Climate Change, Climate Engineering, and the ‘Global Poor’: What Does Justice Require?

ABSTRACT

In recent work, Joshua Horton and David Keith argue on distributive and consequentialist grounds that research into solar radiation management (SRM) geoengineering is justified because the resulting knowledge has the potential to benefit everyone, particularly the ‘global poor.’ I argue that this view overlooks procedural and recognitional justice, and thus relegates to the background questions of how SRM research should be governed. In response to Horton and Keith, I argue for a multidimensional approach to geoengineering justice, which entails that questions of how to govern SRM research should be addressed from the very outset – that is, now.

KEYWORDS:

• Geoengineering
• climate engineering
• climate ethics
• recognition
• procedural justice

‘It’s an engineering problem, and it has engineering solutions.’

-Rex Tillerson, speaking on climate change at the Council on Foreign Relations, CEO speakers series, 27 June 2012¹

‘Does geoengineering raise any ethical issues not already considered by historical figures such as Aristotle, Hume, Kant, and so on? Isn’t the ethics of making decisions that affect others not involved in making the decisions a problem as old as humanity? I just don’t understand how there is anything new here for philosophy…’

-Stanford scientist Ken Caldeira, Geoengineering Google group, April 2012²

‘Our government’s first duty is to its people, to our citizens – to serve their needs, to ensure their safety, to preserve their rights and to defend their values. As President of the United States, I will always put America first, just like you, as the leaders of your countries will always, and should always, put your countries first.’

-Donald Trump, Speech to the United Nations General Assembly, September 19, 2017³

Introduction

A number of prominent proponents of research into solar radiation management (SRM) geoengineering argue on consequentialist grounds that such research is justified because the resulting knowledge has the potential to benefit everyone, particularly the least advantaged. Josh Horton and David Keith, for example, argue that ‘a prima facie moral obligation exists to research SRM in the interest of developing countries, because SRM appears to be the most effective and practicable option available to alleviate a range of near-term climate damages that are certain to hurt the global South most of all’ (Horton & Keith, 2016, p. 89). Horton and Keith provide a consequentialist, distributive justice argument for SRM research: research is justified – or even morally required – in order to save ‘the global poor.’

Horton and Keith acknowledge that the obligation is defeasible. However, they offer little reason to think the prima facie moral obligation they assert might be overridden by countervailing considerations. In this paper, instead of accepting the prima facie obligation and considering concerns that might override it,⁴ I want to challenge the very starting point of Horton and Keith’s argument and its implicit suggestion that the decision to pursue research (or not) can be grounded primarily or exclusively in considerations of distributive justice. In the course of the paper, I will ‘trouble’ their position (in the sense of the Latin, turbidus, to make turbid – or to complicate)⁵ and in so doing, show why I find the position troubling (in the worrisome sense) and offer an alternative approach.

My core concern is this: arguments such as Horton and Keith’s often relegate to the background questions of how SRM research should be governed, implicitly suggesting that the decision to pursue research should happen first, followed by decisions about how to govern it. In my view, however, governance needs to begin at the beginning. Arguments such as Horton and Keith’s, which focus on distributive and consequentialist considerations, should be considered not prior to, but in tandem with, concerns for participatory, recognitional, and multigenerational justice. This, in turn, entails that questions about how to govern SRM research – including questions of who decides whether such research should be pursued and how it should be pursued – should be addressed from the very outset – that is, now.⁶

The Limits of the Distributive Frame: Multidimensional Climate Justice and SRM

Before turning to Horton and Keith’s argument specifically, I want to situate distributive justice within a broader multi-dimensional frame. I will argue for a multi-dimensional conception of justice – which incorporates procedural, participatory, and recognition justice – and show how the multi-dimensional view can broaden discussions of distributive justice in relation to SRM, moving beyond a narrow, consequentialist frame.

Although both academic discussions of climate justice and international climate negotiations through the UNFCCC process have emphasized distributive justice, the discourse around justice has long encompassed more than distribution alone. Some of the earliest accounts of justice in the Western philosophical tradition, for example, extend beyond distribution: Plato describes justice as a form of harmony within and among individuals (see, e.g., LeBar & Slote, 2016), and Aristotle discusses both distributive and rectifactory dimensions of justice. More recent accounts of environmental justice include both distributive and participatory dimensions (Figueroa & Mills, 2001; Shrader-Frechette, 2002), as well as additional dimensions such as recognition, intergenerational justice, ecological justice and capabilities (e.g., Figueroa, 2011; Nussbaum, 2009; Schlosberg, 2007; Whyte, 2011). Nevertheless, despite promising recent work arguing for a multidimensional approach to climate justice (see Kortetmäki, 2016), the ethics of climate mitigation, adaptation, and geoengineering remains centered on distribution. Who wins? Who loses? Who pays?

I will argue that this approach is inadequate, for multiple reasons. Some of these reasons reflect general concerns with conceptions of justice that focus only on distribution; others are more particular to the climate case, and to SRM specifically. Beginning with the general concerns, let’s take a very simple example commonly used to conceptualize distributive justice, the pie case, where one is asked to consider how to fairly divide up a metaphorical pie among various parties. The pie case tends to provoke the intuition that when all else is equal, the pie should be divided equally. If you and I are sharing a pie, for example, it seems fair that we each get half. Of course, all else is rarely equal: for example, where did the pie come from, and who baked it? If I grew pumpkins for the pie and made it, then you stole it from me, it seems doubtful that justice will be achieved if you simply give half of it back. Context matters – and it often matters a lot – in assessing what sort of distribution would be fair. So one problem with the pie case, described in very abstract terms, is that it omits context, and what might at first seem like obvious distributive principles can turn out to be not so obvious after all, once the context is filled in. However, there is another problem with the pie metaphor: by focusing attention on the size of the slices in a fixed pie, the metaphor can distract from further questions about who decides and through what process distributive decisions are made.

