Power in resilience and resilience's power in climate change scholarship

Resilience thinking has undergone profound theoretical developments in recent decades, moving to characterize resilience as a socio‐natural process that requires constant negotiation between a range of actors and institutions. Fundamental to this understanding has been a growing acknowledgment of the role of power in shaping resilience capacities and politics across cultural and geographic contexts. This review article draws on a critical content analysis, applied to a systematic review of recent resilience literature to examine how scholarship has embraced nuanced conceptualizations of how power operates in resilience efforts, to move away from framings that risk reinforcing patterns of marginalization. Advancing a framework inspired by feminist theory and feminist political ecology, we analyze how recent work has presented, documented, and conceptualized how resilience intersects with patterns of inequity. In doing so, we illuminate the importance of knowledge, scale, and subject making in understanding the complex ways in which power and resilience become interlinked. We illustrate how overlooking such complexity may have serious consequences for how socio‐natural challenges and solutions are framed in resilience scholarship and, in turn, how resilience is planned and enacted in practice. Finally, we highlight how recent scholarship is advancing the understandings necessary to make sense of the shifting, contested, and power‐laden nature of resilience. Paying attention to, and building on, such complexity will allow scholarly work to illuminate the ways in which resilience is negotiated within inequitable processes and to address the marginalization of those continuing to bear the brunt of the climate crisis.


| INTRODUCTION
At the time of writing, wildfires are raging across British Columbia, Canada, the heat has extended beyond 50 C (122 F) in Jacobabad, Pakistan, and Hurricane Elsa has left a trail of destruction across the Caribbean. Hurricanes Irma and Maria, Cyclone Amphan, and the burning of Paradise, California loom large in collective memories. As the symptoms of climate breakdown become evident and increase in frequency and magnitude, minds and policy priorities turn to resilience-of communities, buildings, systems, and individuals. Even as we work to reduce emissions, societies must adapt and invest in resilience measures-updating infrastructure, challenging vulnerabilities, changing consumption patterns, and empowering communities.
Although the terminology of "resilience" originated in the 1970s, providing a language to understand the extent to which a system can tolerate shocks and change (see Holling, 1973), it has now become a central priority in contemporary climate politics (Nalau & Verrall, 2021). Across disciplines, the concept of resilience takes shape in many different forms and understandings. In the context of resilience to social and environmental stressors, at times, it is seen as an outcome to be aimed for, even though, in reality, it is difficult to achieve (de Abreu-Mota et al., 2018;Kmoch et al., 2018). Increasingly, however, resilience is seen as a socio-political process that is continuously negotiated between a range of actors with diverse needs, interests, and levels of authority (Harris et al., 2018). For the purpose of this article, we frame resilience in line with the latter; values need to be deliberated and different processes and techniques used to "build" resilience and re-humanize socio-natural relations (Mikulewicz, 2019).
If resilience is indeed apprehended as a socio-political process, a more nuanced understanding of how it is influenced by and inscribed with societies' power relations and inequities is fundamental. Contemporary patterns of vulnerability have deep roots; historical patterns of inequity define who is located in areas of exposure (Ranganathan & Bratman, 2021), colonial relationships and structural dependency complicate recovery efforts from disasters (Moulton & Machado, 2019), and efforts toward resilience often protect privileged individuals and communities while excluding or harming others (Rice et al., 2021). Patterns of political, economic, social, and cultural power (be it visible or otherwise) influence who is resilient and who is constructed as vulnerable-but also who has access to arenas of decision making and, with it, agency in defining solutions (Carr, 2019). A rigorous conceptual understanding of such processes of power is necessary for appreciating the web of ambiguous and sometimes contradictory relations between risk and vulnerability-and, in particular, who is resilient, who is not, why, and with which consequences.
However, similarly to concepts such as "adaptation" (Woroniecki et al., 2019) and "vulnerability" (Camponeschi, 2021), scholarly discussions of resilience demonstrate a lack of clear conceptualization with respect to power (Jordan, 2019). Such reluctance has been criticized for flattening complexity to focus on climatic over social drivers (Arora-Jonsson, 2016;Kashwan et al., 2019), over-emphasizing the stability of current systems and processes (Carr, 2019), isolating resilience from broader socio-natural processes (Nightingale et al., 2020), and presenting associated policies as technical and value-free (Mikulewicz, 2019). All result in a distancing of scholarship from addressing the "messy realities that would challenge established forms of governance and policy-making" (Nightingale et al., 2020, p. 345) and benefit those who are already privileged.
A more nuanced exploration and understanding of the links between resilience and power inequities-found in feminist theory and feminist political ecology scholarship-allows for the conceptualization of the interlocking, intersectional axes of power that shape and restrict resilience and associated decision making (Resurrecci on, 2017;Tschakert et al., 2016a). In doing so, a language of resilience can be reconciled with hopes for transformation or radical change. Drawing on, among other corpora, Black feminist discourses (see Crenshaw, 1989) and work on power and subjection (see Butler, 1997), recent contributions to this field offer an understanding of resilience that foregrounds analysis in its ever-changing and power-laden complexity (Bezner Kerr et al., 2018;Osborne, 2015;Stock et al., 2021;Sultana, 2020). Such complexity leads to resilience being understood as "negotiated" (Harris et al., 2018) and "situated" within specific contexts (Cote & Nightingale, 2012;Tschakert & Tuana, 2013).
Although critical work has problematized an insufficient conceptualization of power in resilience policy and scholarship, few studies provide clear recommendations for how future research can develop a more nuanced understanding. How power is presented, detailed, and scrutinized in studies of resilience has important consequences. For example, elevating certain groups as "powerful" may well neglect variance within them, just as defining other groups as "powerless" can justify normative interventions. In turn, inadequate reflections of power obscure the significant roles that inequalities play in vulnerability and exposure. The deployment of more nuanced conceptualizations of power for envisioning more just resilience processes will allow for transformation-not only rendering people less exposed to risk but also identifying co-benefits and opportunities for change Roberts & Pelling, 2020).
It is against this background that our study presents a systematic literature review of 202 academic peerreviewed articles, published between 2015 and 2020, to explore and detail how power is understood in recent resilience scholarship. We draw from an inductive content analysis approach that adopts a threefold analytical framework for exploring how scholars conceptualize the role of power in resilience, encompassing three interrelated entry points inspired by feminist theory and feminist political ecology: (1) the politics of knowledge; (2) the politics of subject making; and (3) the politics of scale. Through our analysis, we probe multiple, overlapping lines of inquiry: Why does it matter how power is conceptualized in resilience scholarship? What narratives with respect to power in resilience do particular understandings open up and close down? And what are the consequences and for whom?
By probing conceptualizations of power through a comprehensive, replicable analysis, we hope to contribute to ironing out shortcomings in contemporary literature, to illuminate productive work in defining the role of power and advance scholarly efforts to address the oppressive relations that continue to permeate resilience efforts. In particular, we highlight the importance of (1) understanding that power hierarchies are not set in stone, but can be challenged, with people and communities "pushing back" to reclaim meanings of resilience and recalibrate pervasive subjectivities; (2) focusing on power as a process, rather than an outcome, to avoid neglecting the ever-changing "messiness" of uneven relations in resilience, and; (3) addressing our responsibility-as researchers-to avoid perpetuating power inequities and to bolster the capacities of disadvantaged individuals and societies as agents of change. Ultimately, we argue that progress in these spaces is crucial for advancing theoretical inquiry at the nexus of power and resilience, and for foregrounding new, more transformative resilience trajectories.

