Migration as climate adaptation? Exploring discourses amongst development actors in the Pacific Island region

This paper investigates the perspectives of a set of actors devoted to development in the Pacific on climate change, migration, and adaptation. While much of the debate over climate and migration is centred around the Small Island Developing States in the Pacific, little is known about how the debate is articulated at that regional level. Drawing on poststructuralist discourse theory and using semi-structured interviews with a set of development actors working in the region, the paper discerns three distinctive discourses on climate and migration. These are (1) a main discourse that promotes international labour migration as an adaptation response and two alternative discourses that challenge the main discourse’s views, by suggesting (2) that migration is of marginal importance and engagement with socio-economic factors that influence Pacific Islands' vulnerability is more pressing, and (3) that out-migration is undesirable but that communities may have to be relocated within their countries. The paper further explores why the discourse on labour migration may have emerged and why it is being perpetuated by actors that originate outside the Pacific region. The paper concludes by suggesting that significant differentials in economic and political resources exist between the main discourse and the alternative discourses. In addition to these empirical insights, the paper adds new findings to the growing literature on the politics of climate migration discourses. Unlike earlier work that identifies a shift from an alarmist to an optimist framing, it illustrates that both alarmist and optimistic imaginaries operate simultaneously in the discourse on labour migration.


Introduction
One of the most iconic images of climate change is that of people moving from their homes. Debate around so-called climate migration emerged in the 1980s and has since captured interest across academic, policy and popular debate (Oels 2005;Vlassopoulos 2013;Bettini 2014;Ransan-Cooper et al. 2015). This debate is particularly centred around the Small Island Developing States (SIDS), including those in the Pacific. Described as among the most vulnerable regions on the planet (Nurse et al. 2014), the Pacific Island Countries (PICs) are regularly portrayed, in news media and academic texts alike, as future locations of large-scale outmigration as seas rise and coastal settlements become uninhabitable. In this context, Smith and McNamara (2015) speak of a 'geopolitical characterisation' of Pacific nation states as inherently vulnerable. In other words, the Pacific holds a key position in the global imaginary around climate migration.
Diverse institutions including the International Organisation for Migration (IOM), the World Bank, Greenpeace, and the United Nations Development and Environment Programmes (UNDP and UNEP) have begun to engage with the topic through the facilitation of background studies, publications, conferences and policy recommendations (see e.g. Brown 2008;Hamro-Drotz 2011;Wodon et al. 2014;Bedarff and Jakobeit 2017;Opitz Stapleton et al. 2017). Important international donors such as the European Moreover, the diversity of PICs in terms of island geography, settlement structure and economic development makes generic categorisations of the future uninhabitability of islands problematic (Barnett and Waters 2016).
To date, most academic and policy literature on climate migration has focused on explaining or quantifying the role climate change has on the movement of people (Warner et al. 2010;Warner 2011), sought to illuminate interconnections between mobility, environment and social change (Morrissey 2013;Ransan-Cooper 2016); engaged in discussions on the correct terminology (McAdam 2012;Faist and Schade 2013); raised normative concerns about the establishment of a climate refugee protocol Boas 2008, 2010); and considered the implications for destination regions (Gemenne and Blocher 2017).
