Engaged Climate Ethics*

WE are in the grip of a climate emergency. Recently, a letter signed by more than 11,000 scientists from 153 countries, published in the journal BioScience, stated frankly that ‘[a]n immense increase of scale in endeavors to conserve our biosphere is needed to avoid untold suffering due to the climate crisis’. It is in this profoundly disturbing context, at the dawn of what may perhaps be humanity’s most defining decade, that a degree of critical reflection on the role political philosophers can, and should, play in these endeavours seems apt. Over the last three decades, political philosophers/theorists and ethicists have attempted to grapple with the myriad conceptual and normative challenges posed by climate change. They have produced a rich and fascinating canon of scholarship that has pushed the boundaries of moral and political theory, bringing formerly marginal questions—about global duties, future persons, non-human species, diffusely caused harms, risk and uncertainty, for example—into the mainstream of philosophical theorizing. Climate change, in short, has done much for philosophy. But philosophy, we suggest, has not done much for climate change. That is, we doubt whether normative theorizing about climate change (hereafter ‘climate ethics’) has done much to positively influence real-world climate action. This, at least, is our motivating hunch.

Perhaps troubled by this failure, many theoretical interventions in climate ethics over the last decade or so have taken a 'non-ideal' turn. In this article, however, we argue that the standard approach to non-ideal climate ethics limits its capacity to have a positive influence on real-world climate action. Rather, we argue that engaged methods 4 -methods of doing political theory that involve substantial interaction between the theorist and an actual or potential agent of change, or participation by the theorist in such a group agent, where such interaction or participation influences the content of the theorist's normative constructs-are a superior means of achieving that objective.
To be clear, we are not suggesting that engaged climate ethics is the only way climate ethics could legitimately be done. We take no issue with using standard methods as a means to achieving progress towards philosophy's truthseeking objective. However, we believe many of our colleagues pursue climate ethics at least partly because they are concerned about the effects of climate change, and they believe, in a pragmatist vein, that climate ethics can have some influence, however marginal, on social responses to it. Accordingly, this group of philosophers, at least, should be interested in how climate ethics scholarship can best-or at least better-achieve that practical objective. Given the scale and urgency of the climate crisis, we think this is a valuable contribution that normative theorists can make in their capacity as scholars (as distinct from their capacity as concerned citizens); and our purpose here is to spearhead efforts in that direction.
The article is structured as follows. In Section I, we briefly outline some recent examples of non-ideal climate ethics. We identify among such examples the common utilization of a principles-first method of non-ideal theorizing. In Section II, we critically evaluate in detail one illustrative example of the principles-first method-a 2014 article in this journal by Simon Caney. 5 We discuss two major weaknesses of Caney's method qua means of effectively averting the climate emergency (Caney's self-described goal). In Section III, we argue that engaged methods are well suited to overcoming these weaknesses, offering a more promising path to real-world influence through political theorizing. We discuss three kinds of engaged methods-which we call ethnographic, activist, and committee-based engaged methods-about which there has been much recent discussion, often methodologically reflexive, in political theory more generally. We suggest some promising subjects to which aspiring engaged climate ethicists might apply these engaged methods. In Section IV, we respond to a set of 4 The term 'engaged political philosophy' has been used in a similar but distinct sense by Jonathan Wolff, 'Method in philosophy and public policy: applied philosophy versus engaged philosophy', Annabelle Lever and Andrei Poama (eds), The Routledge Handbook of Ethics and Public Policy (Oxford: Routledge, 2019), pp. 13-24. Wolff describes a method in which the philosopher engages with other kinds of experts to develop and evaluate public policy proposals. As discussed in Section III, we see this as but one type of engaged theorizing among (at least two) others. 5 Simon Caney, 'Two kinds of climate justice: avoiding harm and sharing burdens', Journal of Political Philosophy, 22 (2014), 124-49. objections that goes to the legitimacy of engaged methods, which might be levelled by proponents of the more standard, principles-first approach.

I. NON-IDEAL CLIMATE ETHICS: THE PRINCIPLES-FIRST METHOD
Seminal early contributions to climate ethics applied ideal, often cosmopolitan, principles of justice to develop accounts of climate-related rights and duties. 6 Climate change, on these accounts, was seen as a global and intergenerational problem best solved by implementing universal ethical principles. A central question addressed by climate ethicists was how to distribute emission rights between individuals in developed and developing countries, where these individuals were described as having made different contributions and having different capacities to reduce their emissions. 7 However, after the failure in Copenhagen dashed expectations of an ambitious global climate treaty, many climate ethicists 8 turned away from formulating global principles of climate justice and towards non-ideal theory. 9 Political reality seemed unresponsive to issues of equity and justice; it was 'radically non-ideal', as Aaron Maltais put it. 10 Such non-ideal theorizing typically involved taking widely accepted principles of climate justice and considering how they could be implemented under existing conditions.