One might reply that as long as the principles of distribution are fair, then the process doesn’t matter. As long as the pie is fairly divided and everyone gets their share, it doesn’t really matter who cuts and doles out the slices. Perhaps there are some cases in which it doesn’t matter who divides the pie. For instance, if it were always fair to divide the pie equally among the relevant parties, and everyone agreed on this, then perhaps concerns about procedural and participatory justice would be less salient. The trouble is, in many complex situations of justice, there is no ex ante general consensus on which distributive principle(s) to use. Even in very simple, stylized cases, distributive principles can be controversial. And in most real-world cases, complex contextual considerations bear on the choice of relevant distributive principles. In these cases, ex ante agreement on the appropriate distributive principle(s) is unlikely, and certainly can’t be assumed. But without consensus on what would be distributively fair, who has the authority to assert the applicability of a particular principle to a particular case? I want to suggest that in cases where the relevant distributive considerations are complex and controversial, it is morally problematic for an individual or a small group to select a distributive principle or determine a distributive outcome without taking seriously the need for participatory engagement. At minimum, then, it seems that where ex ante agreement is lacking, distributive justice cannot be achieved without a participatory component.

However, participatory justice also matters in its own right.⁷ One way to ground the idea of participatory justice is in the ideal of participatory parity, which requires that people have the opportunity to engage as ‘full partners in social interaction,’ equal in status and standing to other participants (Fraser & Honneth, 2003, p. 36). There are multiple ways to operationalize this ideal from a participatory perspective, but one of the most basic approaches is through the all-affected principle: people should have opportunities to participate in decisions that significantly affect them, or put another way, ‘all those affected by a political decision ought, directly or indirectly, to have a say in its making’ (Näsström, 2011, p. 115).⁸ Although some have objected to this principle on grounds that it is potentially too inclusive, Goodin (2007) argues that it is nevertheless ‘the standard by which the adequacy of [other variations of the principle] is invariably assessed.’ The all-affected principle thus represents an important starting point for participatory justice, and has significant implications for global scale issues like climate change. Adopting the all-affected principle in the case of SRM research and development would extend the community to which obligations of participatory justice apply to all people, given the technology’s intended global reach.⁹

This line of argument shows that distributive considerations alone cannot be determinative in relation to SRM. Instead, arguments from distributive justice should be offered as fodder within a broader, more inclusive conversation about whether and how SRM research and development should take place, a conversation that gives voice to the perspectives of all those who would be affected by SRM, that is, to the perspectives of everyone.

The ideal of participatory parity, along with the associated all-affected principle, not only provides grounding for obligations of participatory justice; it also entails the need for justice as recognition. To enable people to engage ‘full partners in social interaction’ requires not only opportunities to participate in a formal sense in decisions that affect them; it also requires that people have similar status and standing as participants. Recognition justice focuses on these issues of status and standing, rejecting social arrangements and governance structures that silence or denigrate the perspectives of those in particular cultural groups, or with particular gender, racial, or class identities – or in the case of SRM, those who currently lack dominant voices in geoengineering debates. This requires respect – including respect for difference – and mutual acknowledgement. The importance of recognition appears clearly in the space of overlap with participatory justice: meaningful participation requires not just formal opportunities to offer one’s views, but interlocutors, institutions, and processes that take seriously those perspectives.¹⁰

As Kyle Whyte (2011) explains, ‘Recognition justice requires that policies and programs must meet the standard of fairly considering and representing the cultures, values, and situations of all affected parties.’ This entails that the structure of participatory processes as well as the terms of the debate over SRM must take seriously the cultures, values, and situations of all involved. Recognition requires that processes, deliberations, and decisions about whether and how to proceed with SRM research and development engage all of the relevant perspectives without privileging a single normative perspective or worldview. Thus (to provide just one example), if non-consequentialist considerations matter to people in relation to SRM, any decision-making process that presupposes consequentialism will fail to provide adequate recognition by marginalizing perspectives that fall outside the consequentialist frame.

The view of justice just described – which incorporates distributive, participatory, and recognitive dimensions – has been called the trivalent model of justice (see, e.g., Kortetmäki, 2017; Schlosberg, 2007). The trivalent model not only incorporates three elements of justice that have value in their own right; it is a model in which each element supports the others. For example, distributive justice supports participatory justice and recognition by providing adequate resources to enable full participation and engagement of participants on an equal footing: significant disparities in knowledge, money, and power limit access to participatory processes (cf. Fraser & Honneth, 2003, p. 36). Conversely, participatory justice and recognition can play an important role in distributive justice by challenging certain assumptions about distribution and introducing a wider range of considerations and possibilities (Hourdequin, 2018). For example, robust, inclusive participatory processes may provoke important questions such as the following:

• What is being distributed? (Which pie is being divided?) Have all the relevant distributive dimensions been taken into account?

• Is a particular approach to remedying distributive injustice the only or best one? What alternatives exist?

• What principles of distribution are being presupposed?

• How are historical injustices taken into account? For example, do proposed distributive schemes or policies take history into account, and if so, how?

• How are future generations being considered?

• How will distributive decisions and frameworks be assessed and revisited over time?

• Are the values at stake fungible, as distributive approaches often presuppose?

• At what scale(s) are distributive frameworks being developed? What scale(s) are being overlooked?

Many of these questions are relevant to the case of SRM, and to the position articulated by Horton and Keith. For example, it might be the case that SRM could alleviate some of the anticipated harms of climate change for some of the world’s most vulnerable people in the near term. However, participatory processes that take recognition seriously might raise the question of whether these are the only relevant harms. Perhaps some of the vulnerable people in question are also worried about the uneven distribution of power and control over the earth’s climate, and see tradeoffs between alleviating physical/climatic harms through SRM and potentially exacerbating maldistributions of power through the development of this technology. Attention to participatory justice and recognition might also provoke a robust discussion about how research and development of SRM could affect future generations, and how best to address intergenerational justice. Thus, the trivalent model of justice can lead to fuller consideration of distributive questions in all their richness and complexity, and it can also make space for ethical concerns related to SRM that distributive arguments alone might overlook.