A growing body of poststructuralist work takes a different approach. 2 Poststructuralists have utilised the lens of discourse to scrutinise how different assumptions, political interests and the means to which they are employed play a role in how climate migration discourses (CMD) are produced, circulated and interpreted. For example, Kothari (2014:133) suggests that a 'wide range of actors, including scientists, government bodies, nongovernmental [sic] agencies and activist groups, are involved in the ongoing production of the discourse, delimiting the boundaries of the debate and producing shared meanings across different scales'. Most of these important studies have, however, tended either to engage with CMD at a theoretical/conceptual level (Felli 2013;Baldwin 2017) or derive their empirical analysis from debates and policy texts at the international level, for instance, within COP negotiations (Bettini 2014;Methmann and Oels 2015;Oels 2015;Ransan-Cooper et al. 2015;Bettini et al. 2016;Ober and Sakdapolrak 2017;Rothe 2017). Fewer studies have undertaken more localised empirical research to capture ideas about climate change and migration at the national level in SIDS, including the Pacific. These studies have focused on local perspectives, national governments, nongovernmental organisations (NGOs), donor agency discourse, and how Islanders view the existing news media coverage on climate impacts for SIDS (Mortreux and Barnett 2009;Farbotko and Lazrus 2012;Dreher and Voyer 2015;Farbotko et al. 2016;Hermann and Kempf 2017). Outside the Pacific region, Kothari (2014), Arnall and Kothari (2015) and Stojanov et al. (2017) have examined resettlement policies and perceptions of climate migration in the Maldives. Yet, there remains an empirical gap when it comes to the regional level; we know very little about how climate change and migration are discussed in regional development discourse, among organisations devoted to development, even though these organisations are influential in both policy setting and in structuring aid programmes on the ground for SIDS, including those in the Pacific. Beyond the provision and allocation of funding, these organisations are active and vocal agents in imagining, planning and implementing responses. Particularly in SIDS, they have been found to exert influence on regional scientific, political, business and media agendas, such as in the formulation of development and climate adaptation plans (Arnall and Kothari 2015;Denton 2017). This type of influence is not unique to the Pacific, but it is especially acute here because nine of the top 20 most aid-dependent countries globally are PICs (The World Bank 2019). 3 This absence of regional level studies marks the departure point for this paper. Taking the regional hub for development organisations in the Pacific, Suva in Fiji, as an entry point, I examine the way in which climate change and migration are articulated, specifically in relation to climate adaptation, by development actors working in the region. Specifically, I draw on 17 semi-structured interviews with a set of actors engaged in development work in the Pacific. Under the term development actors, I group various actors engaged in the Pacific's regional climate and development community. These include international actors such as donors, bilateral and multilateral development organisations and NGOs, regional and national bodies and CSOs, as well as academic institutions. 4 By regional hub I mean the central role Fiji's capital Suva plays in development work across the Pacific region, with many of these organisations' regional headquarters located there.
1 Two prominent research initiatives funded by the EU and its Commission are the EACH-FOR (Environmental Change And Forced Migration Scenarios) and the MECLEP (Migration, Environment and Climate Change: Evidence for Policy) projects, which received approximately EUR 681846 and EUR 1.9 million, respectively. 2 Poststructuralism is a late-twentieth-century development in philosophy that comprises a variety of reactions by predominantly French philosophers to structuralism. Poststructuralists emphasise the instability of meaning and contingent construction of the world and are interested in the relationship between power and systems of knowledge. The focus of analysis is commonly on discourses and the different ways people make meaning from language.
Developing regional solutions to shared environmental and developmental problems across PICs is a common practice and evident in both concrete policies, such as the Framework for Resilient Development in the Pacific (2017)(2018)(2019)(2020), and in the existence of governance bodies that operate on a regional scale, such as the so-called Council of Regional Organisations in the Pacific (CROP). This makes the analysis of regional development discourse in the Pacific highly relevant.
In the following, I first briefly outline the paper's theoretical framework, which builds on the work of Laclau andMouffe (2001 [1985]), and explain how this was used to analyse the empirical material. Using interviews undertaken with actors devoted to development in the region, I then discern three distinctive discourses on climate migration and describe the assumptions underlying them. These are a main discourse on (1) labour migration as adaptation, in which the PICs' vulnerability is intrinsically linked to their physical geography as well as climate change, and which sees migration as inevitable and international labour migration as a proactive adaptation strategy; another (2) migration is possible but of marginal importance, in which PICs' development challenges and vulnerability stem from a wide range of socio-economic factors, including colonisation and patterns of contemporary globalisation; and (3) migration as anathema, which argues that PICs' vulnerability is caused by climate change alone but that out-migration is undesirable as an option. What is important to note right away is that the two alternative discourses appear to have emerged largely in response to the main discourse and are therefore less elaborate. I further explore why the main discourse that links labour migration to adaptation may have emerged and why it is being perpetuated, from and by development actors that originate outside the Pacific region. I conclude by suggesting that significant resource differentials exist between actors who engage in the main discourse and those in the alternative discourses, and that the former has been appropriated in ways that match the regional social and political climate and align with development organisations' earlier mandates. The paper's findings contribute to the growing literature on the politics of CMD.