Consider three examples. Dominic Roser analyses the current climate-political impasse as an injustice that follows from a lack of individual motivation. 11 He then argues that individuals should 'choose the least unjust option within the bounds of motivation, however insufficient motivation may currently be': 12 that is, actions that reduce greenhouse gas emissions without making the agent worse off, such as some energy efficiency measures. Second, Henry Shue argues that there is no need to sort out the fine details of principles of climate justice, because they will all point in the same direction in the short term: the initial costs of combating climate change should be borne by the rich developed countries and no one's basic needs should be jeopardized. He refers to these as 'minimum standards' and transitional guidelines, compliance with which would minimize injustices in dealing with climate change. 13 A third example is the two-step approach adopted by Alexandre Gajevic Sayegh, in which principles of climate justice are, first, interpreted in the light of empirical facts in order to determine which actions they prescribe in various real-life situations, and, secondly, reformulated to the extent that these implications are found counterintuitive. 14 Common to these non-ideal theories is the utilization of what we shall refer to as a principles-first method of non-ideal normative theorizing. 15 This method is motivated by a concern to think through what should be done to improve the real world. However, the assumption is that this requires recourse to a prior, 'ideal' theory of justice, which consists of general principles that determine duties and rights for a perfectly just society. The aim of ideal theorizing is to discover the true ideal theory of justice through the use of philosophical techniques designed to establish the impartiality or non-rejectability of the relevant normative principles. 16 The normativity of these principles is understood in categorical terms, as generating duties that have a universal, binding moral force. Moreover, these principles are typically specified and justified at a relatively high level of generality (befitting their universality). However, this high level of generality is not thought to strip justice claims of their real-world relevance, or 'action-guiding' potential. 17 Rather, it is precisely their general, universally binding nature and impartial justification that is thought to underwrite the special status and social power of justice claims in political discourse, rendering such claims critically potent and motivationally powerful. 13  , pp. 6-9, who uses the phrase 'ethics first' political philosophy. As with the reconstructions of these authors, our intention is not to caricature, but to present an accurate description of a common method in analytic political philosophy that many readers would recognize and subscribe to, at least in significant part. 16  It is the role of non-ideal theory, on this approach, to implement the principles prescribing these rights and duties in a manner that is thought, in light of empirical facts about constraints and opportunities, to be feasible and to bring society closer to the normative ideal. 18 But still, the recommendations of non-ideal theory are issued at a level of generality that leaves much undetermined. Most notably, they are typically issued to the citizenry (or human beings, or states) in general, with little to no attention given to the role of particular agents of change 19 and their epistemic standpoint, their position within power structures, or their unique constraints and opportunities. Less attention still is given to thinking through the relationship between the philosopher's theorizing and the practice of such agents. 20 We claim that this principles-first method of non-ideal theorizing has limited capacity to have a positive influence on real-world climate action. To substantiate this claim, in the next section we critically evaluate in detail another prominent example of non-ideal climate ethics that clearly adopts this principles-first method, that of Simon Caney in his 2014 article 'Two Kinds of Climate Justice: Avoiding Harm and Sharing Burdens'.
We focus on the method used in that article for three reasons. First, Caney is undoubtedly one of the most eminent contemporary climate ethicists, and the article was published in this journal, one of the top journals in the field. Second, Caney is explicit in that article that the goal of his theoretical scheme of responsibility-allocation is to effectively avert the climate emergency, which means there can be no doubt that he shares our objective of using theory to positively influence real-world climate action. And third, he is a particularly clear and explicit proponent of (what we are calling) the principles-first method, affording us a transparent target with which to assess the potential for that method to achieve its self-described objective. To be clear, the focus is on Caney's 18 On one variant of the principles-first method, the precise formulation (or 'interpretation') of the principle emerges from studying relevant empirical facts. This, for example, is how we understand Gajevic Sayegh, 'Justice in a non-ideal world', discussed above. 19 Unless the particular agents of change themselves happen to be among those who have duties of justice. On this point, see Laurence, 'The question of the agent of change', pp. 371-4; and Tommie Shelby, Dark Ghettos: Injustice, Dissent, and Reform (Cambridge, MA: Harvard University Press, 2016). 20 That is not to say that non-ideal theorists deny that more will need to be done in order to implement their theories. However, this is typically seen as requiring merely further empirical specification (about the institutional context, the subjective interests and relative power of relevant agents, etc.) to a sufficient degree to generate particular recommendations about what some specific agent or group should do now (e.g. what course of action an individual should take; what policy a government should adopt, etc.); the normative elements are assumed to be supplied by the earlier stages of theorizing, and ultimately by ideal theory. The need for, and methodology for undertaking, such additional steps is rarely specified, but see, 2014 article qua exemplar of the principles-first method widely adopted by climate ethicists, not qua exemplar of Caney's thought more generally. 21

II. CANEY'S PRINCIPLES-FIRST METHOD OF CLIMATE ETHICS: A CRITICAL EVALUATION
By way of exposition, the steps in Caney's method are as follows: 1. Caney begins with an analysis of the agents which-in accordance with ideal principles (which Caney specifies in other publications)-bear responsibilities 22 to mitigate climate change by reducing their greenhouse gas emissions, 23 which he calls 'first-order responsibilities'. 24 2. He then argues that when agents fail to discharge their first-order responsibilities (as is manifestly the case with respect to climate change), various 'second-order responsibilities' kick in. Second-order responsibilities are responsibilities to take actions that alter the structural (social, political, economic, and technological) contexts in which other agents make emissionsaffecting decisions. The function of second-order responsibilities is to ensure that, by so altering the structural context, other agents comply with their first-order responsibilities. 25 3. Caney proceeds to derive these second-order responsibilities by identifying the 'tasks' involved in so altering the structural context. 26 The tasks he identifies include 'enforcing', 'incentivizing', and 'enabling' (for example, through technological innovation) compliance by others with their firstorder responsibilities, as well as creating new social norms, undermining political resistance, engaging in civil disobedience, and enacting otherwise desirable demographic/population policies. 27 4. From the list of required tasks, Caney's next step is to infer the agents who have the 'capabilities' to efficaciously carry out those tasks. 28 It is these agents who bear second-order responsibilities. 29 5. Finally, Caney justifies the allocation of such second-order responsibilities to those capable agents. To do so, he appeals to the 'power/responsibility principle'-a capacity-based principle focused on second-order 21 Caney has grappled seriously with the non-ideal nature of climate change over many years, and different methods are used and espoused across his work, which spans both academic and policyoriented work. As we note in Section III, the latter encompasses engaged theorizing, too. 22 The notion of 'responsibility' invoked by Caney here, and assumed throughout our article, is the forward-looking kind; see Caney, 'Two kinds of climate justice', p. 127. 23 Caney also refers to responsibilities to enable climate adaptation and to compensate the victims of climate damages. For simplicity, we will refer only to mitigation (i.e. emissions reduction) responsibilities. 24 Caney, 'Two kinds of climate justice', p. 134. 25 Ibid., pp. 134-9. 26 Ibid., p. 135. 27 Ibid., pp. 136-9. 28 Ibid., pp. 139-40. 29 Or at least do so prima facie, subject to countervailing considerations; ibid., pp. 143-6.
responsibilities-and elucidates the conditions under which it applies. Those conditions are: (1) there exists an emergency; (2) certain agents can effectively contribute to its avoidance or mitigation; (3) those agents have a crucial and privileged causal role in so contributing; and (4) there are no sufficiently weighty countervailing considerations. Caney argues that these conditions are satisfied in the case of climate change. 30 It is evident from these steps that Caney endorses a principles-first approach to determining who should do what in order to effectively avert the climate emergency. This involves starting with an independently justified normative theory or principle, and working down to an allocation of second-order responsibilities via a causal analysis of the tasks necessary to ensure compliance with first-order responsibilities and the capabilities of agents to perform such tasks. Caney asserts that this is the 'most logical way' to proceed. 31 We claim that this method ultimately fares poorly relative to Caney's selfdescribed objective of averting the climate emergency (that is, positively influencing real-world change). Specifically, we raise two objections below.