| METHODOLOGY
This study is based on a systematic review of academic scholarship that, in line with Gough et al. (2017), entailed a rigorous and standardized set of procedures to identify, screen, and analyze recent literature at the nexus of power and resilience. Three main stages (Table 1) were completed for collecting and screening the literature, coding identified articles, and examining how power is conceptualized across the sample.
First, a comprehensive search procedure was performed using Scopus ( Figure 1). The basic search criteria required that articles were published between January 2015 and March 2020, were written in English, and comprised empirical, theoretical, methodological, or mixed contributions to scholarship. An initial search using firstorder key words ("climate" AND "resilience" AND "power") returned 251 articles (Table 2). To strengthen the reliability, a small group from the broader research team (composed of 16 members located in eight countries) screened articles for duplicates and relevance, with those returned due to overlapping terminology rather than relevance removed. The screening process returned 80 remaining articles. These were scanned for additional terms related to resilience (second-order key words) and dimensions of power and inequality (third-order key words); this exposed that few of the articles addressed the central elements of uneven relations in resilience (such as "race", "disability", "ethnicity", and "subjectivity"; see Figure 1 for further details). Accordingly, a second Scopus search was conducted using the first-order key words of "climate" and "power" (but excluding "resilience"), combined with the secondorder key words (e.g., "adaptation" or "transformation"). The second search yielded an additional 122 articles, for a total sample of 202 articles.
Second, we developed a process for tracking and coding the sampled literature. The first step involved the creation of a tracking table (using Excel) with each article listed as a separate row and different categories listed as separate columns. In total, there were 38 categories grouped under six clusters (Table 3). Before coding commenced, a small group went through the tracking table and tested it, with changes made to ensure it was fit for purpose. After the tracking table was finalized, all team members were assigned between 10 and 18 articles to read in full and code. Full-text coding across the 38 categories was then entered into the tracking table (including some entries labeled as "n/a" or "not specified"). The focus of this review process was on which dynamics of power and (re) negotiation were studied, how they were understood and depicted by authors (both conceptually and through empirical examples), and at what scales. Detailed guidance documents were provided to the entire team to ensure standardization and all contributions were checked for consistency. We conducted a second round of thematic coding on specific categories and clusters to extract major themes, nuances, and outliers. Themes were identified inductively and coding was achieved using manual color-coding tools in Excel. All information under the 38 categories was arranged horizontally in Excel so that all codes for the 202 articles could be arranged side by side for viewing and making preliminary visualizations.
Third, for the purpose of this article, we utilized a critical content analysis approach (Vaismoradi & Snelgrove, 2019) to closely examine data coded under the "power" cluster categories (Table 3), as well as some others for cross-checking results and ensuring critical insights were not missed (within the same spatial scale; axes of inequality; negotiation, deliberation, and contestation). We developed a threefold analytical framework (Section 3), based on themes that emerged inductively through the thematic coding of the power categories and T A B L E 1 Stages of the systematic review

Stage
Step Process Outcome In-depth inquiry into how power is conceptualized in resilience scholarship critical insights from feminist theory and feminist political ecology. This framework facilitated a comprehensive and nuanced analysis, justified by the delineation of its components according to existing themes in the literature and the cognizance of feminist scholarship with respect to the operation of power in resilience processes (Section 3.1). Through our analytical approach, we were able to assess how authors engage with key elements of power, such as knowledge politics, cross-scalar interactions, and the construction of subjects. This allowed us to probe, in detail, why the conceptualization of power matters in resilience scholarship, the kinds of narratives related to power and (re)negotiation that particular conceptualizations might open up and close down, and the consequences (and for whom) with respect to how resilience is framed and enacted. First-order key words (core concepts of review) *power* AND *climate* AND *resilience* Second-order key words (main themes related to resilience) *mitigation* OR *adaptation* OR *disaster* OR *transformation* OR *security* (water, food) OR *governance* OR *capacity* OR *livelihood(s)* OR *urban* OR *agricultural* OR *rural* Third-order key words (main themes related to power and inequalities) *ethnicity* OR *race* OR *class* OR *gender* OR *Indigenous* OR *age* OR *disability* OR *justice* OR *participation* OR *inequality* OR *identity* OR *stakeholder* OR *conflict* OR *knowledge* OR *rights* OR *authority* OR *subjectivity/ies*

| ANALYTICAL FRAMEWORK
In this section, we first outline critical debates in feminist scholarship, in particular feminist political ecology (FPE), that have helped shape nuanced engagements with resilience as a power-laden process. For the purpose of this article, we view "power" as representing the fluid 1 and dynamic relations, both oppressive and liberating, that move between everyday individuals, institutions, and modes of governance (Allen, 2018;Butler, 1997;Resurrecci on, 2017;Sultana, 2009), and that shape how resilience is negotiated and enacted across levels and scales of decision making (Nightingale, 2017). In line with this understanding, we propose a threefold analytical framework for analyzing how power is conceptualized in recent resilience scholarship (Sections 3.2-3.4).