I have argued elsewhere for the importance of a trivalent model of justice, citing a number of reasons to pay particular attention to recognition:

“[T]aking recognition seriously would encourage deeper dialogue about the principles of distributive justice and about what climate policy aims to distribute. It might prompt more careful thought about how climate change interacts with diverse cultures and ways of life, such that climate losses and vulnerabilities would not be reduced to economic losses and vulnerabilities [see McShane (2017), Preston (2017), Hourdequin (2015) (unpublished ms.)]. Additionally, it would require greater engagement with the perspectives and needs of those who lack significant political, economic, and cultural power.” (Hourdequin, 2016, p. 37)

Participatory justice and recognition are important to climate mitigation and adaptation, but they play a particularly critical role in relation to SRM, for at least four reasons (Hourdequin, 2016): 1) SRM is global in scope and scale (in contrast to adaptation, which can occur on various scales); 2) SRM poses significant risks in comparison to mitigation (these risks include regional climate effects, but also geopolitical risks such as rogue or uncoordinated geoengineering); 3) SRM would represent a significant shift from past practices by exerting intentional, ongoing control over the climate system at the global scale; and 4) there is no consensus on the acceptability or desirability of SRM research, development, and deployment, and no settled mechanism for achieving an agreement on how to proceed. No international policy framework governs SRM research and planning, and both research on and discussions of SRM thus far have been relatively ad hoc and highly concentrated within a very small group of actors from wealthy countries such as the United States, England, and Germany.

For all of these reasons, the trivalent model of justice is critical to ethical discussions of SRM research and development, and to the establishment of governance frameworks for research. Distributive justice arguments for geoengineering research need to be embedded in a broader conversation that takes seriously issues of representation and recognition.¹¹

Some Troubles with SRM to Save the Global Poor

‘[O]pposition to research on SRM threatens to violate principles of justice by effectively condemning developing countries to suffer the consequences of activities of which they have not been the primary beneficiaries.’ – (Horton & Keith, 2016, p. 80)

The considerations offered above focus on the general insufficiency of arguments based on distributive justice alone to settle ethical questions regarding SRM research, development, and deployment. In this section, I want to return to the specific argument given by Joshua Horton and David Keith. After briefly summarizing their argument, I describe two key concerns – paternalism and parochialism – and describe potential links to the broader problem of expertise imperialism. I then suggest that a multidimensional approach to justice offers resources that can mitigate these concerns.

Before describing my concerns in more detail, it will be helpful to lay out Horton and Keith’s argument, as I understand it. Here is a basic outline:

1. Global climate change will disproportionately burden the global poor. This is because some of the most severe impacts of climate change are falling and will continue to fall on poorer regions and countries, and because the people in those areas have the fewest resources to adapt to climate change.

2. Global climate change is primarily the result of fossil fuel emissions that have primarily benefited the rich.

3. Thus, the benefits associated with the use of fossil fuels have accrued primarily to the wealthy while the costs are being borne primarily by the poor.

4. This is unfair, because ‘intuition tells us that the requirements of justice are violated when an activity benefits wealthy countries at the expense of poorer ones’ (Horton & Keith, 2016, p. 80).

5. The unfairness of global climate change generates an obligation to ‘take steps to reduce harms falling on the most vulnerable nations’ (80). There are three potential paths for reducing climate harms: mitigation (emissions reductions and/or carbon dioxide removal), adaptation, and SRM.

6. Mitigation won’t effectively reduce climate harms in the near term (81). Additionally, the near-term costs of mitigating will be borne more heavily by the poor ‘for whom energy is a larger fraction of expenses’ (81). So mitigation may actually further burden those who are already least advantaged.

7. Adaptation can reduce climate harms, but it works locally (not globally) and is more expensive than SRM. Because the benefits of adaptation are local, it is local actors who have the ‘strongest incentives for action’ but ‘the poor…will have fewer resources’ for adaptation than the rich. Adaptation’s ‘theoretical potential to help the poor most of all is undermined by the practical consequences of its costly, local character’ (82).

8. SRM is preferable to adaptation as a way of addressing near-term harms to the least advantaged, because its benefits are global in scale, and it is cheaper than adaptation. As Horton and Keith explain, ‘Since a given unit of climate protection would benefit the poor disproportionately, the cost differential between adaptation and SRM imparts a comparative greater redistributive potential to the latter response option’ (83).

9. SRM is the ‘most effective and practicable option to alleviate a range of near-term climate damages that are certain to hurt the global South most of all’ (89).

10. Therefore, there is an obligation ‘to research SRM in the interest of developing countries’ (89).

With this outline of the argument in view, I now turn to the concerns, the first of which is paternalism.

Paternalism

‘Now, with these new technologies that evolve always come a lot of questions. Ours is an industry that is built on technology, it’s built on science, it’s built on engineering, and because we have a society that by and large is illiterate in these areas, science, math and engineering, what we do is a mystery to them and they find it scary. And because of that, it creates easy opportunities for opponents of development, activist organizations, to manufacture fear.’

–Rex Tillerson (speaking on concerns regarding hydraulic fracturing), Council on Foreign Relations, CEO speakers series, 27 June 2012, https://www.cfr.org/event/ceo-speaker-series-conversation-rex-w-tillerson

Paternalism is a concept that has been central to discussions in bioethics, where questions arise about when and whether it is appropriate for physicians to make decisions on a patient’s behalf, and relatedly, when and whether it is appropriate to withhold information from patients, on grounds that certain things are better for them not to know. Sjöstrand, Eriksson, Juth, and Helgesson (2013) characterize paternalism like this: ‘Paternalism refers to courses of action (including decisions) that are done in the assumed interest of a person, but without or against that person’s informed consent.’ Gerald Dworkin offers a similar characterization, arguing that a person or institution acts paternalistically towards another when their act interferes with the other’s autonomy, is done without consent, and is done with the intent of improving the welfare or advancing the interests of the other¹² (Dworkin 2017, section 2). By adding the ‘interference with autonomy’ condition, Dworkin’s analysis has the virtue of circumscribing the scope of paternalism such that buying a gift for a friend or picking up a piece of trash from a neighbor’s front lawn would not fall under its scope.