A. Indeterminate Responsibilities: The Need for Local Normative Guidance
In this section we discuss our first objection. We argue that Caney's method is unlikely to be effective in averting the climate emergency for reasons stemming from its indeterminacy. We advance this argument in three steps. First, Caney's method alone is incapable of yielding a determinate allocation of secondorder responsibilities; at the very least, it must be supplemented with localized information in order for it to be applied by actual agents to actual situations. Second, when local information is added into the analysis, it becomes clear that agents face a wide range of choices as to how to influence the structural context in which other agents make emissions-affecting decisions. Crucially, these choices will be normatively complex, in the sense that they are not determinable by the mechanical application of Caney's normative principle to the relevant facts. Accordingly, these choices would benefit from normative guidance-guidance that falls outside the normative prescriptions supplied by Caney's method. Third, given the empirical and normative complexity facing individual agents when deciding what to do about climate change, the generalized, indeterminate responsibilities generated by Caney's method are likely to have limited positive influence on real-world climate action. To see how Caney's method alone is incapable of yielding a determinate allocation of second-order responsibilities, consider first the very high level of generality at which Caney specifies the ultimate goal and describes the tasks needed to achieve it. The specified goal is to avert the climate emergency by staying within the 2°C global warming limit. 32 The high level of generality at which the goal is specified leaves very wide-ranging possibilities as to the combination-sets of tasks that would be sufficient to attain that goal (such 'combination-sets' are often called 'pathways'; we will use the terms interchangeably). This complexity is masked by the high level of generality at which Caney specifies his 'tasks': these are really broad categories of tasks, since each 'task' would need to be broken down into countless sub-tasks before it could feasibly be undertaken by capable individual agents (and thus serve as the basis for an allocation of second-order responsibilities to those agents).
Because of this complexity, there will inevitably be very many agents who are capable of performing tasks that are necessary elements of some hypothetically sufficient combination-set. Consider Caney's task category of 'enabling' (for example, by innovating low-carbon technologies). An engineer may possess engineering skills in hydrogen energy systems that would be valuable in one (hydrogen-heavy) energy decarbonization pathway. But these would be useless in an energy decarbonization pathway that eschews hydrogen. To take another example, consider Caney's task category of 'civil disobedience'. The possibilities for civil disobedience campaigns are very wide indeed, and very many people could play various possible roles in any given hypothetical civil disobedience campaign. But the civil disobedience actions by which a given agent will most effectively contribute to achieving the ultimate climate goal will be highly contingent on local factors, including on actions already taken by others in a particular context.
Moreover, for any given sufficient combination-set, there will inevitably be very many agents who are capable of performing any of the necessary tasks in that set. Say a particular civil disobedience action requires 1 million protestors to achieve its political objective. I may have the capabilities necessary to perform the protesting task, but so might 5 million of my compatriots; the total capabilities needed to execute the task would be surplus to requirements.
The wide range of possibilities-with respect to both the choice of pathway and the allocation of tasks within the chosen pathway-means that efficaciously achieving the ultimate climate goal will require a high degree of coordination among agents (division and specialization of labour, sequencing of tasks, and so on). 33 Yet the absence of any process at the global level for doing the coordination means that most of it must occur at more localized scales (national and lower). Agents will have to make choices that are highly contingent on and responsive to the choices of others in these localized contexts. These choices will require information that must be sourced by engaging with the politics of climate action where, so to speak, the action is. Now, we do not wish to imply that Caney is committed to denying the need for local information in order to specify a given agent's second-order responsibilities. Indeed, it would be open for Caney to agree completely with the analysis so far, without any harm being done to the truth of the power/ responsibility principle. 34 However, even if all one needed in order to derive one's particular second-order responsibilities were to combine the principle with a set of empirical facts, it would not follow that this would be the best normative method for positively influencing climate action (we suggest below some reasons why it is unlikely to be). Still, we do not think that one's second-order responsibilities could be determinately derived merely by adding a set of empirical facts to the power/responsibility principle. Rather, the choices facing agents about how to alter the structural context in which others make emissions-affecting decisions will often be normatively complex. Caney's method cannot provide normative guidance to those choices, since it allocates the responsibilities that are meant to constitute such guidance in a way that leaves open the very choices in need of normative guidance.
There are at least five kinds of normatively consequential choices that agents face if they are to know what their specific second-order responsibilities are. The first two follow from the above points about multiple possible pathways and surplus capabilities, respectively.
First, under conditions of uncertainty about the best pathway to the goal, capable agents face a choice between how to allocate their capabilities to tasks that might fall into different possible pathways. For example, should a talented engineering student specialize in hydrogen cells, electric batteries, or carboncapture techniques? Given the wider normative implications of different pathways to the climate goal, this is a normatively complex choice.
Second, agents must choose whether to contribute to a task when doing so may be superfluous. Perhaps anyone with the capabilities to perform a necessary task has a second-order responsibility to do so, even if the total actions performed result in over-performance (that is, even if not every capable agent's performance of the relevant task is a necessary element of the sufficient combination-set). 35 But this is implausible, since performing tasks will typically be costly to the agents who perform them, and Caney recognizes that such costs, if sufficiently normatively weighty, militate against the agent having a second-order responsibility to perform a given task in the first place. 36 If everyone is required to perform every necessary task they are capable of performing, then pretty soon the personal costs will overwhelm most agents, relieving them of any further second-order responsibilities (for example, in respect of other necessary tasks). 34 See above, n. 20. 35 Caney says, 'Though it may be too strong to say that the intervention of all these agents is necessary, it is plausible to suggest that the action of many of them in concert is necessary (and sufficient) to provide the right kind of choice-architecture to avert dangerous climate change'; 'Two kinds of climate justice', p. 143 (emphasis in original). 36 Ibid., pp. 143-6.