| Feminist theory and feminist political ecology
FPE scholars, and feminist theorists more broadly, have played a central role in embracing and advancing conceptual thinking on power and resilience, by demonstrating the political and normative components of socio-natural struggles and relations. Central to this work, feminist political ecologists, such as Eriksen et al. (2015) and Ahlborg and Nightingale (2018), have developed nuanced understandings of the workings of power through novel engagements with the social, economic, and political processes that make people vulnerable to socio-natural changes. Drawing on early debates centering on power, subjectivity, 2 and vulnerability (Allen, 2002;Butler, 1997;hooks, 1989), this work has made substantial contributions to feminist scholarship that aims to push examinations of power beyond discussions of who holds power over whom, toward the study of power in action in contexts of resilience (e.g., Cuomo, 2011; Garcia et al., 2020;Gonda, 2019;Kaijser & Kronsell, 2014;MacGregor, 2010;Nightingale, 2017;Tschakert et al., 2016a). As a vital contribution, critical feminist work has illuminated how knowledge (both about and within) resilience is embedded in everyday symbolic and material practices (Bee, 2016;Kerr et al., 2018;Ravera et al., 2019). For example, Bee (2016) demonstrated how women in central Mexico utilize cultural knowledge on food gathering to protect their families during drought periods. Drawing on Haraway's (1988) theory of situatedness, this and other compelling evidence highlights the need to position resilience approaches within the lived and experiential knowledge of social actors (Nightingale, 2016;Nightingale et al., 2021). An important constituent for feminist scholars concerned with knowledge-power interactions is "reflexivity" in resilience research; it entails critical reflections on the positioning of scholars within complex power dynamics that influence how knowledge is produced and how methodologies are put into practice (Sultana, 2007). Reflexive approaches provide openings for trust, rapport, and collective action between diverse stakeholders with varying degrees of power in resilience. Still, as Parsons et al. (2017)  The visible, hidden, and invisible power categories under the "power" cluster were adopted in line with Gaventa's (2006) conceptualization of forms of power, as an accessible method for categorizing power dynamics across a diverse team of scholars and practitioners during the literature tracking and coding process.
For the purpose of this article, a more complex framework at the nexus of scholarship on the politics of knowledge, politics of scale, and politics of subject making was developed (Section 3).
operate to afford specific knowledge claims over resilience authority, while disregarding others, through historical marginalizations and entrenched (ne)colonialisms. Across FPE and other feminist disciplines, the underprivileging and/or blatant exclusion of particular knowledges through top-down structures of power is often linked to the construction of pervasive subjectivities; power operates to oppress and/or empower social actors through material and discursive processes that (re)produce specific, often punitive subjects in resilience efforts (Nightingale & Ojha, 2013). In turn, these actors internalize and (re)produce uneven power in relation to one another and external modes of governance (Nyantakyi-Frimpong, 2017;Van Aelst & Holvoet, 2016). For example, Bezner Kerr et al. (2018) highlight how rural actors in Malawi embody colonial notions of "destructive farmers" that allow state authorities and industry actors to shift the blame for environmental impacts. For feminist political ecologists, these processes are intersectional (Kaijser & Kronsell, 2014;Nightingale, 2011;Tschakert, 2012), with everyday resilience inscribed by oppressive relations that subjugate groups and individuals along the lines of overlapping social differences, such as race, class, gender, age, and Indigeneity (Nightingale, 2011). Such a focus is rooted in Black feminist discourses that called for new frameworks capable of comprehending the coconstitutive nature of inequality (Choo & Myra, 2010;Crenshaw, 1989).
In resilience studies, feminist scholars have embraced an intersectionality lens to move away from dichotomous framings, still pervasive in the climate change literature, that reinforce false binaries between those who are men or women, resilient or at risk, and powerful or powerless (Djoudi et al., 2016;Ravera et al., 2016;Sultana, 2010;Thompson-Hall, 2015;Section 4.3). In contrast, an intersectional approach widens the analytical scope to expose how inequalities and social differences are bound up in a "multiplicity of axes of privilege and oppressions" that (re) produce differential opportunities for resilience (Gonda, 2017, p. 176). Although some intersectional studies have been criticized for conceptualizing overlapping social dimensions as fixed, failing to account for the fluid and everchanging nature of power (Kaijser & Kronsell, 2014), others demonstrate how power dynamics and the multidimensional inequalities they produce shift and change with cultural and temporal mobilities (Carter, 2019). Such intrepid frameworks confront the complexities of power and grasp the dynamic nature of intersecting inequalities (e.g., Elmhirst, 2011;Gonda, 2019;Leder & Sachs, 2019;Osborne, 2015;Sultana, 2013). Crucially, they also provide ample space for examining counter-narratives and counter-conducts, and the diverse ways in which groups and individuals push back against resilience processes that, at times, reassert oppressive power relations (Eriksen et al., 2015;Gonda, 2019;Tschakert et al., 2016a). F I G U R E 2 Summary of the threefold analytical framework used to assess how power is conceptualized in recent resilience scholarship, drawing on feminist theory and feminist political ecology provides nuanced glimpses into the relational construction of subjects and entrenched hierarchies in resilience decision-making processes. As early work by Rocheleau et al. (1996) contended, engaging with scale to discuss power in resilience is crucial; dominant actors deploy power within scales to influence climate change policy and governance (Yates, 2012). Such deployment often occurs through the delineation of subjects and knowledge claims that travel unevenly across levels of resilience, illustrating the transience of power that persistently relegates those most vulnerable to climate impacts to the lowest ranks of decision-making pecking orders (Agarwal, 2001;Nagoda & Nightingale, 2017).
Through a feminist perspective of power in action, resilience comes to represent the interconnected, dynamic, and at times, contradictory processes of power that mandate differential capacities to respond to socio-natural changes. In particular, feminist inquiry has illuminated the discursive and material processes that maintain top-down operations of power and authority, and alter resilience action through elevating or constraining knowledge and access to decision making (Nightingale, 2017;Shrestha et al., 2019). They have also shown the contested nature of resilience by illustrating how oppressive structures are challenged, and subjectivities recalibrated, in the contextual web of resilience-as-aprocess (Rao et al., 2019;Steen, 2011). In drawing on these critical debates, we offer a multi-pronged analytical framework, comprised of three interrelated strands of thinking ( Figure 2).