Research framework and approach
There are different approaches to the study of discourse. This paper draws on Poststructuralist Discourse Theory (PDT). Developed by Laclau and Mouffe (2001) in the 1980s, PDT has received little attention within climate migration research (with the notable exceptions of Methmann 2010; Methmann and Rothe 2012;and Bettini 2015). For discourse theorists the notion of discourse goes beyond mere rhetoric or representation to include all social practices and relations. Importantly, discourses do not merely describe or reflect a reality that is 'out there' but construct and produce it as reality in the first place. Discourses establish a picture of how things are and what needs to be done, and in this way, they not only impose limits on what can be thought or said but also draw the boundaries around what people perceive as rational policy responses (Torfing 2005). These premises have implications for research on climate migration. First, this perspective views climate migration first and foremost as a discursive phenomenon, which is to say that the relationship between climate change and migration is socially constructed. Second, it contends that discourses among development actors are productive political sites where meanings and interpretations are formed and negotiated. Third, an analytical perspective based on discourse theory is not interested in 'truth claims' of each discourse or their 'true applicability' to PICs. Rather, I focus on how ideas about climate change and migration are conceptualised and by whom, how they are structured, and what political ends they pursue.
To study discourses, understand and critically explain their internal workings, I draw on three explanatory categories, referred to as logics, which are simultaneously operative in a discourse (Glynos and Howarth 2007; see also Remling 2017 for a more detailed description of the methodology). These are social, political and fantasmatic logics. Social logics, describe norms of behaviour that structure a certain discourse and capture the assumptions that are taken for granted and considered to be unproblematic in that discourse. I focus my analysis of social logics on (i) assumptions about the Pacific's present economic, social, and environmental situation, as well as its future, (ii) general factors that are seen as creating social vulnerability, (iii) recommended courses of action, and (iv) the general view on migration.
Political logics capture the constant and dynamic change in discourses and their inherent struggle to fix meaning. They indicate where the unsettled borders of a discourse or between competing discourses lie. Political logics do something to the social logics considered to be worthy (or in need) of contestation or defence (Glynos et al. 2014). In this paper, I explore moments of tension, contradictions and spaces that appear to be unsettled in the discourse to denote political logics. These might be incidents where interviewees question the assumptions of others, attempt to defend tensions in their own argument, or where different interviewees seek to fill the same term with incompatible meanings.
Fantasmatic logics, account for the emotional, non-rational investment of subjects in a particular discourse (Glynos and Howarth 2007). This is based on the premise that in order to understand discourses, it is insufficient to study only the form of language; it is also necessary to examine what this language moves in people. Applied analytically, fantasmatic logics capture that through which a discourse speaks to its audience and thus derives its affective force. Such fantasies might mobilise hope by promising some future enjoyment (the beatific dimension) or mobilise fears or anxieties through dystopian imaginaries (the horrific dimension). To apply these logics as a heuristic framework for analysing CMD in the Pacific, I pose a series of questions to my interview material as presented in Table 1.