The inevitable result of this inefficient allocation of effort would be a shortfall between what gets done on the basis of actual second-order responsibilities and what would need to get done to avert the climate emergency.
Putting this in the terms of our civil disobedience example: if 5 million people protest and 4 million of them are surplus to requirements, then the 4 million surplus could have allocated their resources more efficiently in terms of achieving the climate goal, that is, in ways that make the goal more likely to be achieved. Of course, no one can know in advance exactly how many protestors will be needed to make a difference (or exactly which energy system will turn out to be best, and so on). The point is that, given such uncertainties, many agents will face a normatively complex choice about how to allocate their limited time and capabilities to different tasks within a given sufficient combination-set.
The third normatively consequential choice many agents applying the power/ responsibility principle would need to make arises from inevitable trade-offs between discharging emissions-intensive second-order responsibilities and complying with their first-order responsibilities. For example, imagine a leading solar energy engineer from an industrialized country who needs to fly around the world extensively in order to produce and disseminate her world-leading research into breakthrough solar energy technologies. Would she be relieved of her first-order responsibility to reduce her personal greenhouse gas emissions in these circumstances? Perhaps first-order responsibilities would take precedence (they are, after all, 'first-order'). This would have the virtue of clarity, but it seems counterintuitive in the case of our solar engineer (and no doubt many others), and therefore objectionable as a general principle.
Caney, by contrast-perhaps motivated by cases like this-suggests a different solution: 'those who undertake burdensome second-order responsibilities can be compensated by allocating to them reduced first-order responsibilities'. 37 Caney's solution seems intuitively preferable, but has the drawback of introducing uncertainty into the content of agents' first-order responsibilities, since some further principle is then needed to determine the conditions under which secondorder responsibilities displace first-order ones, and by how much. What that missing principle might be, Caney does not say. 38 Moreover, if performance of burdensome second-order tasks can displace first-order responsibilities, then the respective composition of agents' first-order responsibilities would be constantly in flux, further complicating the derivation of second-order tasks (the function of which, remember, is to ensure agents comply with their first-order responsibilities) and hence also complicating the identification and allocation of other agents' second-order responsibilities. Simple reflection on the calculations and compromises between these objectives that those of us working on climate change make every day should remind us of their normative complexity.
Fourth, responsible agents must choose between how much of their time and capabilities they spend on second-order climate tasks as opposed to discharging other 'imperfect' moral duties and responsibilities (for example, reducing poverty).
Fifth, and finally, agents must choose how much of their time and capabilities they spend in total on discharging moral responsibilities given the personal costs and opportunity costs associated with such activities. The scope for choice in this regard arises from normative uncertainty as to the extent of one's personal prerogative (or as to the level of personal costs one is expected to bear before one is relieved of one's imperfect moral duties).
Given the empirical and normative complexity facing any agent with respect to their second-order (climate) responsibilities, we question the effectiveness of Caney's method in positively influencing climate action-however closer it may bring us to moral truth (on which we take no stand here). The theory of change implicit in this method is a general and impersonal one: some people may read the article, feel the force of the normative argument for second-order responsibilities, and become motivated to navigate the empirical and normative complexity involved in determining what specifically they should do to change the structural context. We do not doubt that this could have a desirable effect on some people, but the complexity left open by the method surely blunts its transformative potential.
Moreover, the kinds of action most needed to shift the structural context in which emissions-affecting actions occur are those involving groups of peoplesocial movements, labour unions, NGOs, corporations, and, of course, governments. In group contexts, not only will the empirically and normatively complex choices left open by Caney's method need to be made in one way or another among individuals who may potentially disagree about which choices to make, but the group is likely to face an additional source of normative complexity that Caney's method cannot accommodate: the virtual inevitability of disagreement about the truth of the power/responsibility principle. Assuming the truth of, and applying, a particular principle such as this may entail ignoring or concealing value conflicts among the group, excluding alternative normative perspectives, failing to present what is really at stake in the normative conflict, or failing to generate solutions capable of being broadly accepted by the relevant group. 39

B. Unperformed Responsibilities: The Need for Motivated Agents
Let us now assume for the sake of argument that second-order responsibilities could be determinately specified according to Caney's method and that, if fulfilled, 39 For more extended critiques along these lines, see Goodhart the tasks performed as a result would be sufficient to achieve the ultimate climate goal. Even then, Caney's method would face a second problem: many of the second-order responsibilities so allocated would go unfulfilled. Consequently, the method's application under real-world conditions would leave a large gap between the allocation of second-order responsibilities and the hypothetical allocation of responsibilities that would most efficaciously avoid the climate emergency, meaning the method fails on its own terms.
This performance gap can be appreciated by reflecting on an obvious empirical fact that Caney seems to overlook: the vast majority of powerful agents in the world-those with the relevant capabilities to undertake second-order tasks, in Caney's terms-are not only failing to discharge their first-order responsibilities to reduce their greenhouse gas emissions, but are also failing to act on their second-order responsibilities (a fact that is not likely to change as a result of others exhorting them to fulfil their second-order responsibilities). Perhaps we could allocate third-order responsibilities to those agents who are most capable of making agents comply with their second-order responsibilities. But what happens when those third-order agents shirk their (third-order) responsibilities? There seems to be a problem of regress that Caney's method has no way of escaping.
The problem arises from the fact that capable agents will often lack the motivation to discharge their (first-+ nth-order) responsibilities. Since Caney's method allocates responsibilities based solely on capability, it can never gain a foothold in agents' motivation to perform relevant tasks, and so to the extent that second-(+ nth-)order responsibilities are allocated to the capable-butunmotivated, the method cannot get off the ground.