| Politics of knowledge
The first side of our analytical framework focuses on how power operates to provide certain knowledge claims on resilience authority while dismissing others. Specifically, this lens zooms in on the ways in which scholarship probes and delineates processes of knowledge production, dissemination, and interaction that can reinforce certain worldviews and uphold the exclusionary status quo (Wijsman & Feagan, 2019). We frame knowledge hierarchies as providing a conduit for dominant regimes to delimit which epistemologies are valid and valued, and which are not (Anderson et al., 2019)-for example, by defining how climatic processes are understood (Bremer et al., 2017) or which language is used in resilience decision making . Contemporary climate change mitigation and adaptation policies are inscribed by broader patterns of authority that privilege certain forms of knowledge (Eriksen et al., 2015). This influences who is deemed resilient (or not), what requires intervention, and what solutions should be prioritized. This is often evident in a supremacy of cognitive knowledge deemed "scientific" or "expert" over everyday, experiential, and emotional knowledge, as highlighted across feminist scholarship (Bee, 2016;Nightingale et al., 2021;Ravera et al., 2019). In line with Cast an Broto (2020), we view the "messiness" of multiple and situated forms of knowledge as intrinsic focal points for the imagining, and subsequent carving out, of alternative and liberatory resilience futures.

| Politics of subject making
The second side of our analytical lens scrutinizes if and how authors engage with the politics of subject making. This lens encompasses how specific subjects in resilience, both oppressive and empowering, are made and remade through material and discursive processes (Nightingale, 2017). The politics of subject making allows scholars to examine how multiple axes of power, such as those embedded in patriarchal and (ne)colonial agendas, subjugate specific actors through complex interrelations between individuals, institutions, and modes of governance. Moreover, it exposes how such subjugation is mobilized through the internalization of power asymmetries and embodied performances of social difference, compellingly shown in feminist political ecology scholarship (Bezner Kerr et al., 2018;Gonda, 2021; Section 3.1). At the same time, in line with nuanced expositions of how disenfranchised actors leverage the fluidity of power (Garcia et al., 2021;Steen, 2011;Tschakert et al., 2016a;Van Aelst & Holvoet, 2016), we view it as crucial to probe how groups and individuals can reconfigure their subject positions through subtle and outward contestations of oppressive structures. Through such inquiry, scholars can account for the shifting and ambiguous character of subjectivities in the making.

| Politics of scale
The third side of our analytical framework pays close attention to how scholarship interrogates the choices resilience programs make about scale. Such choices shape which resilience processes emerge as causal, which impacts are taken seriously, and which levels are prioritized in decision-making processes. In particular, this lens centers on the exclusion of specific spatial and jurisdictional modalities (such as communities and local governing bodies) from decision-making spheres, in turn producing "scalar mismatches". Mismatches represent the sites at which uneven power crystalizes through divergent views of what resilience means, for whom, and with what consequences (Holland, 2017). For example, top-down resilience initiatives often clash with the interests, values, and aspirations of local residents, thereby thwarting bottom-up efforts toward empowerment (Nightingale, 2017). Moreover, when resilience projects fail to address cross-level and cross-scale interactions, scalar politics become the means through which pervasive inequalities are (re)produced, as seen in the case of social hierarchies in Nepal crossing scalar boundaries to stifle already disenfranchised actors in adaptation committee meetings (Nagoda & Nightingale, 2017). We align with Fisher and Dodman (2019, p. 245) to view such uneven relations as contributing to divergent understandings of resilience, as "[T] he framing of where processes and change happens are imbued with normative constructs of scale".

| THE CONCEPTUALIZATION OF POWER IN RESILIENCE SCHOLARSHIP
Through the use of our multi-pronged framework, three overarching themes emerged from our analysis with respect to how authors interpret, present, and scrutinize uneven relations of power in resilience processes: (1) hierarchizations in knowledge-production processes; (2) movements of power and (re)negotiation; and (3) competing views of power as an outcome or a process. In the following sections, we illustrate how conceptualizations of power fall across these themes, with varied levels of nuance and complexity. In doing so, we highlight how specific ways of engaging with power open up (or close down) space for critical scholars to better comprehend resilience as a power-laden process that is perpetually in-the-making. It is crucial to acknowledge here that although this analysis provides pertinent insights on "blind spots" and "bright spots" within the recent literature focused on power and renegotiation, the analyzed sample covers a modest area of climate resilience studies. Much of the broader resilience research fails to make the much-needed space for nuanced understandings of power, or follows uncritical, positivist conceptualizations of how it operates. Progress in this space is of vital importance for scholarship to overcome oversimplified framings, within critical studies and beyond, that risk perpetuating entrenched marginalizations and to foreground more just approaches for actualizing meaningful resilience for disadvantaged groups and societies.

