical scales do not inhibit or contradict each other, but rather create synergistic benefits (Selin 2010).
Our analysis of the mercury systems shows the benefits of designing a polycentric governance approach across global, regional, national, and local scales. Efforts to limit the risks of methylmercury exposure resulting from the consumption of seafood benefit from targeted, locally specific advice, but much seafood is traded on regional and global markets, requiring some coordination across governance scales. The effectiveness of pollution-control technologies for capturing mercury emissions from industrial point sources depends on technical characteristics that may be specific to an individual plant. The Global Mercury Partnership and cooperation under the Minamata Convention facilitates the development and diffusion of pollution standards globally. The Minamata Convention sets out ambitious provisions for the phaseout of mercury in products and processes; the implementation and enforcement of these depend on actions at the national and local level. Addressing mercury use in ASGM involves a combination of community-level capacity building, changes to national mining laws, and efforts to change the behavior of transnational gold market participants.
The governance literature tends to focus on political, economic, and technical factors when characterizing the structure of environmental problems. From this perspective, environmental problems have been classified on a spectrum from “benign” (relatively easy to address) to “malign” (much harder to solve) (Miles et al. 2002). Other problem categories include “wicked” and “super-wicked” problems (Levin et al. 2012; Head and Alford 2015; Grundmann 2016). Super-wicked problems have been defined as comprising four key governance features: (1) time is running out; (2) those who cause the problem also seek to provide a solution; (3) the central authority needed to address it is weak or nonexistent; and (4) policy responses discount the future irrationally (Levin et al. 2012). These types of characterizations of problem structure, however, are much too general and simplified to fully capture the dynamics of many sustainability issues, including the dangers of mercury to human well-being, and to inform any deeper understanding around interactions and interventions.
It is important that a governance approach to an individual sustainability issue take into consideration its unique biophysical as well as societal characteristics. A key aspect of many chemicals issues is how substances cycle through the environment and society. But there are important differences even between chemical substances, which means that individual substances need their own detailed description of biogeochemical cycling and of societal uses and flows. Factors that influence governance can also vary among different aspects of a single substance, which we see across the mercury systems. Governments can, if they have the political support to do so, set pollution controls on a smaller number of major stationary sources relatively readily (as effective control technology already exists). It is much harder to address the activities of tens of millions of ASGM miners and community members operating in the informal sector. Designing institutions thus requires a detailed engagement with different aspects of what might seem to be a single issue.
The Minamata Convention attempts to engage different aspects of the mercury issue by setting out a global-scale legal framework for action on the full lifecycle of mercury (Selin et al. 2018). The negotiations and early implementation of the Minamata Convention drew political attention to mercury, and triggered some countries to take administrative and regulatory actions that they otherwise would not have taken (or at least not as early as they did). International law, including multilateral treaties such as the Minamata Convention, however, is only a “thirty-percent solution” to environmental problems (Bodansky 2010, 15). The general language in several Minamata Convention articles creates a need for further specificity, such as the formulation of national emission standards and national action plans on ASGM. Initiatives outside the scope of the Minamata Convention such as international certification schemes and codes are also important governance instruments. Stakeholders range from individual consumers of skin-lightening creams to ASGM miners and gold market participants to operators of fossil fuel–based energy infrastructure. This means that future governance will require efforts by a large number of public, private, and civil society actors, in a multitude of global, regional, national, and local forums, to realize the objective of protecting human health and the environment from mercury.
Governance Strategies
The literature on governance strategies for complex systems suggests that the identification of components and interactions that are located at different places within a system can help guide efforts toward effecting system change (Meadows 1999; Abson et al. 2017). These efforts include attempts by interveners to intentionally move systems toward greater sustainability. Many early efforts to address air and water pollution of mercury and other hazardous substances focused on ways to prevent emissions and releases through the application of end-of-pipe technologies. In contrast, a growing number of policy initiatives over the past few decades emphasizes the importance of interventions further upstream through modifications to underlying production processes that produce pollution (Browner 1993; Mayer et al. 2002). Upstream approaches are particularly influenced by discussions about the precautionary principle that focuses on preventing harm before it occurs, even when scientific knowledge about environmental and human health impacts of pollution may be incomplete or uncertain (Harremoës et al. 2013).