Horton and Keith’s argument for SRM research seems to fit quite neatly within Sjöstrand et al.’s model of paternalism. The intent to advance others’ interests is clear: Horton and Keith argue that SRM research ought to proceed in order to benefit ‘the global poor,’ and further, that there exists an obligation to conduct SRM research on their behalf (implicitly, it seems that this obligation is one they expect to be borne by researchers in wealthy countries¹³). However, Horton and Keith don’t describe this obligation as contingent on the consent of those they seek to benefit: as noted above, they nowhere mention any data or consultation process that support the idea that the ‘the global poor’ share the view that SRM research is the best way to advance their interests in relation to climate change. By assuming it is possible to know that SRM research is in fact the best way to advance the interests of those in poorer parts of the world with respect to climate change, Horton and Keith make a paternalistic argument, insofar as paternalism involves acting to benefit another without their consent.

Horton and Keith’s position also seems to fit Dworkin’s more restrictive characterization of paternalism, with its ‘interference with autonomy’ condition. First, unless inclusively designed and governed, SRM research centered in wealthy countries (as it is now) will exacerbate power imbalances between powerful, wealthy countries and less powerful, less wealthy ones. The power to intentionally manipulate the global climate through the deployment of SRM is not inconsiderable, and SRM research thus far has amassed that power primarily within North America and Western Europe. While Horton and Keith might argue that they can sidestep concerns about interference with autonomy by advocating only for research, this is far from clear. In the case of nuclear weapons, for example, it is not only the actual deployment of these weapons that shapes the behavior of other nations, but the possession of the power to deploy them. The credible capacity of some countries to use nuclear weapons restricts the behavior of other countries; similarly, the developed capacity for wealthy nations to engage in global-scale climate engineering would constrain the actions and positions of less wealthy and less powerful countries, introducing another complex dimension into climate negotiations and another set of SRM-related justice concerns. Additionally, although Horton and Keith ostensibly argue only for an obligation to research SRM, this obligation gets its force directly from the assumed benefits of actually using SRM to address near-term climate harms. In other words, Horton and Keith seem to be arguing that we (in wealthy countries) should conduct SRM research so that we (in wealthy countries) can deploy SRM to help the global poor. They offer no reasons to think that SRM research alone will benefit the poor; without the link to deployment, it is not clear how SRM research discharges the obligation to ‘take steps to reduce harms falling on the most vulnerable nations’ (p. 80).¹⁴ And even if research on SRM does not interfere with the autonomy of the global poor (which I have offered reasons to doubt), it is hard to see how even benevolent deployment of SRM by wealthy nations would avoid such interference.

Taking paternalism seriously highlights that even if SRM research were the best way to advance the interests of ‘the global poor,’ this doesn’t entail that proceeding with such research without their consent or participation would be just (Jamieson, 1996). On the trivalent model of justice described above, arguments from distributive justice don’t override the need for participatory justice and recognition.¹⁵ Thus, from the perspective of multidimensional justice, paternalistic arguments for SRM are ethically problematic insofar as they attempt to circumvent the need for participatory processes and assume that the development of SRM can be justified based on assumed benefits to those most heavily burdened by the physical effects of climate change.

Although I want to acknowledge the limits my capacity to understand the diverse and context-dependent perspectives of the many people that Horton and Keith categorize as the ‘global poor,’ it seems that the following simple allegory might capture some aspects of the current situation in relation to SRM, seen in the context of the history of climate change more broadly¹⁶:

• I take actions that benefit me and (incidentally) harm you. I benefit a lot from these actions, which vastly increase my power and wealth.

• Later, I argue that you should not take similar actions to benefit yourself, because they might harm me, or others.

• At the very least, I say, if I am going to curtail my actions you should curtail yours, even though your actions are modest in comparison to mine and you’ve only recently started acting in the way that I’ve been acting for a long time.

• I continue to act in the ways that benefit me in the short term while harming you and others, and I determine that it is too expensive and inconvenient for me to stop doing so.

• Harms to you continue to accumulate, and your situation is becoming increasingly dire.

• I could provide resources to you that would enable you to better cope with your situation, but doing so would be expensive (though not so expensive as to significantly compromise way of life, which is luxurious compared to yours, or my power, which is vast compared to yours).

• So I argue that I should investigate a technology that I believe would benefit us both, alleviating some of the harm to you in the near term – though it might leave our grandchildren worse off.

• The technology is risky, but I am confident that my research will reveal the relevant risks, that I’ll be able to control them, and that my research will be used to advance your well being most of all.

• Without consulting you, I argue that I have a moral obligation to investigate this technology, and that not investigating it would be wrong.

• I acknowledge the need to think about whether and how to use the technology I’m developing, but insist that we don’t really need to work out those details now, because I’m just doing research, not actually advocating use of the technology.

• I embark on research (again, without consulting you, or at most, consulting you in a perfunctory way) and argue that it’s ok to start experimenting with the technology in the real world, because the experiments I’m proposing don’t really count as testing or using the technology (though the distinctions between research, testing, and deploying the technology are controversial).¹⁷

Recent empirical, social scientific research in climate-vulnerable regions supports the plausibility of this allegorical description in relation to climate engineering. In a study exploring the perspectives of people from the South Pacific, North American Arctic, and Sub-Saharan Africa, Wylie Carr and Christopher Preston found that although many of the interviewees did not categorically oppose geoengineering research or its consideration as one possible response to climate change, their ‘overarching concern…[was]…that climate engineering could further erode the already weakened self-determination of vulnerable populations due to a long history of oppression’ (2017, p. 764, emphasis added). Interviewees not only expressed concern that they would be further disempowered by climate engineering, they worried that geoengineering research would not account for the climate variables most important to their lives and livelihoods (p. 767); that unintended and uncompensated impacts would fall on those already most vulnerable to climate change (p. 767); and that the costs of geoengineering research and deployment would be used by ‘affluent technologically advanced societies [to] justify their control over the technologies’ (p. 768). As one interviewee put it, ‘It’s scary as hell to be dependent on some other person to dictate the weather or climate change’ (Carr and Preston, 2017, 770).