| Hierarchizations in knowledge-production processes
Hierarchies in resilience emerge as a core focal point in the literature surveyed, making visible how authors engage with who acquires power in resilience processes, and at what scales. In particular, as especially evident in studies of knowledge politics, the command that state-level actors and central institutions have over decision-making and knowledgeproduction processes is a prevailing line of inquiry. Close to a quarter of the 202 sampled manuscripts detail how the state (broadly defined to include institutions, agencies, and government actors) exercises power that stipulates which kinds of knowledge are defined as legitimate, and in turn, are welcomed in policies of climate change adaptation and resilience (Bastakoti & Davidsen, 2017;Chu, 2018;Eriksen et al., 2015;Geun Ji, 2019). This can be seen in work exploring how institutional definitions of what is "acceptable" land use in Uganda influence wider processes of land grabbing and access to natural resources that contribute to the subjugation of local communities (Carmody & Taylor, 2016). Such dynamics can lead to clashes between competing visions of what resilience might be, with notions of resilience at one level contradicting those at another (Taylor & Bhasme, 2020). As Ajibade (2017) demonstrates in the case of Eko Atlantic city in Nigeria, hierarchizations that define the type of knowledge needed for resilience support elite interests over local knowledge and the values of poor residents. The sampled literature shows this is not a new process and, instead, has deep roots in the settler-colonial project, which establishes norms and legal structures to constrain alternatives (Bell et al., 2019), restrict Indigenous mobility (Whyte et al., 2019), and erode traditional belief systems (Hirons et al., 2018).
Such imperative recognition notwithstanding, a focus on state institutions in studies of scalar hierarchies can neglect and downplay the role of other actors who apprehend and employ power. Research has explored how scientific knowledge can neglect end-users (Harvey et al., 2019) and how simplified metrics flatten the complexity of lived experience, everyday life, and positionality (Grabowskia et al., 2019). Hierarchies of knowledge and scale shape resilience measures-dictating risk and vulnerability through numerous means, including classification, measurement, and definition and problem framing (Fraser, 2017). As a result, the work of researchers, consultants, and donors involved in resilience projects and policy design represent important sites at which hierarchies of decision making and knowledge production play out. For example, national governments' dependence on external funding can lead to donor organizations having an outsized influence on project development and governance (De Roeck, 2019; Lebel et al., 2018;Wilbanks, 2015). In Thailand, for instance, NGOs attempt to influence government policy and shape decision-making processes (Ober & Sakdapolrak, 2020). In such contexts, the preferences of NGOs, organizations, and employees from the Global North may well be prioritized over those of local communities (Wood et al., 2016).
Several authors, however, provide more hopeful expositions of how researchers and consultants can center local knowledge in resilience research and interventions. Morchain et al. (2019), for example, illustrate how Oxfam's approach to assessing vulnerability and risk to climate impacts in the Global South provides a voice to otherwise marginalized groups and individuals through cross-scalar and cross-cultural collaboration. Other studies highlight the role of researchers in paving new ways to include local perspectives in resilience knowledge building, in particular through participation and co-development (Bremer et al., 2017;Grabowskia et al., 2019;Kmoch et al., 2018). As a notable example, Bremer et al. (2017), in their work on the co-production of knowledge in Bangladesh, provide practical accounts of how narratives can be harnessed to illuminate, reflect on, and organize diverse understandings of climate variability. A handful of studies stipulate reflexivity as fundamental to such engagements; Tschakert et al. (2016a, p. 189) utilized a reflexive approach to collective learning in their study of flood adaptation in India to "build trust, […] address power imbalances, and shift from a practitioner/expert position to one of a participant and co-learner". Reflecting on what and how diverse actors know, and on the political dynamics that inevitably permeate our work, critical scholars can crack open much-needed spaces to address our roles in "enabling or dismantling socio-economic and socio-political structures of inequality" (Wijsman & Feagan, 2019, p. 74).
Yet, despite such valuable contributions, attempts to turn the critical mirror inwards to scrutinize the ways in which scholars tap into and affirm power asymmetries are few and far between, thus obscuring the power rooted in academic knowledge and expertise. Just as the priorities and viewpoints of actors who dominate resilience hierarchies often overwhelm local voices, values, and aspirations, researchers may be unwittingly guilty of perpetuating that same process through the make-up of research teams and authorship of publications. In addition to the constant pressure to produce "more" and "better" "data", research projects often prioritize quantity over quality, in particular over the establishment of affective relations that may help to collectively re-think resilience. Within our sample, most theoretical and empirical articles are authored by researchers from the Global North (as per their institutional affiliations), with only a handful of theoretical studies including researchers who live in the Global South (C ordoba Vargas et al., 2020;De Jong et al., 2017;Lebel et al., 2018;Paprocki & Huq, 2018;Sayer et al., 2015;Tanjeela & Rutherford, 2018;Taylor & Bhasme, 2020;Tschakert et al., 2016b). A lack of reflexivity with respect to such dynamics is especially concerning considering the various types of privilege Global North scholars have access to (geographical location, education, class, ethnicity, able-bodies, and so forth) that are persistently denied to others, including the individuals and communities that are frequently enlisted as "subjects" for analysis (108 articles featured case studies from the Global South, compared to only 53 articles from the Global North). "Experts" of resilience, particularly researchers, have a way to go in further exploring and conceptualizing their own complicity in uneven operations of power-as well as their ability (or lack thereof) to recognize and mitigate such positions.