Insights from the mercury systems show that a combination of interventions is often necessary to enhance human well-being, for both present and future generations. Approaches such as recycling and environmentally sound disposal helped reduce mercury discharges. Actions further upstream that prevent mercury discharges completely by eliminating all mercury uses also remain critical to reduce mercury-related problems. However, the further upstream the intervention, the slower it may be to propagate through the system. People have altered the environmental cycling of mercury to such an extent that environmental concentrations will not return to natural levels on human-relevant timescales. Thus, while technology-based approaches that avoid generating pollution can help prevent further contamination, they will need to be paired with other actions to protect human well-being. This includes other kinds of downstream initiatives such as the issuing of global recommendations for maximum allowable mercury concentrations in food by the World Health Organization (WHO), together with more nationally and locally tailored dietary recommendations for specific populations and geographical locations.
Different actors with varying abilities to effect change engage in multiple ways in governance for sustainability. The mercury systems feature large and well-organized interests in energy generation and industrial production, and more diffuse groups such as consumers, medical patients, and ASGM miners. As in other issue areas, less powerful groups of actors often find it difficult to make their voices heard, as seen in Minamata, Grassy Narrows, Kodaikanal, and elsewhere. Many narratives surrounding pollution focus on populations who are on the receiving end and have no control over the processes that create it. The same is true in many cases for mercury, including for subsistence fishers and others who depend on local fish for important nutrients. At the same time, whale consumers in the Faroe Islands have much individual freedom to make their own dietary choices, and Arctic indigenous peoples collaborate in transnational networks to shape regional and global institutions, including the Minamata Convention. This is consistent with studies that show that victims of pollution may have opportunities to both individually and collectively exercise power and influence (Fernández-Llamazares et al. 2020).
It is sometimes possible to design interventions on one sustainability issue so that they positively affect another one as well. The application of pollution-prevention technology on industrial point sources can address sulfur, particulate, and mercury emissions simultaneously, having multiple benefits for environmental quality and human health. Reducing dependency on fossil fuels addresses both mercury and carbon dioxide emissions; slowing down temperature increases limits the negative impacts of climate change and may lead to less remobilization of legacy mercury from environmental storage. The same strategy for addressing land use change might benefit biodiversity, prevent runoff of nitrogen and hazardous substances into waterways, and create conditions that reduce the formation of methylmercury in aquatic ecosystems. Increasing resources for climate change adaptation can help prevent conditions that drive people to become ASGM miners, whose actions lead to mercury discharges. All of these examples illustrate the importance for analysts (and decision-makers) to consider opportunities for enhancing cross-sectoral synergies when designing governance structures and strategies to support a move toward greater sustainability.
At the same time, it is important to recognize that efforts to enhance governance synergies across sustainability issues can have drawbacks as well. Global cooperation on mercury may not have resulted in a legally binding treaty if solutions to mercury emissions were linked to restrictions on coal burning (necessary for addressing climate change). Enhancing linkages between the Minamata Convention and the Basel, Rotterdam, and Stockholm Conventions has been a topic of much discussion. Supporters argue that it would create political and administrative synergies, but skeptics fear that it may result in reduced attention to the Minamata Convention. Analyses of the efforts to link the Basel, Rotterdam, and Stockholm Conventions show that such efforts can facilitate policy-making and implementation as well as transfer political disagreements from one forum to another (Selin 2010; Allan et al. 2018). This suggests that efforts to craft “win-win” solutions to sustainability governance should consider political as well as material connections among issues. Less-coordinated actions can have a greater influence on human well-being than carefully optimized strategies where decision-making and implementation are contested or incomplete. A realistic governance approach may include both a higher-level and a longer-term focus on sustainability transitions in combination with an “honest recognition of the realities of near-term incrementalism at the same time” (Patterson et al. 2017, 4).
Evaluating Effectiveness
Evaluating whether governance strategies enhance present and future human well-being is analytically challenging. Analyses of the effectiveness of international environmental treaties like the Minamata Convention that focus on institutional dynamics include qualitative case studies, quantitative case studies, and large-number quantitative analyses (Young 2017). A nuanced understanding of institutional fit and problem structure is needed not just when designing institutions (as we discussed earlier), but also when examining the effectiveness of environmental treaties. Paying more careful attention to the role of technical and biophysical systems, however, poses significant research design challenges (Young 2011). Accounting for problem structure raises the issue of endogeneity, in which an environmental treaty is influenced by problem structure and in turn aims to influence it (Mitchell 2006). This makes it analytically difficult to study environmental treaties—and their impacts and effectiveness—as a dependent variable. Many effectiveness evaluations also rely on indicators that do not capture the data necessary to trace key elements linking policy to environmental outcomes (Wöhrnschimmel et al. 2016).