In the absence of a major reorientation that takes seriously these concerns, I submit that conducting SRM research in the current context is paternalistic, and not benignly so.

Cultural Parochialism

Because paternalism makes assumptions on others’ behalf, it risks parochialism. Of course, all of us are positioned in ways that limit our view, and any argument is bound to reflect particular conceptions of the world. The trouble lies in a lack of reflexivity, and in assuming that one’s own views are widely or universally shared (or ought to be so). It is in dialogue with others that we can come to recognize alternative perspectives, and this dialogue is particularly important for positions that prescribe courses of action – such as research on SRM – that are likely to have significant consequences for the world as a whole: for present and future generations, for people with diverse values and ways of life, for animals, plants, and ecosystems.

There are two particular aspects of Horton and Keith’s argument that I believe reflect parochial assumptions that deserve further scrutiny. First, Horton and Keith assume a consequentialist framework in which risks and benefits can be traded off against one another.¹⁸ For example, they assess mitigation, adaptation, and SRM in terms of their effects on climate risk and climate protection, focusing on the physical effects of climate change in various regions. They favor SRM over the other options because it is (they argue) the most cost effective way of reducing climate risk in the near future.¹⁹ The language, throughout, reflects a consequentialist, cost-benefit frame. Although this frame is widely deployed throughout the world, it is arguably a culturally parochial one that encodes certain dominant Western presuppositions about the nature and fungibility of value.²⁰

A second aspect of Horton and Keith’s argument that embeds parochial assumptions has to do with what they take as ‘fixed’ and what they see as changeable. For example, Horton and Keith argue that in the near-term, the costs of both mitigation and adaptation will fall disproportionately on the poor, and therefore SRM – which is cheaper and whose (modest) costs are more likely to be borne by those in wealthier parts of the world – is a more just response to the near-term impacts of climate change. They further note that ‘local actors pursuing local interests through the use of SRM might, if the intervention was properly designed, benefit the rest of the world (especially the global poor) as a virtual by-product of their otherwise self-interested use of solar geoengineering’ (Horton & Keith, 2016, p. 83). There’s a fair bit to unpack here. Clearly, the costs of adaptation and mitigation don’t necessarily have to fall on the world’s poorest people: what Horton and Keith take as fixed need not be so. If nations such as the United States truly cared about the well being of ‘the global poor,’ they could transfer resources to poorer countries burdened with the costs of adaptation or could shoulder a larger proportion of the global burdens of mitigation. Horton and Keith seem to assume that neither is likely to happen, thus SRM is a better option. But under a scenario in which the wealthy countries care too little about the poor to do their fair share with respect to mitigation and adaptation, how likely is it that research and development of SRM will prioritize the interests of ‘the global poor’? And how confident do ‘the global poor’²¹ feel about the optimistic prospect that they will be the primary beneficiaries of an SRM program centered in and controlled by wealthy nations ‘pursuing [their own] local interests’?

A broad and inclusive discussion of SRM research could illuminate the degree to which the assumptions and framings offered by Horton and Keith are shared by others throughout the world, and particularly by those their position purports to benefit.

Expertise Imperialism

The worries described above would be less troubling if Horton and Keith located their argument as one perspective among many on the complex and contested question of whether and how to pursue research on SRM. However, Horton and Keith instead present their argument as determinative, insisting that those who disagree and oppose SRM research are ‘effectively condemning developing countries to suffer the consequences of activities of which they have not been the primary beneficiaries’ (Horton & Keith, 2016, p. 80). They not only criticize the narrowness of perspective of their opponents – which seems to me a legitimate line of critique – they also question their opponents’ ‘deeper motives’ (p. 90) and suggest that these motives eclipse concerns for those who are least well off. They claim that many of the opponents of SRM research are ‘rich-country commentators criticizing solar geoengineering in an effort to shore up mitigation…while ignoring the potentially huge distributional advantages SRM might confer on the world’s poorest in the global South,’ and this, they hold, is ‘ethically disturbing’ (Horton & Keith, 2016, p. 90).

Yet it’s not clear what gives Horton and Keith the epistemic position from which to understand their opponents’ motives, and it seems a bit odd to hang both their arguments in favor SRM research and against those who oppose it on the importance of prioritizing the interests and needs of the least advantaged without consulting those same people about what their interests and needs really are. Despite arguing that the primary reason to engage in SRM research is to help the ‘global poor,’ Horton and Keith nowhere explicitly discuss the possibility of consulting or involving the ‘global poor’ or their representatives.²² If, indeed, wealthy and privileged opponents of SRM research are resting their arguments on the suffering of those from other, less wealthy parts of the world, a robust conversation that includes those others may be the best way to bring this out. It might also expose limitations or lacunae in the arguments of Horton and Keith.

Paternalism and parochialism in discourse around SRM may reflect and be bound up with what Allen Buchanan calls expertise imperialism. Buchanan (2002) describes expertise imperialism as ‘the tendency of experts to appeal to their genuine expertise in one area to justify their exercise of control in areas to which their expertise is in fact irrelevant.’ Buchanan (2002) offers as an example the phenomenon of medical paternalism, which he suggests is enabled by various institutional factors, including the power and privilege associated with physicians’ ‘elite status,’ which tend to ‘insulate [them] from criticism’ and undermine their own capacities for self-criticism (p. 133). However, Buchanan also notes that the phenomenon extends beyond the medical realm:

“…not just medical professionals, but socially recognized experts generally, exhibit a combination of characteristics pregnant with the possibility of self-serving bias. On the one hand, the members of such groups have a common interest in sustaining and, if possible, expanding the privileges that they enjoy; this corporate interest can be in conflict with the interests of other individuals and of the community at large. On the other hand, the privileges that experts enjoy tend to insulate them from external criticisms that might serve as checks on the tendency to indulge in self-serving rationalizations.” (Buchanan, 2002, p. 134)