| Movements of power and (re)negotiation
The disproportionate representation of specific entities that occupy the "highest ranks" of knowledge and decisionmaking hierarchies is inevitably linked to authors' understandings of how power moves within and across spatial and jurisdictional levels of resilience. Here, we portray two main ways in which authors present, detail, and probe such movements, illustrating a pronounced juxtaposition.
On the one hand, a propensity to focus on the local as downtrodden, seen in half of the articles that engage with processes of subject making, feeds into the portrayal of communities as "sites of subjugation". In particular, authors emphasize how powerful actors and institutions position local actors as "wrong doers" (e.g., Haverkamp, 2017;Weinstein et al., 2019), "vulnerable and helpless" (e.g., Bertana, 2020;Mirumachi et al., 2020), and "unknowing and uneducated" (e.g., Lindegaard, 2018;Ravera et al., 2019) with respect to their capacities to build resilience in the face of climate vagaries and participate in resilience planning. For instance, Weinstein et al. (2019) show how authorities and planning instruments in India justify the forced removal of residents from informal dwellings by attributing flood responsibility to the poor. Although many of these studies provide pertinent insights into the operation of power through subjectivity (Section 4.3), depictions of uneven relations as solely "pushing in" on marginalized groups and individuals risk delineating power as unidirectional. Such a focus shrouds the fluid nature of power and politics that transcend perceived boundaries between actors and institutions.
On the other hand, a growing body of scholarship foregrounds the capacities of disadvantaged individuals and communities to "push back" on oppressive structures to reclaim meanings of resilience and recalibrate power differentials, as seen in several studies across our sample. This work showcases how disenfranchised actors can leverage the fluidity of power to reconfigure pervasive subjectivities (e.g., Andersson et al., 2018;Blackburn, 2018;Eriksen et al., 2019;Gonda, 2019;Kaika, 2017;Tschakert et al., 2016a). Kaika (2017), for example, highlights how residents in New Orleans utilize community-level campaigns to challenge entrenched notions that urban inhabitants should simply "be resilient" in the face of climate hazards. This and other examples demonstrate how clashes over scale mismatches can create space for questioning the status quo. Tschakert et al. (2016a), in their study of the micropolitics of flood adaptation in India, illustrate how contestations emerge from encounters between actors at different scales often constructed as separate. In their study, community members employed collective learning spaces to contest exclusion and subordination, and elevate themselves as serious actors that deserve consideration and justice. As Wijsman and Feagan (2019, p. 72) argue, paying attention to how contestations of power emerge in practice creates space for marginalized communities to become "critical vantage points for naming and challenging structurally dominant knowledge practices, blind by design to their own ignorance".
Divergent portrayals of the ways in which power travels to subjugate and/or empower specific actors have serious consequences for how resilience is framed in climate change scholarship, and in turn, how resilience is enacted in policy and practice. With respect to unidirectional framings, engaging with power as exclusively "pushing down" a spatial scale promotes a "proximal level bias" that can manifest as an overreliance on resilience building at the individual level. Such biases are denounced by Jones (2019) in her work on food pedagogies in the United States, which illustrates how the burdening responsibilization of individuals for their own health renders healthcare a personal, environmental, and moral pursuit. This conceals racialized power and environmental responsibilities that affect marginalized groups at other interacting scales, such as gentrification and pollution, while reinforcing racist and classist assumptions about the connections between health and food consumption (Jones, 2019). Proximal level biases take time and attention away from a focus on how multidimensional vulnerabilities, as well as capacities for resilience, are constituted through complex and shifting interactions within and across scales.
The obscuring of cross-scalar movements of power also feeds into what can broadly be labeled as a "depoliticization" of resilience that strips a problem of its political complexity (Mikulewicz, 2019). Across scholarship, key resilience concepts and terms such as those related to efficiency (Cast an Broto & Westman, 2020), evidence-based planning (Fainstein, 2015), and community (Buggy & McNamara, 2016) have all been presented as outdated tropes that neglect the "socio-political" in resilience processes and forward vertically imposed solutions. Hence, depoliticizing discourses provide cover for entrenched biases and political agendas, for instance, techno-managerialism associated with a neoliberal political economy (Kaika, 2017;Mikulewicz, 2019), the securitization of climate problems (Mirumachi et al., 2020), and the incorporation of pre-existing norms to limit the space for discussion and dissent (Bastakoti & Davidsen, 2017). Such processes have been criticized for neglecting how resilience is both socially contingent and power laden (Davidson et al., 2016), and eroding notions of (in)equity, (in)justice, and contestation (Mikulewicz, 2019;Scoville-Simonds et al., 2020).
In contrast, studies that delve into the many ways in which subjugated actors challenge and (re)negotiate power can pave the way for reclaiming one's or a community's position and responsibility over resilience within scalar relations. It is at the local level, as Holland (2017) argues, that possibilities for transformational change emerge within resilience processes. Such lines of inquiry are often grounded in situatedness (Section 3.1), emphasizing the everyday in local contexts, rather than monolithic notions of "the global" or "the average". Thus, the lived experiences of communities and individuals "called to be the 'sites' of resilience activities" are foregrounded for questioning, contesting, reworking, or resisting normative resilience activities (Betteridge & Webber, 2019, p. 948). Conceptualizations of resilience as situated help to question the analytical compatibility of certain concepts (e.g., vulnerability, transformation, adaptation, and even resilience itself), often constructed as universal, despite growing acknowledgment of the varied experiences of groups and individuals across dimensions of inequality and geographic and cultural contexts (Wijsman & Feagan, 2019). Still, limiting this scope can engender resistance where there was none previously, especially in communities with high stocks of social capital and expectations of engagement. Mismatches between expectations and priorities across levels and scales can bond members of local societies by prompting collective action and strengthening their position through working with wider links to social, cultural, and economic networks (e.g., Holland, 2017).
Studies that provide the much-needed space to assess how subjugated actors push back on oppressive systems of authority also shed light on the ambiguous, and at times, contradictory nature of power and subject making in resilience, as seen in a handful of articles in our sample. In particular, Eriksen et al. (2015Eriksen et al. ( , 2019 and Gonda (2019) argue that individuals can both uphold and resist multiple axes of power while (re)producing multiple, at times antithetical subjectivities. For example, as Eriksen et al. (2019) show, women in western Uganda utilize climate-smart adaptation activities to both assume their subjugated positions as "supplementary farmers" and acquire more active subjectivities through experimentation with new farming techniques. This and other examples from the literature sample (Anderson et al., 2019;Bezner Kerr et al., 2018;Ciplet, 2017;Ravera et al., 2019) demonstrate that resilience is embroiled within ambivalent and unpredictable processes of power.
In summary, despite the crucial insights the literature surveyed offers with respect to (re)negotiations of power, there remains room to explore how and where alternative knowledge, scalar arrangements, and subjectivities can (and do) reconfigure dominant ways of knowing, understanding, measuring, and securing resilience. Cross-scalar interactions, and especially scalar mismatches, seem to be interesting openings to explore the possibilities for these reconfigurations. Tracing such openings, even if spatially and temporally confined, is crucial for understanding resilience as a fluid and shifting process, one in which new spaces can be cracked open for contesting, (re)negotiating, and reconfiguring the uneven relations that remain pervasive and self-producing in resilience efforts.