This weakness in Caney's method can be seen as an instance of what Ben Laurence has diagnosed as a more general weakness common to much non-ideal theorising: the misspecification of the agent of change. 40 Laurence notes that the agents who cause injustice 'are often politically mobilized defenders of the status quo', and therefore 'the forces they can bring to bear are among the obstacles to a solution that must be overcome through political action'. 41 Laurence notes that governments, political parties, and corporations often fall into this category. 42 Many of the agents Caney explicitly or implicitly identifies in his discussion of 'tasks' that give rise to second-order responsibilities are governments or intergovernmental bodies (like the WTO, IMF, and World Bank) 43 -powerful group agents who are, for the most part, 'part of the problem'. 44 So how should we think about the agent(s) of change? Like Caney, Laurence recognizes that potential agents of change require suitable capabilities ('capacities and resources', in Laurence's terms) if they are to feasibly overcome the barriers 40 Laurence, 'The question of the agent of change'. 41 Ibid., p. 358. 42 Ibid., pp. 358-60. 43 Caney, 'Two kinds of climate justice', pp. 139-40. 44 Laurence, 'The question of the agent of change', p. 358. that block the path to justice. 45 But, unlike Caney, Laurence argues that capabilities are insufficient: in identifying plausible agents of change, one must also attend to questions of motivation-for feasibility is conditional on an agent trying to bring about the change, and it is unrealistic to think that all agents will do this. A plausible agent of change is a potential agent of change who is, or might realistically come to be, willing to pursue change to address the relevant injustice. 46 We agree that identifying agents of change requires taking account of not only agents' capability to perform a relevant task, but also their motivation-or rather what we shall call their motivatability, that is, their potential to be so motivated (what Laurence refers to as agents who 'might realistically come to be' willing to pursue change). Consideration of agents' motivation and motivatability is thus central to the engaged methods of climate ethics that we will advocate in Section III.
A proponent of principles-first methods may respond that one's motivation (unlike one's capability) is not a normatively justified basis for allocating moral responsibility for undertaking a task. But this misunderstands our purpose: we are not arguing for an alternative ground for allocating moral responsibilities, but rather critically evaluating the potential for principles-first methods (which do focus on moral responsibilities/duties) to positively influence real-world climate action. The response assumes that the task of the philosopher is to allocate duties of (or responsibilities for) justice, and that such an allocation has categorical moral force. But this goes to the heart of our concerns with principles-first methods: if categorical duty-of-justice allocations are essential to normative theorizing, then it is difficult to see how this form of theorizing could have much real-world traction, given facts about agents' present motivations and motivatability. In Section IV, we explain how proponents of engaged methods can respond to the claim that normative theorizing about agents of change requires categorical duty-of-justice allocations.

III. ENGAGED CLIMATE ETHICS
We argue that, insofar as normative theorists aim to contribute to real-world climate action through their philosophical work, they should use an engaged method, which we define as follows: Engaged method. A way of doing moral or political philosophy that involves substantial interaction between the theorist and some actual or potential agent(s) of change, or participation by the theorist in a group agent of change, where such interaction or participation influences the content of the theorist's normative constructs (as opposed to merely providing the empirical facts needed to apply predetermined normative constructs).
Engaged climate ethics, then, involves the use of engaged methods to positively influence real-world climate action. Starting one's theoretical enquiry by scanning the horizon of agents, to consider the political possibilities of their agency given their capabilities and motivations, and then engaging with them to think through the politics and ethics of their predicament in a situated and detailed way, is essential to overcome the deficits that principles-first theorizing inevitably leaves.
We substantiate this claim through a discussion of three families of engaged methods that have been widely applied in other areas of philosophical enquiry. In each case, the 'family' can be conceptualized as lying on a spectrum between two ideal-type social roles: at one end, the ideal-type of the philosopher as traditionally understood; at the other end, the ethnographer, the activist, and the policy-maker.
The first of these families of engaged methods we call ethnographic engaged methods. Here, the theorist uses ethnographic fieldwork or other situational experience as a basis for philosophical reflection about contemporary institutions, organizations, and practices. This method overlaps with the traditional role of the ethnographer insofar as it serves a descriptive/interpretive function. However, insofar as they are interested in description/interpretation, normative theorists have a narrower focus, and they bring unique skills and perspectives to this task. 'In the context of normative inquiry', write Herzog and Zacka, 'adopting an ethnographic sensibility means being attuned to how individuals understand themselves as situated moral and political agents', to the array of normative demands those individuals face in their particular social roles, and to the barriers to and opportunities for moral and political agency that they face. 47 Moreover, ethnographic engaged methods of moral and political theorizing have additional functions beyond description/interpretation, which can encompass conceptual and normative innovation, and practical recommendations for reform. 48 In this regard, it is notable that much recent ethnographic work in normative theory has focused on agents operating within, or proximately affected by, particular institutional and organizational settings-including state institutions, 49 business corporations, 50 and NGOs 51 -and is typically animated by a practical desire to see such institutions and organizations transformed for the better. We see considerable potential in such methods to advance climate ethics.
We see promise, for example, in greater normative engagement by climate ethicists with workers adversely affected by decarbonization and with labour unions. Workers in emissions-and energy-intensive industries face serious socio-economic risks from decarbonization, including unemployment, loss or devaluation of skills, displacement from their communities, and the many flow-on effects of such losses. Beyond advocating the protection of human rights, subsistence needs, or 'rights to development' in the course of decarbonization, climate ethicists have paid little attention to the transitional risks and harms facing such workers, or to their normative entitlements and obligations. 52 One of the strengths of ethnographic-based philosophical reflection is that immersion in a particular context can illuminate novel forms of harm. 53 Another strength is that it can diagnose under-appreciated obstacles to moral conduct. 54 In this vein, future work could usefully uncover the harms experienced by, and obstacles facing, workers in the declining fossil fuel sector and other carbon-intensive industries. 55 This, in turn, might reveal possibilities for new forms of solidarity, new political alliances, and more effective approaches to climate policy.