It is difficult to judge whether and when ‘self-serving rationalizations’ are at work in any particular case, and I don’t think it would be reasonable or fair to accuse Horton and Keith of expertise imperialism based on their article, ‘Solar Geoengineering and Obligations to the Global Poor.’ However, the structure of their argument is more confrontational than invitational, and seems to focus more on boxing out objectors to SRM research than on engaging seriously with their concerns.²³ Other scientists working on SRM have displayed a similar tendency. Take, for example, one prominent scientist’s remark on the Google Geoengineering Groups listserv (also quoted above):

“Does geoengineering raise any ethical issues not already considered by historical figures such as Aristotle, Hume, Kant, and so on? Isn’t the ethics of making decisions that affect others not involved in making the decisions a problem as old as humanity?…In my conception, philosophers… develop theory and we all, when faced with moral problems, attempt to apply this theory in our moral reasoning. I drew the analogy with mathematics, where mathematicians develop theory and we all apply this theory when buying groceries….To me, a moral philosopher of geoengineering is like a mathematician of the grocery checkout line.”²⁴

The suggestion here is that the necessary ethical tools and considerations are already widely available and agreed upon, thus engaging philosophers in these debates is like having a theoretical mathematician supervising the clerks in the grocery checkout line: clearly unnecessary. Notice that not only does this scientist assume that it is straightforward to apply philosophical theories to the climate engineering context, he also seems to assume that canonical Western theories are sufficient to cover the ethical concerns, and that ‘the ethics of making decisions that affect others not involved in making the decisions’ is something that’s been fully worked out. Moreover, the entire framework here is a technocratic one: philosophical experts (though perhaps only long-dead ones) develop ethical theories that they pass on to scientific experts who develop technologies to which these theories can then be straightforwardly applied.

So one strategy for limiting ethical discussion and calls for broader engagement in SRM governance is to argue that the relevant ethical considerations are available and at hand, and scientists can easily apply them to the climate engineering case. A second strategy suggests that there’s nothing special or distinctive about SRM research at its current stage that warrants special attention or a special governance regime. The need for governance at a later stage is often acknowledged, though the criteria for determining when we will have arrived at the relevant stage are often unclear, and may shift. Note David Keith’s comments at the 2017 international Climate Engineering Conference in Berlin on Harvard University’s planned 2018 SRM field experiments:

‘There are things that it’s not. It’s not the first solar geoengineering experiment, there have been at least two…Second, it’s not a field test. To me, a test is something you do well down the engineering system, where there is a serious decision to deploy that system and you want to test that in a binary way. That’s not what we’re doing. What we’re doing is something much earlier, it’s really applied atmospheric science.’ (Quoted in Dunne, 2017, emphases added)²⁵

Is this expertise imperialism? Again, it’s not easy to make that judgment in any particular case, nor is it necessarily helpful. But at the broader level, there are patterns worthy of attention. Much of the ethical discourse surrounding SRM – particularly among those in the scientific community engaged in SRM-related research – seems to focus on enumerating various ethical concerns primarily in order to develop replies to them that clear the way for further research and/or to provide reassurance that the ethics issues are covered, and there’s no need for outsiders to worry, or intervene. Insofar as expertise imperialism has the potential to arise in any community of experts who are insulated from criticism, it is perhaps more important to take steps to guard against it than to diagnose it in any particular case. Such steps might address institutional structures, individual virtues, and their intersection, and could reveal ways in which existing institutional structures facilitate or hinder criticism and reflexivity within discussions of SRM.²⁶

Conclusion: Multidimensional Justice and Governance – Developing Institutions and Cultures of Recognition

Paternalism, parochialism, and expertise imperialism are salient not only in relation to one particular position or argument, but in relation to discussions of SRM research and development more broadly. A multidimensional approach to justice, grounded in recognition and participatory parity, can mitigate against these troubles. Institutions that aim to instantiate recognition and participatory parity can directly combat insular or self-serving tendencies by self-consciously aiming to counteract the inequities that lead a few ‘experts’ to have outsized roles in guiding decisions involving normative concerns that deserve broader public deliberation.²⁷

Taking seriously concerns about paternalism, parochialism, and expertise imperialism requires taking seriously their roles in knowledge production. Like all research, SRM research is not ‘neutral,’ and decisions about whether and how to pursue SRM research will affect not only the likelihood that SRM is deployed, but how it is developed, what concerns are taken into account, and what concerns are overlooked. SRM research raises questions of epistemic power and how asymmetries in knowledge and expertise can generate and exacerbate status inequality that in turn undermines recognition and participatory parity. Concerns of distributive justice in relation to SRM thus extend beyond the distribution of harms and benefits in relation to climate effects; they also encompass questions about the distribution of epistemic power, and relatedly, the power to direct and control responses to climate change (for further discussion, see Carr and Preston, 2017, Whyte, 2012a, 2012b). As Kyle Whyte (2012b, p. 175) explains in an article discussing indigenous sovereignty and research on SRM, the push for ‘even early research represents an emerging crystallization of a commitment that will give some people greater capacity to impact the climate system…[T]his is an equity issue; it is also an issue of whose capabilities to dominate the environment and other people will be strengthened.’

A multidimensional perspective on climate justice suggests that the current concentration of power in relation to SRM research and development is untenable, and arguments from distributive justice do not preclude the need to address participatory justice and recognition, which can generate forms of accountability that the current system lacks. In imagining what multidimensional justice might mean for SRM research, it is critical that both the terms of participation and the structure of research be considered from the perspective of participatory parity. From this point of view, it seems unlikely that forms of participation in which researchers or governments merely ‘consult’ or survey members of the public about their views will be sufficient, especially if the terms of the information gathering are already set by the ‘experts.’ With respect to research, participatory parity will require addressing disparities in knowledge and power, and decentering research processes and institutions: if the justification for pursuing research on SRM is to benefit those who are most unfairly burdened by climate change, then why should SRM research be conducted primarily by those whose lived experience intersects relatively little with these burdens?