Methodological strategy and empirical material
The analysis of discourses on climate migration within the Pacific region is based on fieldwork carried out between November 2015 and January 2016. I conducted 17 semistructured interviews with 18 people from different local, regional and international NGOs, donor agencies and academic institutions engaged in development work (see Table 2, names of interviewees are withheld to ensure confidentiality). Fourteen interviews were conducted in Suva. Beginning with contacts made through my previous work on community-based development in Fiji, interview participants were identified by snowball sampling. Four additional interviews were conducted in Australia with participants from the Australian National University, the Migration Council Australia, the University of Sydney and an independent journalist who has been working on development questions in the Pacific for over a decade and has authored reports for Oxfam and other organisations. These additional interviewees had repeatedly been referred to by different participants in Suva; therefore, they seemed to belong to the regional development discourse in which I was interested. Interviews centred on the interviewees' understanding of the relation among climate change, migration and adaptation in the Pacific. In addition to the interviews, I observed two meetings of development organisations during the same time period (at the regional office of the United Nations and the Pacific Islands Forum Secretariat). These meetings added context to the interview material and further clarified different interviewees' positions.
The selection of interview participants is part of a much larger landscape of organisations and actors. Specifically, perspectives from national governments and NGOs are missing, and as interviewees centred mainly around the hub of Suva, they do not represent all the diversity of perspectives across the Pacific development landscape. 5 Nevertheless, I suggest that my sample represents an important set of regionally significant development aid actors and is a useful indicator for an exploratory study of the ideas about climate change, migration and adaptation that are circulating in different organisations in the region.
Interviews were recorded through extensive note taking with the informed consent of the participants. 6 Transcribed interview notes were coded manually with the help of MAXQDA data analysis software. For coding, the previously introduced social, political and fantasmatic logics were used as heuristic devices. Coded segments were extracted and then clustered around converging themes, and finally commonalities and differences among the interviewees were identified. It should be noted that the people interviewed do not necessarily reflect the official stance of their respective organisation.

Social logics
What is the interviewee's understanding of the relation among climate change, migration and adaptation?
What is described as the core cause for concern, what creates vulnerability in the context of climate change?
How does this understanding relate to migration?
What is the proposed response? Political logics What does the interviewee challenge or defend?
Where are boundaries to 'inappropriate' opinions and responses drawn? Fantasmatic logics How is the particular view on climate and migration sustained and justified?
What images are employed to make his/her perspective seems appealing?
What consequences for Pacific Islands does the interviewee promise or warn against?
What vision of a better or worse future is projected? 5 (Other organisations were contacted but not available for an interview, including (in alphabetical order) the Climate Change Division of the Fijian Government, Diverse Action for Voices and Equality (DIVA), the Foundation of the Peoples of the South Pacific International (FSPI), the German Development Cooperation Agency (GIZ), the Pacific Islands Association of Non-Government Organizations (PIANGO), the Pacific Islands Development Forum (PIDF), the Pacific Youth Council, Partners in Community Development Fiji (PCDF), the Secretariat of the Pacific Regional Environment Program (SPREPwhich hosts a seconded staff member of the Nansen Initiative), and the United Nations Development Programme (UNDP)). 6 Audio recordings appeared to make the participants uncomfortable, wherefore this method of capturing interviews was abandoned after the first couple of interviews.
Exploring three climate migration discourses amongst development actors As noted, debate over the relation of climate change, migration and adaptation at the international level has been prevalent for decades. Yet, the interviewees suggested that it only recently emerged in discussions among organisations devoted to development in the Pacific. 7 Rather than listing all the different understandings offered by the interviewees, and acknowledging that the analysis of discourses is interpretative, in the following I cluster the views expressed into three distinct discourses. By 'clustering' I mean that the three discourses demarcate overlapping key positions, but I do not suggest that they are entirely coherent or mutually exclusive.