One reason we see such workers as being fruitful subjects of situated normative reflection is that they are often already-or readily able to be-mobilized and coordinated through labour unions or alternative worker organizations. This gives them considerable political power in relation to climate change. At present, that power is sometimes used to block climate action, with some unions using their power to forge intra-industry coalitions with carbon-intensive corporations and industry associations. Yet other unions and labour confederations are embracing a more progressive stance, arguing for a 'just transition' out of fossil fuels for workers and their communities-demanding social dialogue, social security, retraining, and investment in green industries. 56 Given their potential motivatability, their access to coordinating structures, and their often significant capabilities to influence politics (particularly social-democratic political parties), we suggest that engaged theorizing about the 'just transition' agenda through fieldwork-based interaction with unions and workers in affected industries would be a fruitful focus of climate ethics. 57 Our second family of engaged methods, which we label activist engaged methods, is firmly rooted in the politics of existing social movements fighting palpable injustice. 58 It can be distinguished (from traditional philosophy and from ethnographic engaged methods) by a conscious and explicit identification by the theorist with the cause of the movement with which they are interacting (often as an active participant). Most obviously, much 'critical' theorizing is engaged in this sense, including much feminist, queer, disability, race, indigenous, and postcolonial theory, as well work on class and economic injustice. 59 Among proponents of 'activist'-type methods, numerous recent works of political theory have explicitly theorized about methodology. 60 A common methodological tenet of such work is the combination of a critical-analytical function with a prescriptive-normative function. On Goodhart's 'bifocal method', for example, the theorist switches between an 'analytical lens' and a 'partisan lens'. Through the analytical lens, theorists aim to produce a more realistic account of the status quo by placing political disputes in their historical and social context, and illuminating the ideologies and power relations that lie behind the various positions in a political dispute. 61 The theorist's analysis is deepened through engagement with others experiencing injustices, who bring to bear unique interpretations and criticisms of present conditions. 62 In light of such engagement, analytical theorizing aims at 'interpolating people's disparate experiences and understandings of injustice into a richer-though still provisional-theoretical account of it'. 63 Doing so in turn 'enriches our partisan thinking about how best to articulate arguments, critiques, and prescriptions in ways likely to influence people's behavior, to 57 Worker and union mobilization on climate change has been central to the (overlapping) 'just transition' and 'environmental justice' movements, illustrating the potential for ethnographic engaged theorizing to blend into the activist-type engaged theorizing we discuss below. Ethnographic engaged methods may be more appropriate in cases where workers/unions have potential to mobilize on such issues but are not yet mobilized (or are mobilized in opposition to climate action). 58 61 Goodhart, Injustice, pp. 155-9. 62 Ackerly, Political Theory and Feminist Social Criticism; Goodhart, Injustice, pp. 153-4. There is clearly overlap here with the descriptive/interpretive function of ethnographic engaged theorizing, discussed above. 63 Goodhart, Injustice, p. 155.
win their hearts and minds'. 64 Through the partisan lens, theorists work 'on or with a particular substantive normative view, consciously as ideologists, to organize and clarify concepts, articulate their meaning and relation to one another, formulate critiques animated by those conceptual frameworks, and recommend political action'. 65 The 'environmental justice' movement offers something of a template for an activist-engaged climate ethics. 66 The movement emerged in the USA in the early 1980s from community experiences of inequitably distributed environmental harms, such as the dumping of toxic waste in low-income, predominantly African-American neighbourhoods. 67 Fusing civil rights activism with environmental activism, the movement articulated a distinctively humanistic vision of environmentalism focused on the environmental conditions of everyday life and their implications for people's health and human rights. 68 The concerns of the movement multiplied in subsequent years, addressing many topics that would come to intersect with climate concerns-such as energy, transportation, green spaces, urban planning, and natural disaster management 69 -in significant part through the path-breaking work of scholar-activists such as Bob Bullard and Julian Agyeman. 70 Indeed, the movement went on to articulate a distinctive set of concerns in the climate change debate, emphasizing but extending beyond the imperative to reduce greenhouse gas emissions. This set of concerns has encompassed, among others, the effect of fossil fuel industry co-pollutants on public health in exposed communities, and the racial and class inequities in the way that climate policies (re)distribute such harms. 71 For example, environmental justice advocates have highlighted how some carbon cap-and-trade schemes end up increasing hazardous co-pollutants in low-income communities of colour-a risk that tends to be overlooked in applied philosophical discussions of cap-and-trade 72 -and have proposed substantive and procedural amendments to policy programmes to address such inequities. 73 On this and other fronts, the movement has had a significant positive impact on real-world change, including in the design of major state-and city-level climate policy programmes in California 74 and New York. 75 In this vein, one key theme we see as holding promise for activist-engaged climate ethics is anti-fossil-fuel scholar-activism. A critical analysis of the climate problem that takes history, power relations, and ideology seriously reveals the overwhelming importance of fossil capital in creating and sustaining the structures of our present predicament, and illuminates the mechanisms through which it does so. 76 Such analysis uncovers the role of the fossil fuel industry in multiple forms of injustice throughout the world, extending well beyond the harms caused by climate change itself-human rights abuses, indigenous dispossession, corruption, economic volatility, ill-health, and multiple forms of environmental degradation. 77 And it reveals the ideologies and practices that hide, deny, excuse, and normalize these effects. 78 Yet, at present, we see very limited philosophical attention paid to the distinctive, structural role of fossil fuel companies with respect to climate change; on most normative analysis, they are merely one source of greenhouse gas emissions among many, and are therefore rarely singled out for special attention. 79 Such a critical analysis, in turn, provides signals about where (not) to look for potential agents of change, and suggests possibilities for normative-prescriptive work that can be done with such agents. Because of the multiple harms and injustices caused by fossil fuel practices, mobilized anti-fossil-fuel social movements are already widespread, such as the fossil fuel divestment movement and campaigns aimed at stopping new fossil fuel infrastructure. 80 Such groups are already coordinated, capable, and motivated to take action aimed at changing political structures and norms of ethical behaviour. Accordingly, by interacting with or participating in such movements, normative theorists can learn from them, improving their own normative theorizing by enriching it with insights from the field-including the theoretical and normative insights of participants. They can also contribute analysis and ideas into an existing network of changeagents, thereby improving the odds that their theorizing will have a desirable real-world impact on the climate problem.
In terms of the content of such interactions, we see opportunities for theorizing with social movements to develop normatively informed strategies and tactics for framing and movement-building-both of which are central tasks for the leaders of social movements. 81 One possibility concerns the articulation and evaluation of straightforward deontological injunctions against harmful practices, such as the 'anti-fossil-fuel norms' advocated by the fossil fuel divestment movement and by the campaigns against new fossil fuel infrastructure mentioned above. 82 Normative theorists are skilled at constructing and evaluating normative principles, especially deontological injunctions. Accordingly, they have skills that would make them valuable participants in movements aiming to develop such normative constructs. Yet norms developed on the basis of ideal principles of justice are often too abstract, complex, or technical to be of use to social movement leaders who need to mobilize masses of ordinary people and attract attention in a competitive media landscape. Sustained interaction with anti-fossil-fuel social movements can sensitize normative theorists to the situated demands and obstacles faced by climate activists, yielding opportunities to develop synthetic normative constructs that have a genuine chance of influencing climate action, yet also withstand critical reflection.