Multidimensional justice for SRM requires not only specific institutions and governance structures, but also certain practices and habits of mind among those engaged in geoengineering research, development, and governance. In this regard, I want to highlight the importance of cultivating cultures of recognition. As I have argued previously, cultures of recognition encompass ‘particular habits of interaction that allow for mutual respect, which in turn rest on the development of individual and institutional virtues’ (Hourdequin, 2018). Virtues that support recognition include patience, open-mindedness, care and concern, epistemic humility, and a commitment to sustaining relationships through disagreement. Recognition also requires self-awareness and receptivity, particularly on the part of the wealthy and powerful, who – as noted above – may be particular vulnerable to certain forms of insensitivity, though they often control the development of critical discourses, bodies of scientific knowledge, and governance systems. Multidimensional justice requires a multifaceted approach: the development of virtues, practices, and cultures of recognition can work synergistically with efforts to re-envision both research and governance institutions in ways that constructively respond not only to the direct harms of climate change, but to the marginalization of those harmed in critical decisions that shape their and others’ futures.

Acknowledgments

I am grateful to Stephen Gardiner, Augustin Fragnière, Alex Lenferna, and Thomas Ackerman, as well as participants at the fall 2017 University of Washington conference on Geoengineering, Political Legitimacy, and Justice for very helpful questions and comments on an earlier version of this paper. I would also like to thank students and faculty at University of Colorado-Boulder and University of Washington for discussion and feedback on additional presentations of this work.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. Available at: https://www.cfr.org/event/ceo-speaker-series-conversation-rex-w-tillerson [accessed 5 March 2018].

2. The quotes above are drawn from a thread with multiple posts and replies; they can be found at https://groups.google.com/d/topic/geoengineering/hECCEn5AG4E/discussion and https://groups.google.com/d/topic/geoengineering/TnO6Cce66fE/discussion [accessed 5 March 2018].

3. Available at: https://www.whitehouse.gov/the-press-office/2017/09/19/remarks-president-trump-72nd-session-united-nations-general-assembly [accessed 5 March 2018].

4. Though some of the concerns adduced here could also be relevant under this frame. It is worth noting also that the assumption that SRM is ‘the most effective and practicable option available to alleviate a range of near-term climate damages’ that would fall most heavily on the poor (Horton & Keith 2016, p. 89), on which the prima facie obligation is based, is itself subject to challenge on distributive and consequentialist grounds.

5. For further discussion, see Staying with the Trouble (Haraway, 2016, 1).

6. I am far from the first to advocate for wide-ranging, participatory engagement in geoengineering governance, including research governance. For other examples, see: Carr et al. (2013) (a paper on which David Keith is a co-author), Corner, Pidgeon, and Parkhill (2012), Carr, Yung, and Preston (2014). What this paper seeks to add is a critique of pro-research arguments that focus solely on consequentialist and/or distributive concerns, and an argument for a broader, multi-dimensional framework for geoengineering justice.

7. I recognize that I am making a normative argument for participatory justice here, the basis of which is contestable on the same grounds as I contested the legitimacy of an individual’s or small group’s choice of distributive principles without broader participatory engagement. My argument can thus be seen as one among many normative arguments that others might bring to the table. The same is true of the normative arguments made in this paper more generally. As Sophia Näsström (2011) explains with respect to the all-affected principle, discussed below, ‘When we appeal to the all-affected principle we are therefore not only doing political philosophy. We are also engaged in a conflict with other citizens about the future boundaries of democratic life.’

8. The all-affected principle, as I understand it, does not entail that every person be directly consulted on every decision that affects them: it permits delegation of authority to others to make decisions through various systems of representation.

9. The all-affected principle, as described here, is not the same as a requirement of consent – it is, instead, a principle requiring opportunities for those affected to substantively engage in decisions that affect them. Nevertheless, concerns that Stephen Gardiner (2013a, pp. 28–29) has raised regarding consent-based justifications for geoengineering might be raised in relation to the all-affected principle: if the principle is construed as requiring engagement or representation of present people only, then the principle excludes future generations and non-human beings, who also matter morally. To this concern, I offer two replies: 1) in my view, honoring the all-affected principle, if interpreted to require engagement or representation of present people, is an important starting point for participatory justice, but is not necessarily sufficient to justify SRM. As I argue throughout, multiple dimensions of justice should be considered. What’s more, there are normative considerations other than those of justice that are relevant to SRM; 2) it is worth considering how the all-affected principle might be elaborated so as to take account of future generations and non-human beings. If the all-affected principle were taken seriously for present persons, it seems likely that concerns for future generations and non-humans would enter into the dialogue more fully than they have to date. However, this is clearly contingent on present persons’ concerns for future generations and diverse forms of life, and a more robust instantiation of the principle would seek to represent directly these constituencies.

10. Although I focus here on the way in which recognition and participatory justice interact, recognition extends beyond the participatory realm. Recognition, as I understand it, is fundamentally relational, with respect as its normative core. Thus, recognition requires relations of respect and non-domination, and misrecognition occurs when these relations fail at the interpersonal level, as well as where institutions and structures exclude, marginalize, or ignore those they purportedly serve. Participatory justice plays an important role in correcting failures of recognition, but misrecognition occurs outside of contexts involving participatory justice. For instance, misrecognition manifests frequently in everyday interactions, particularly where certain stereotypes or unwarranted assumptions block respectful engagement with others. Claudia Rankine’s recent book, Citizen: An American Lyric, provides numerous examples.

11. I want to acknowledge other dimensions of justice not foregrounded by this paper. For example, a fourth dimension of the multidimensional justice framework might focus on the temporal axis. We might call this multigenerational justice, and consider under this dimension justice to future generations as well as ways of considering and responding to historical injustices in the present. This fourth dimension deserves treatment in a separate paper, which might also address the question of whether a multigenerational justice lens can sufficiently account for structural injustices in the present that grow out of historic patterns of marginalization and oppression. Questions of multigenerational justice might also be considered as involving the scope of justice, rather than a distinct dimension (I thank Augustin Fragnière for this point), but for the purposes of this paper, I leave this issue aside.