Discourse 1: a discourse on managed labour migration as adaptation The most prominent discourse-in terms of the number of interviewees affiliated with it (n = 7), the coordination and linkages among these interviewees, and their financial and agenda-setting capacity-comprises representatives from the UNESCAP, the ILO, the PIFS, Oxfam Australia, the EU's Delegation for the Pacific, and the Australia-based ANU and MCA (see Table 2 for an explanation of the abbreviations). This discourse's basic premise is that PICs are small and incapacitated. Associated with factors such as 'smallness', 'remoteness' and 'living in a hazardous region', the islands are viewed as having an inherent fragility that inhibits socioeconomic development. Several proponents of this discourse conceptualise small PICs as so-called MIRAB economies: intrinsically poor in natural resources, offering few opportunities for development or economic self-sufficiency, and thus destined to rely heavily on external aid and remittances. 8 Following this conception, social vulnerability is seen as inevitably rooted in the region's geographic features and is therefore considered to be an inherent precondition of PICs. People on the islands are thought to have little capacity or agency to adapt to climate change, at least not sufficiently to meaningfully counter its impacts. Consequently, this discourse suggests that people will be forced to move from their islands, which will create a humanitarian crisis that risks becoming traumatic for Pacific peoples. Accordingly, this discourse sees climate-induced migration as a natural progression of climate impacts and, therefore, as almost inevitable.
To avoid, or at least ameliorate, such unplanned migration and associated negative consequences, the interviewees suggest that people should begin migrating now to provide labour to nearby countries. As a means to generate remittance flows to the region, such orderly, well-managed and early labour migration is seen as leading to improved livelihoods and thus 7 A notable exception is the PCC, whose members initiated a discussion on climate change and migration already in 2007. However, the focus of this debate was the protection of forced climate migrants and the possibility of resettling entire communities (Pacific Church Leaders 2009). 8 MIRAB stands for migration, remittances, aid, and bureaucracy, and is an idea that originates with Bertram and Watters (1985). Originally introduced as a refinement of mainstream development practice and to give more emphasis to context, its associated ideas of 'smallness' and 'island vulnerability' have been extensively critiqued as belittling, paternalistic and disempowering (see Hau'ofa 1993;Farbotko 2010  development on the islands. This, in turn, is expected to act as a buffer to improve the resiliency of home communities and enable action on adaptation. In the long term, labour migration is expected to help develop Pacific Islanders' skills and establish connections to overseas places, potentially becoming a steppingstone for more permanent migration in the future. In a nutshell, this discourse hopes to forestall the large-scale climate-induced migration of many, by facilitating the early and voluntary labour migration by selected few. This call for enabling labour migration should not be mistaken as support for uninhibited movement. On the contrary, promoters make clear prescriptions about where Pacific Islander labour should be deployed, namely, to Australia and New Zealand (the so-called Pacific Rim). They also envision a specialised form of mobility, predominantly unskilled and lower-skilled labour migration in a circular or temporary fashion that builds on existing migration pathways, for instance in the horticultural agriculture and tourism industry. Interviewees also emphasise the importance of rigorous eligibility criteria for temporary work visas that must align with the interests and demands of recipient countries. Notably, migration as a form of self-help development is viewed as an opportunity and form of successful adaptation, provided it is promptly addressed and properly managed through international labour mobility schemes.
The discourse assumes that Pacific people, communities and governments will see the benefits of migration if given the right information. Assessing the need for and proposing suitable schemes is seen as the responsibility of international organisations. Governments also play a role but only the destination-country governments of Australia and New Zealand, which decide who can enter their territory, on what basis, for what duration and with what attendant rights, benefits, and obligations. Pacific Islanders' role in the process is reduced to mere recipients of help.
In regard to political logics, two key internal tensions within this discourse are evident. The first tension arises between the humanitarian imperative to assist the most vulnerable and the reality that the likely beneficiaries of the proposed migration schemes will be a small group of young and better educated people (see also points made by Discourse 2 further below). Different arguments are used to reconcile this tension around an 'elite bias' and defend the promotion of labour migration. One interviewee suggested that it is simply a matter of getting the schemes' setup and the selection processes 'right'. Another interviewee who identified strongly with the labour migration discourse contended that as 'communist, tribal societies' (Interviewee #9), Pacific Islanders would justly distribute remittances, thereby re-allocating benefits and balancing any elite bias. Two other interviewees associated with the discourse simply accepted this incompatibility as unavoidable; one suggested that a humanitarian imperative might actually be a hindrance in selecting suitable workers, and the other interviewee (based in Australia) drew attention to the fact that Australia would design policies that primarily benefit Australian businesses, not Pacific Islanders. A second tension concerns uncertainty. Most interviewees not only acknowledge that climate models contain some uncertainty and lack accuracy for the Pacific region but also that disentangling migration decisions based on climate change from other factors is difficult if not impossible. Officially, no interviewee suggested a direct causal relationship between climate change and migration; however, the proponents of Discourse 1 nevertheless insist that migration is not only plausible but a highly likely future scenario.