The environmental justice movement also teaches us the importance of theorizing climate change politics in an intersectional way, from the perspectives of minority and/or oppressed groups, whose members tend to experience climate 80 See Andrew Cheon and Johannes Urpelainen, Activism and the Fossil Fuel Industry (London: Routledge, 2018); Georgia Piggot, 'The influence of social movements on policies that constrain fossil fuel supply', Climate Policy, 18 (2018), 942-54. 81 Goodhart, Injustice, pp. 192-5; Fergus Green, 'Anti-fossil fuel norms', Climatic Change, 150 (2018), 103-16. 82 Green, 'Anti-fossil fuel norms'. We note that Caney, 'Two kinds of climate justice', pp. 137-8, argues that persons with relevant capabilities have a moral responsibility to work to change social norms. We are not seeking to allocate moral responsibilities; rather, we are encouraging climate ethicists themselves, insofar as they are motivated to contribute to positive real-world climate action, to work with agents of change to shift social norms. change and climate policy differently from members of majority and privileged groups. Work by indigenous philosopher-activist Kyle Powys Whyte provides one example. Whyte has theorized the unique perspectives of Native American indigenous peoples on climate change in light of their collective experiences of adaptation to natural ecological variability and of the injustices wrought by settler colonialism and carbon-intensive industrialization. Building on this knowledge and experience, Whyte envisages alternative strategies for responding to climate change by and with indigenous communities. 83 A second example highlights one of the many points of intersection between climate change and feminist concerns. Alyssa Battistoni shows how pro-labour decarbonization plans often focus on male-dominated, blue-collar jobs while ignoring 'pink-collar' service sector work-work that is labour-intensive, rapidly expanding, mostly low-carbon, crucial to human flourishing in a climate-stable world, and done disproportionately by women, immigrants, and people of colour. Battistoni envisages a decarbonization strategy that centres such work-including advocating raising its pay, conditions, and status. She argues that such a strategy both embodies an inherently more equitable agenda and pragmatically expands the possibilities for political mobilization relative to a narrower decarbonization strategy focused only on infrastructure. 84 These two examples merely scratch the surface of the critical, normative, and pragmatic-political potential of intersectional, activist-engaged theorizing.
A third, and quite different, family of engaged theorizing uses committeebased engaged methods. Here, philosophers join multidisciplinary committees established by organs of government or non-government bodies, such as research councils or NGOs, and engage in the co-construction of normatively informed, evidence-based recommendations. Jonathan Wolff has written extensively on the political philosophy of various public policy issues in light of his experience of this type of committee work, 85 and has set out the methodological steps conducive to doing it well. 86 It must be remembered, however, that we are here interested in the work of philosophers qua engaged philosophers, that is, work that mixes elements of the traditional philosophical role with elements of another social role-in this case, that of the policy-maker. Accordingly, one kind of committee-based work that would not count as 'engaged' in our sense is where a philosopher simply applies 83  an existing normative theory or principle to the concrete issue at hand in order to generate a policy recommendation. 87 This is not to say that philosophers doing committee-based engaged work would not draw on the inherited stock of concepts, principles, theories, and arguments when undertaking their analysis, evaluation, and recommendation-inevitably they will do so. 88 The point is that, as we emphasized in our discussion of Caney's method, real-world agents-and indeed, policy committees-face not only empirical, but also normative complexity that any single abstract principle cannot mechanically resolve. Engagement-in this case with other experts as part of a committee-is the means through which that complexity is embraced, and a normatively desirable way forward is co-constructed.
Another example of committee work that would not count as engaged on our definition is where the philosopher's role is merely to translate or synthesize existing philosophical theories, principles, concepts, and arguments from the philosophical literature. 89 This kind of work may be valuable in its own right, but it would not meet our definition of an engaged method. 90 Climate-related examples of committee-based engaged methods of which we are aware include Caney's work as part of the Nuffield Council for Bioethics' working party on ethical issues concerning biofuels 91 (a role that also illuminates the diversity of methodological approaches Caney has taken), and the work of Catriona McKinnon and Andrew Light on the governance of solar radiation management for the Forum for Climate Engineering Assessment. 92 The possibilities for establishing such committees in the climate realm are endless, but the opportunities to do the work are obviously limited to the committees that are actually established, and to the philosophers who are invited to join them. Since there is limited scope for philosophers to pursue this work on their own initiative, we will not endeavour to suggest fruitful areas of future enquiry here.
This concludes our discussion of each family of engaged methods. We end this section by summarizing the pay-off from engaged theorizing in terms of overcoming the limitations of principles-first non-ideal methods that we identified earlier. First, engagement with potential agents of change brings theorists into contact with a concrete, informationally-rich site from which feasible 93 and normatively desirable pathways of political transformation can be co-constructed with actual or potential agents of change. From this vantage point, theorists and agents of change can and must grapple with the dense empirical and normative complexity that is left open by principles-first non-ideal theorizing. 94 Moreover, they can do so in a manner that is more democratic and sensitive to pluralist values, since they do not need to assume the moral truth of any given theory or principle. Second, by doing one's theorizing with agents whose early moves along that pathway are themselves realistic, because the agents possess the necessary capabilities and motivation (or at least motivatability), the gap between the desired actions and the realistic agency of change-agents is reduced or eliminated, as is the gap between the agent of change and the theorist.
It is true that these conditions circumscribe the possibility space in which normative theorists can do valuable work. Yet, in another sense, the kind of work they can do is dramatically expanded by the richness of material with which they will be confronted in the field, and by the transformative potential inherent in theorizing on the vanguard of social or institutional change.