| Competing views of power as an outcome or a process
The work detailed above illustrates the significant shift in climate change scholarship to acknowledge and address the fact that resilience is indeed imbued with power. Yet, the ways in which uneven relations are discussed, particularly with respect to scalar configurations and subjectivities, reflect a lingering tendency to focus on seemingly static outcomes of power and resistance, rather than on complex processes of subjugation and empowerment that are perpetually in-the-making. Here, we provide two examples of such normative depictions, juxtaposed with more nuanced examinations that compellingly demonstrate how power in resilience materializes in intersecting and enigmatic ways.
First, a focus on the subject positions of disadvantaged actors, rather than on processes of subject making, is evident in around half of the articles that examined subjectivities. For instance, several studies provide examples of state authorities and powerful community-level actors stereotyping disadvantaged citizens as vulnerable and helpless, without detailed consideration of the material and discursive processes that (re)produce vulnerable subjects and further entrench disproportionate vulnerabilities to climate change. Studies that engage with subjectivities and stereotypes in such a way risk reducing subject making in resilience to an "emergent property" of socio-natural systems. Such oversimplified conceptualizations obscure the inner workings of the socio-political in resilience thinking and privilege the identification of, and response to, biophysical threats.
Several authors, however, provide detailed and nuanced expositions of how subjects are made and remade within resilience efforts, for instance through dominant narratives and the internalization of pervasive subjectivities (see Alvarez & Cardenas, 2019;Bezner Kerr et al., 2018;Eriksen et al., 2019;Gonda, 2019;Jackson et al., 2020;Leap, 2018). As a notable example, Leap's (2018) study examines how societal notions of women as caretakers and men as breadwinners in Sumner, Missouri, perpetuate heteronormative ideologies with respect to who should lead, and be included in, resilience planning and action. Recent work further illustrates how pervasive subjectivities are embodied through the thoughts and feelings of subjugated actors themselves. Jackson et al. (2020) show how rural actors in the Western Province of Papua New Guinea internalize their subject positions as "helpless", thereby reinforcing dominant narratives steeped in historical and colonial marginalization. Such examples provide pertinent insights into the symbolic elements of power that permeate resilience, with very real, material consequences for the actors involved.
Second, a disproportionate focus on the vulnerabilities of certain population groups (often discussed at the community level), rather than intersectional processes of subjugation, contributes to the (re)production of those vulnerabilities. Indeed, scholarly work on the marginalization of groups and individuals in resilience efforts has long embraced intersectionality as an analytical lens (e.g., Sultana, 2010;Thompson-Hall et al., 2016;Tschakert & Machado, 2012;Vinyeta et al., 2016). Yet, particularly in empirical contributions, more than two-thirds of the sampled articles that examined the construction of subjects in resilience dealt exclusively with one dimension of inequality (e.g., Betteridge & Webber, 2019;Kaika, 2017;Ober & Sakdapolrak, 2020;Ramalho, 2019;Weinstein et al., 2019). This can in part be explained by the convenience of simplified binaries between those who are resilient and at risk, privileged and marginalized, and powerful and powerless, simplistic enough to be easily employed in resilience scholarship, policies, and projects. In their discussion of the politics of climate-smart agriculture (CSA) in Uganda, Eriksen et al. (2019) problematize this trend, illustrating how rural households are homogenized through notions of "scraping by", "stepping up", "stepping out", and "hanging in". This superficial typology is not only internalized by farmers but used by local agricultural organizations as a way of exerting authority. A failure to capture how power operates across multiple, overlapping dimensions of inequality risks the use and presentation of caricatures of vulnerability that lead to the inaccurate identification of problems (Gonda, 2019;Osborne, 2015). In turn, resilience scholarship risks misdiagnosing, and contributing to inadequate responses to, the triggering processes that result in inequitable resilience outcomes.
Some notable exceptions can be seen in studies that recognize and engage with intersectional and multidimensional subjectivities (e.g., Eriksen et al., 2015;Gonda, 2019;Osborne, 2015), with some examining particularly complex constellations (Jones, 2019;Leap, 2018). For example, Jones (2019) explains how food pedagogies based on subjectivities related to race, class, and health in Austin, Texas undermine the capacities of Black youth to build resilience to climate change. Moreover, Gonda's (2019) examination of state-level adaptation projects and policies in Nicaragua shows how entrenched ideologies (both societal and institutional) exclude women living in poor rural areas from decision-making processes that mandate access to resilience actions. Such nuanced examinations of the intersectional nature of subjectivities, as noted by Osborne (2015), are imperative for capturing how power operates across multiple, overlapping dimensions of inequality and subject making to constitute distinct experiences with resilience.
With power often stipulated in the sampled literature as either an outcome or a process, or both, it is necessary for critical work to widen its lens for more nuanced assessments of how power operates across different scales and intersectional forms of inequality. This requires shifting the focus from the effects inequity may create or exacerbate, to operations of power determined by relational and shifting processes of interaction. Such a refocusing is preempted by those studies that exemplify the many forms power takes to (re)produce oppressive dynamics in resilience, such as the internalization of dominant narratives and the operation of power across multiple axes of social difference. Continued progress in this space is imperative for advancing scholarly understandings of how power is deployed, contested, and (re)negotiated in the cracks and crevices of resilience-as-as-process.