Discourse 1 exhibits an interplay of two fantasmatic scenarios that speak to both fear and desire. It invokes dystopian imaginaries to suggest that climate-induced migration is seemingly imminent and that such unplanned migration will lead to a humanitarian crisis in the region. In this way, and despite the absence of connotations of islanders as 'victims' or 'refugees' and 'hard' security implications (see commonalities among the three discourses further below), the discourse implicitly renders 'Pacific Islanders as vulnerable victims, as "climate refugees" in waiting' (McNamara and Gibson 2009). Set against this spectre of desolation, the discourse moves on to provide a source of hope: that proactive, voluntary and planned labour migration to Pacific Rim countries will prevent forced, unmanaged and unplanned migration. This beatific fantasy not only promises the pre-emption of climate-induced migration but also offers economic prosperity and a 'triple win' for everyone involved. For Pacific communities, prosperity results from the reduced pressure on local resources and ecosystems and the diversification of household income through the inflow of remittances. For the 'adaptive migrant', prosperity results from actively choosing his/her fate. Finally, for the country receiving migrants, prosperity results from the ability to fill labour shortages. Consequently, the discourse essentially confronts Pacific Islanders with two options; either migration forced by climate catastrophe in the (near) future, or seeking labour migration opportunities in the present. Remaining on the islands is not portrayed as a viable future option.
While Discourse 1 is the most elaborate and established discourse, it is critiqued by two distinctly different discourses. Rather than being coherent discourses in and of themselves, these appear to be constituted largely in relation (and opposition) to Discourse 1.

Discourse 2: migration as possible but of marginal importance
The second discourse contends that the current challenges facing PICs are irreducible to climate change or their smallness and remoteness. Interviewees representing this group come from the IUCN, UN Women, and the SPC and include four researchers and the journalist (n = 8 in total). Proponents centrally point to many different sources of social vulnerability among PICs rather than assume that vulnerability is primarily geographical and climatic in nature. PICs are not seen as inherently vulnerable, rather, the region's challenges and vulnerabilities are connected to socio-economic and political factors including the processes of colonialism and globalisation that have and continue to disrupt social structures and economies. Furthermore, the interviewees point out that local activities and unsustainable development practices in PICs significantly contribute to their vulnerability.
On migration, the interviewees express a more ambivalent perspective than Discourse 1. They acknowledge that migration can lead to positive livelihood outcomes and that it has been a livelihood diversification strategy in the Pacific since humans first settled on the islands, but caution that positive outcomes are by no means a given. Specifically, interviewees raise three concerns regarding the promotion of out-migration. First, migrants themselves might face adverse consequences, with them ultimately becoming more vulnerable than they were before moving, through the loss of community, culture, traditions, land rights, and livelihoods, as well as the difficulties of adjusting to and integrating into new places. Second, migrants might find themselves in more climate-vulnerable locations or in positions of economic exploitation. Finally, interviewees raise concerns over adverse consequences for the communities left behind. Losing the able-bodied working population may mean that communities struggle not only to meet their collective labour needs but also to uphold important cultural and social functions of a village, which might eventually erode the local capacity to respond to climate impacts and therefore be maladaptive.