12. According to Dworkin’s analysis, a paternalistic act is one done ‘only because [the actor] believes [the act] will improve the welfare of [the other]…or in some way promote the interests, values, or good of [the other’ (Dworkin 2017, emphasis added). But the stipulation that the other’s interests or welfare be the sole motive for a paternalistic act may be too strong, as Dworkin himself acknowledges, so I have slightly loosened the third condition here.

13. See p. 88, where Horton and Keith argue that ‘to close off research into SRM is to shirk the Northern responsibility to address the full range of climate risks destined to affect the global South most of all.’

14. One could imagine an argument for SRM research to benefit the ‘global poor’ that was decoupled from deployment and preserved autonomy. Such an argument might emphasize that SRM research has the potential to provide a greater array of options for responding to climate change, but it seems that this argument would work only if SRM research expanded the range of options available to the most vulnerable.

15. On a related note, Stephen Gardiner (2013b) argues that acting beneficently (to benefit another) can be morally problematic when such actions impinge on that other’s rights. Breaking into your neighbor’s house in order to clean it is problematic, even if your neighbor would like a cleaner house and you leave the house in significantly better condition than that in which you found it.

16. Stephen Gardiner (2013c) also offers an allegory for thinking through the decision context for climate engineering. This is the case of ‘Wayne’s Folly,’ where a chronic philanderer takes steps to address the risk of spreading sexually transmitted disease (STD) to his wife and others by investing in a drug company working to improve treatments for STDs. Wayne’s response to his promiscuity, argues Gardiner, bears important resemblance to certain contemporary arguments for climate geoengineering: both involve moral schizophrenia, a divergence between reasons and motives for action, and more specifically, ‘creative myopia,’ where ‘an agent invokes a set of strong moral reasons to justify a given course of action, but this course of action is supported by these reasons only because the agent has ruled out a number of alternative courses of action strongly supported by the same reasons, and where this is due to motives she has that are less important, and are condemned by those reasons’ (Gardiner, 2013b, 19).

17. Note David Keith’s comments at the Berlin Conference (fall 2017) on Harvard’s planned SRM field experiments, (quoted below, p. 14), which he insists should be categorized not as ‘field testing’ but rather as ‘applied atmospheric science.’ See: https://www.carbonbrief.org/geoengineering-scientists-berlin-debate-radicaly-ways-reverse-global-warming [accessed 5 March 2018].

18. For example: ‘The ethical implications of these responses [to climate change] turn on the particular distribution of benefits and harms associated with each’ (Horton & Keith 2016, p. 81).

19. With respect to mitigation, they argue that ‘in the near term, mitigation has significant costs compared to only modest benefits, and a disproportionate share of the costs may fall on the poor’ (p. 81), and in relation to adaptation, they explain, ‘Since a given unit of climate protection would benefit the poor disproportionately, the cost differential between adaptation and SRM imparts a comparatively greater redistributive potential to the latter response option’ (Horton & Keith 2016, p. 83).

20. What’s more, even if one accepts the cost-benefit frame, Horton and Keith’s conclusions about the cost effectiveness of SRM in comparison to other climate responses is controversial. For a succinct and helpful discussion of the costs of SRM, see (Lenferna, 2017). Lenferna points out, for example, that cost estimates that focus only on the costs of delivering aerosols to the stratosphere fail to consider the full panoply of costs associated with SRM.

21. Throughout the paper, I typically refer to ‘the global poor’ with scare quotes. Although it sometimes makes sense to speak in terms of broad categories (‘the global North,’ ‘the global South,’ or ‘the global poor’), it is important to recognize that these sweeping terms obscure vast diversity within these categories. So the question, as phrased above, ‘how confident do “the global poor” feel…?’, is a rhetorical one that mirrors the categories through which Horton and Keith frame their argument, though truly addressing the concerns of those who Horton and Keith see as falling into this broad category warrants a much more nuanced and fine-grained discussion.

22. Horton and Keith mention procedural justice very briefly at one point in the paper: this is in discussing the possibility of a global cost-benefit analysis (CBA) for SRM. After pointing out that some of the negative impacts of SRM on vulnerable regions might be offset by ‘gains attributable to other avoided damages from climate change,’ Horton and Keith acknowledge the difficulties in conducting a CBA for SRM, noting that ‘any utilitarian assessment would need to ensure robust procedural and substantive protections so as not to violate fundamental principles of justice.’ (Horton & Keith 2016, p. 86). However, they offer no further specification of what such ‘procedural protections’ would entail, and these protections are not highlighted in the context of their argument as a whole, but only in association with this brief discussion of the challenge of quantifying SRM’s costs and benefits.

23. For example, Horton and Keith repeatedly impugn the motives of their opponents, insisting that those who advocate against SRM research care more about their ‘rich-world political agendas’ than about the suffering of the poor (2016, p. 91).

24. Available at: https://groups.google.com/d/topic/geoengineering/hECCEn5AG4E/discussion [accessed 5 March 2018].

25. https://www.carbonbrief.org/geoengineering-scientists-berlin-debate-radicaly-ways-reverse-global-warming, emphases added.

26. In this regard, see Allen Buchanan’s (2002) work on social moral epistemology and José Medina’s writing on the ways in which the privileged are susceptible to the vice of meta-blindness, an ‘insensitivity to insensitivity’ which makes the privileged not only inattentive to the limits of their own perspectives, but also unaware of their own inattentiveness. Medina further describes meta-blindness as ‘a special difficulty in realizing and appreciating the limitations of [one’s] horizon of understanding’ (2013, p. 75).

27. I should note that moral philosophers are not exempted here – though in my view the greatest current risk is not that SRM policy is set by philosophers, but by a small group of technical experts working at the interface of geoengineering research and policy.