IV. SCEPTICISM ABOUT ENGAGED METHODS: A RESPONSE
So far we have argued that engaged methods have a greater potential to contribute to desirable real-world change than the standard, principles-first approach to non-ideal theorizing about climate change. We now consider and respond to a set of potential objections to engaged methods, namely that such methods are illegitimate, mistaken, or misguided means of influencing real-world change through academic philosophy. To be clear, we envisage that the objectors accept our assumption that seeking to influence real-world change-or what they might prefer to call 'action-guidance'-is a legitimate objective of academic philosophizing, but reject engaged methods as means of pursuing that objective. We shall assume that the objections stem from holding an alternative view about the right way for philosophers to achieve this objective, namely the principlesfirst approach that we reconstructed in Section I. If this is the case, the objector is likely to see various aspects of engaged methods as illegitimate, mistaken, or misguided.
First, the objector might claim that engaged methods lead to prescriptive theorizing of insufficient generality to be action-guiding in all but the particular cases to which they apply. Second, the objector might claim that engaged methods lead to prescriptive theorizing that lacks categorical moral force. This is problematic, so the objection might go, because it reduces the philosopher to a partisan activist and/or undermines the practical effectiveness of normative theorizing, leaving the philosopher with no objective basis for critiquing and persuading morally responsible agents, or for motivating them to fulfil their duties of justice.
At least three general strategies for responding to this family of objections are possible. The first involves rejecting the principles-first approach altogether, claiming it to be a fundamentally mistaken means of contributing to real-world change. 95 This is the strategy adopted by Mills 96 and Goodhart. 97 Goodhart, for example, argues that all justice claims are necessarily ideological, 98 and that their normativity is the hypothetical normativity that derives from the contingent subscription to the relevant ideology from which the justice claim emanates; 99 hypothetical imperatives 'tell agents what to do given particular aims and circumstances'. 100 For Goodhart, ideological convictions are far more powerful motivators; a more effective route to political change is thus to work with motivated change-agents to do the analytical and partisan work that sharpens their ideas and expands their power. 101 Such work will involve theorization, argument, and persuasion, to which the normative theorist should be well equipped to contribute, but this will inevitably look somewhat different to the argumentation typical of the principles-first approach. 102 Goodhart does not lament the 'loss' (relative to the principles-first approach) of objectivity and moral force that his method implies, because he denies that such objectivity is possible and that such moral force exists. 103 A second strategy is more accommodating of aspects of the standard approach, but holds the latter to be insufficient to the practical-transformative project of non-ideal theory. On this view, engagement with agents of change is a necessary corrective to this deficiency. This is the tack taken, for example, by Shelby 104 and Laurence. 105 Both authors think there are categorical duties of justice, but do not think that non-ideal theorizing can be limited to working out what these are and exhorting the duty-bearers to follow them.
Laurence notes that in cases of pressing injustice, many agents will lack the capabilities or motivation to act on those duties. The theorist, he argues, must therefore search for and work with (capable and motivated or motivatable) agents of change-theorizing their predicament, and constructing feasible-anddesirable responses to it, together. 106 This engagement requires eschewing what he calls 'the technocratic posture', according to which 'the political philosopher is a neutral knower, who stands above the political fray, offering expert guidance to an unspecified subject'. 107 Instead, he argues, non-ideal political theorists must embrace the view that their work is necessarily a partisan and political act. 108 For both Shelby and Laurence, moreover, the courses of action recommended by the engaged political theorist are not exhausted by identifying duties of justice and their bearers; many agents of change will and must take up the cause of fighting injustice despite not having any duty of justice to do so. 109 A third, pluralist strategy is to side-step the standard approach, perhaps accepting it as one legitimate form of theorizing, perhaps not; but in any case advocating engaged theorizing as legitimate and practically valuable in its own right. Something like this stance is taken by numerous proponents of ethnographic engaged methods. Responding to the objection that ethnographic engagement might compromise theorists' objectivity, Herzog and Zacka acknowledge that there are risks of perspectival absorption and bias that arise from ethnographic work, but argue that these can be guarded against through appropriate epistemological and methodological practices. 110 At the same time, they note that no standpoint is objective: armchair theorizing carries its own risks of bias and myopia. 111 Responding to the objection that ethnographic-based normative theorizing will result in critical analyses and prescriptions that are relatively concentrated on the proximate and the near-term, Longo and Zacka embrace this particularity, envisaging it as but one mode of valuable philosophical enquiry along with others that are more idealized and general. 112 The common denominator of these three response strategies is that engaged methods are at least a legitimate means of seeking to influence desirable change in the real world (and of getting a better grip on our present predicament), with virtues and strengths at least different from those of the principles-first approach. They reveal, moreover, that engaged methods are not reducible to mere political pragmatism, activism, or rhetoric, but rather involve distinctive modes of analytical, interpretive, critical, and prescriptive theorizing. These modes look, and are, different from the principles-first approach, but are at least no less legitimate.

V. CONCLUSION
This article has argued that, insofar as contributing to real-world climate action is a legitimate goal of climate ethics, engaged methods are superior to principlesfirst methods. Principles-first non-ideal theorizing, we argued, has limited capacity to contribute to such action. We highlighted two limitations of the principles-first method through a detailed analysis of an article by Simon Caney, which we took to be representative of this method. We then discussed three families of engaged methods. We argued that the substantial interaction with, or participation in, agents of change that engaged methods entail enables theorists to overcome the limitations of the principles-first approach, better equipping them to contribute to real-world climate action. Finally, we defended engaged methods against a set of objections that calls into question their legitimacy as means of achieving realworld change via academic political philosophy.
We note in closing that some theorists may feel uncomfortable with engaged theorizing, or feel ill-equipped or ill-supported to do it in light of current academic structures. We sympathize with these reservations. But they do not amount to arguments against engaged climate ethics. Rather, if our arguments are sound, they speak in favour of challenging and reconfiguring existing academic structures, and broadening the methodological training that is offered to philosophers (a reform project that should lead us to reflect on our own role as potential agents of change). Of course, even if such changes were made, we have no doubt that many readers would still prefer to stick to principles-first methods of climate ethics. This may allow them to make excellent philosophical contributions to the objective of seeking truth, and in this endeavour, we genuinely wish them well. But for those of us wishing to do academic theorizing that makes significant contributions to real-world efforts to avert the unfolding climate emergency, the time has come to roll up our sleeves.