Institutionally focused evaluations of treaty effectiveness can use combinations of outcome and process indicators (Selin et al. 2018). The use of outcome indicators, focusing on environmental conditions, requires having the scientific ability and data to document complex pathways of interactions. First, it is necessary to determine whether a particular intervention caused a change in behavior or operation. For example, was it mercury-specific regulations that caused the mercury emission reductions that were observed in the United States in the 2000s? Or were these reductions the result of other air pollution controls, or market forces that made the burning of coal less economically profitable? Finding the answer requires expertise in policy analysis and economics, coupled with a deep understanding of the technologies involved. Next, it must be determined whether the decrease in mercury emissions have resulted in changes in the amount of mercury entering nearby and remote ecosystems. Measuring deposition trends in the context of both short-range and long-range transport of emissions as well as ecosystem variability and environmental change is difficult, and attributing deposition trends requires integrating environmental measurements with models and counterfactual analysis.
The challenge of attributing environmental change to its causes is much discussed in the scientific literature (Deser et al. 2012; Selin et al. 2018). Time trends for mercury in biota may not track concurrent trends in atmospheric concentrations or deposition because of the strong influence of other drivers on conversion to methylmercury and uptake in biota (Wang et al. 2019). It is well appreciated that many interactions, especially in the environment, are complex and often poorly understood. Much knowledge is still developing about the environmental behavior of different forms of mercury. For example, uncertainty remains about the chemical reactions that mercury undergoes in the atmosphere, the degree to which it is taken up by land ecosystems, and the timescales of its cycling in the environment (Selin 2009). Even more uncertainty surrounds links between environmental mercury cycling and climate and other large-scale environmental changes (Obrist et al. 2018). In addition, there is a growing appreciation that mercury, its behavior, and its impacts are variable in time and space (Eagles-Smith et al. 2018; Giang et al. 2018; Hsu-Kim et al. 2018).
The use of process indicators offers an additional way to evaluate treaty effectiveness. Process indicators provide a way to gather complementary information to the typically incomplete environmental and human health data that are generated through application of outcome indicators. For process indicators, the focus is not on whether a specific intervention has had a direct impact on mercury concentrations in fish in a particular body of water, for instance, or in a specific group of people, but rather if practical steps that are likely to have had a positive impact have been taken. In the earlier example of emissions in the United States in the 2000s, it can be assumed that measures such as the closure of old point sources that burn coal and/or the greater application of emission control technology will result in the avoidance of at least some mercury emissions that otherwise would have entered the atmosphere. This does not come nearly as close to tracking the effects from specific policy measures as using outcome indicators, but the latter may not be scientifically feasible at present. Process indicators can thus provide a rough approximation with much more straightforward analysis.
David C. Evers and colleagues (2016) suggest a combination of outcome and process indicators to evaluate the effectiveness of the Minamata Convention. Their evaluation framework consists of specific metrics for measuring changes in mercury levels that are derived from environmental monitoring, including of hot spots (outcome indicators), as well as a suite of short-, medium-, and long-term metrics that are related to the control articles in the Minamata Convention (process indicators). The use of process indicators also makes it possible to highlight multiple aspects of international cooperation ranging from the number of countries that have taken regulatory measures to the levels of financial flows in support of treaty implementation. Work on the first Minamata Convention effectiveness evaluation, scheduled for 2023, is examining a variety of outcome and process indicators. Hybrid approaches that associate process indicators with proxies for impact can also be developed. Based on experiences from the Montreal Protocol on stratospheric ozone depletion, it has been suggested that indicators estimating the impact of present-day mercury emissions on ecosystems in the future could help policy-makers better account for the timescales of the mercury problem when considering different policy options (Selin 2018).
The fact that planet Earth is under much (and growing) pressure from human activities makes sustainability an urgent challenge. For those researchers who engage in systems analysis, insights from the mercury systems underscore the need to apply multidisciplinary perspectives when considering human, technological, environmental, institutional, and knowledge components together, to account for adaptation in understanding sustainability-relevant systems, and to critically consider the applicability of concepts like the Anthropocene and planetary boundaries. For those who aim to advance definitions of sustainability and who explore sustainability transitions, human interactions with mercury over time and space offer fodder for thinking about definitions of sustainability and the dynamics of incremental change versus fundamental disruptions. For those who focus on governance challenges related to sustainability, analysis of the mercury systems shows how environmental and societal factors together affect institutional design, and how efforts to evaluate institutional effectiveness require paying attention to both types of factors.