| CONCLUSION
This advanced review illustrates how the resilience literature engages with power. In particular, it scrutinizes how the mobilization of pervasive subjectivities, cross-scalar arrangements of power, and entrenched biases in the production of knowledge are dealt with in resilience scholarship and practice to advance particular notions of expertise, responsibility, and agency. Despite a growing focus on the politics and uneven outcomes of resilience among critical scholars, our findings demonstrate that power is still often conceptualized as "power over"-relegating local communities to perpetual victimhood. Such insights are consistent with other, albeit limited, analyses of the conceptualization of power in resilience (and by association adaptation) scholarship. For instance, Woroniecki et al. (2019) highlight that "power over" occupies a core focus in the climate change adaptation literature, referring to the process of "one actor […] influencing the behavior of another, and in indirect expressions, where one actor influences the context in which another may operate". The overemphasis on particular actors "holding" power over others risks preserving common tropes and stereotypes that obscure the agency and capacities of historically marginalized groups, reinforcing the need for intervention by external actors who allegedly possess the knowledge and tools to front resilience processes. Such dynamics are also visible in the composition of research teams and directionality of research efforts whereby the Global North secures its place at the epicenter of climate change knowledge production while the Global South continues to be subjugated as a site for empirical verification. As Woroniecki et al. (2019, p. 7) argue, "the risk is that [power] is perceived as immutable and owned by different actors, rather than as determined through relationships that can be shaped through collective processes". Seeing power as immutable can seriously hamper capacities for envisioning alternative, transformative futures of resilience.
These persistent blind spots notwithstanding, our review also provides important insights on more progressive and radical research engagements-or "bright spots"-that advance nuanced understandings of how power is articulated, reproduced, and challenged within and through resilience processes. These include careful and caring approaches that foreground everyday, experiential, and emotional knowledge, lending themselves to the decentralization of "knowledge hegemonies that characterize contemporary thought in climate change governance" and decision making (Cast an Broto, 2020, p. 242). Equally as important are authors' illustrations of the cross-scalar interactions that challenge disembodied accounts of resilience and their tendency to homogenize unevenly positioned groups; they reflect a focus on "power to", also described by Woroniecki et al. (2019) as prevalent in studies of power and adaptation, albeit with varied nuance and complexity. In contrast, our review finds that examinations of resistance and the recalibration of power differentials remain underdeveloped. Nonetheless, we want to acknowledge, and draw inspiration from, those approaches that probe the multidirectional, interdependent, intersectional, and ambiguous ways power operates to support and/or hinder resilience building in place, calling attention to local and decolonizing perspectives, crosscultural alliances and solidarities, and collective action. In the process, they recalibrate harmful subjectivities by demonstrating how seemingly powerless actors can and should be conceived as agents of change.
Even if incipient, this body of work provides promising new avenues to contest the unequal relations of power that continue to permeate resilience efforts. It also invites more nuanced investigations of resilience-as-a-process, highlighting our roles as reflexive resilience researchers in reconciling resilience with the urgent need for transformational change. There is growing recognition that equitable and just resilience entails emancipatory approaches that recognize colonial legacies of exclusion, domination, and exploitation, and incorporate the needs, values, and desires of historically marginalized groups in framing resilience and climate adaptation more broadly (Klenk et al., 2017;Porter et al., 2020;Roberts & Pelling, 2020;Veland et al., 2018). Here, recent critical and decolonial approaches to adaptation provide entry points to expose how resilience is undermined by reductionist hegemonic framings and how more pluralistic engagements can promote the co-production and co-existence of multiple climate realities, with attention to emplaced power asymmetries Goldman et al., 2018;Johnson et al., 2021;Nightingale et al., 2020).
Such invigorating entry points provide the foundation from which we offer three overarching recommendations for advancing the conceptualization of power in resilience scholarship. First, we suggest that more earnest confrontations with the various points of privilege and marginalization of the diverse players involved, specifically researchers and participants, are crucial to avoid perpetuating power inequities and to bolster the capacities of disadvantaged individuals and societies as agents of change (Owens et al., 2018;Searle & Muller, 2019). Second, unidirectional framings of power should be replaced with more nuanced understandings that demonstrate power hierarchies are not set in stone, but can be challenged, with people and communities "pushing back" to reclaim meanings of resilience and recalibrate pervasive subjectivities (Cobarrubias, 2020;Kythreotis et al., 2020). Third, concerted efforts are needed to continue advancing engagements with enigmatic processes of power, such as the internalization of normative ideologies and the operation of power across multiple axes of social difference, to avoid neglecting the ever-changing "messiness" of power in resilience (Gonda, 2017;Sultana, 2020). We view these steps as essential for critical scholars to embrace, support, and come to grips with the fluid realities from which resistance to oppression and transformational change emerge (Nightingale et al., 2020).
Although the recommendations above may well be compelling from a conceptual as well as theoretical perspective, the methodological angles of how to examine and overcome power asymmetries in daily resilience building remain ambiguous and the harnessing of scholarly advances for practical achievements challenging. There is an abundance of empirical work that demonstrates how citizen participation in climate resilience decision making can go unquestioned (Cast an Broto & Westman, 2020) and how community-led committees and other seemingly participatory engagements often fail to be inclusive (Koslov, 2019;Woroniecki et al., 2019) or reinforce existing hierarchies across and within scales (Buggy & McNamara, 2016), for instance via elite domination (C ordoba Vargas et al., 2020;McDonnell, 2020). To overcome such pervasive shortcomings, recent methodological advances in resilience scholarship underscore the need to nourish political capabilities and political spaces; the aim is to actively contest exploitative and oppressive power dynamics while mobilizing alternative, emancipatory subjectivities that are needed to negotiate inclusive and equitable resilience in practice and deliberate possible transformative pathways (e.g., Ensor et al., 2021;Matin et al., 2018). Along similar lines, as Wijsman and Feagan (2019) suggest, a practitioner's commitment to positionality, situatedness, and responsibility in decision-making processes is better suited to help trace the intersecting complexities of power than misguided or disingenuous participation.
The many encouraging efforts to nourish political capabilities notwithstanding, genuine inclusiveness requires a constantly watchful eye to discern and address, early on, deep-seated power hierarchies and potentially exclusionary modes of governance. Even when employing methodologies that embrace co-learning, such as cooperative inquiry or participatory action research, we as researchers and practitioners need to commit to iterative and agency-enhancing approaches, the validation of multiple knowledges and worldviews, and the fostering of deliberation and solidarity with disadvantaged populations. The ability to disrupt patterns of privileging does not come from anywhere; it demands candid and courageous self-reflexivity and mastery, in the everyday spaces of resilience practice.

CONFLICT OF INTEREST
The authors have declared no conflicts of interest for this article.

DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analyzed in this study.

RELATED WIREs ARTICLES
Power, realism, and the ideal of human emancipation in a climate of change Adaptation and resilience: responding to a changing climate Resilience implications of policy responses to climate change Resilience isn't the same for all: Comparing subjective and objective approaches to resilience measurement The framing of power in climate change adaptation research