Another related concern raised by interviewees is that proposals for labour migration ignore that some people might not be willing or allowed to enter labour migration schemes. Being socially selective, such schemes will not be accessible to everyone equally, especially not 'the most vulnerable' who lack access to information and the needed economic and social capital to migrate, which raises questions over what will happen to people too young, old, poor or uneducated to move (see Black et al. 2011 for a related discussion of 'trapped populations'). These concerns over an elite bias directly challenge the policy response proposed by Discourse 1. Interviewees representing Discourse 2 imply that other strategies would be more suitable to address development deficits and vulnerabilities and, thereby, better prepare the region for climate change. They suggest addressing unsustainable development practices and urgent problems such as high population growth and rapid urbanisation trends in PICs first, before considering out-migration. Labour migration to foreign countries is not considered to be a principal, or particularly favoured, adaptation option. More critical of outsiders' interventions than Discourse 1, they also suggest evaluating and revising current development paradigms and decision-making patterns, and to enable more Pacific self-determination in the development of suitable policy responses.

Discourse 3: migration as anathema
A third discourse was articulated by two members of CSOs (n = 2). Although it is the most minor of the three discourses, it reflects the only grassroots organisations interviewed and is distinct from the other respondents' views; thus, it is worthwhile highlighting.
This discourse sees climate change, caused by the emissions of developed countries, as a major new threat to the region that will have significant negative consequences. Social vulnerability for this discourse is directly and primarily prompted by climate change. Similar to Discourse 1, proponents of this discourse also see climate change as a direct driver of migration and take for granted that it will lead to dislocation of Pacific Islanders in the future. However, viewing migration largely unfavourably and opposing overseas migration, interviewees suggest that out-migration would be a measure of last resort and were vehement in stressing this as an option only after all other adaptation options have been exhausted. Instead, they see it as necessary to prepare for the planned short distance relocation of entire communities within PICs, driven by national governments. Such relocation is seen not only as being less disruptive to people's lives but also more in line with local customs and traditions. Furthermore, against the (then) upcoming COP 21 in Paris and drawing on wider debates around climate justice, interviewees strongly contended that PICs have a right to financial compensation through international mechanisms such as the UNFCCC Loss and Damage mechanism.

Interpreting the three discourses and the role of development actors therein
There are a number of commonalities across the three discourses. Most notably, all three exhibit little endorsement of ideas around 'climate refugees' or any association with a security lexicon typical of global debates on climate migration (see Oels 2015;Ransan-Cooper et al. 2015;Bettini et al. 2016). Migrants are never framed as being a potential source of conflict or insecurity in the Pacific or elsewhere. When interviewees allude to security implications, these are rendered a problem of human security rather than an issue of hard national or military security. This avoidance of unpopular notions of climate victims and refugees (see McNamara and Gibson 2009;Smith and McNamara 2015) might help secure the compliance of relevant parties and to, thereby, develop a regionally acceptable migration discourse.
Another shared trait is that interviewees across the discourses refer to Anote Tong's aforementioned call for 'Migration with Dignity'. By doing so, they not only create an association with this widely known slogan but also define its meaning in a way that closely relates to their own perspective. The term is being used by the participants who strive for labour migration as adaptation, by those who despise such proposals and emphasise the many other contextual challenges that need attention, and by those who call for planned community relocation. In other words, different interviewees articulate and attempt to define 'Migration with Dignity' in their own particular way. This can be interpreted as an indicator of a discursive field that is still unsettled.
The greatest differences between the discourses are threefold. First, while all converge around the idea that the Pacific is faced with significant challenges in regard to future climate impacts, the root causes of these challenges are understood as fundamentally different. The discourse on labour migration reduces island vulnerability largely to the 'natural' geographical features (i.e., MIRAB + climate change = problem), alternative-Discourse 2 emphasises historical and current, political and economic influences that render Pacific communities vulnerable and Discourse 3 blames climate change alone. Second, the interviewees express different views on migration and its potential as an adaptation measure. Whereas the main discourse is centrally concerned with reducing climate risk by managing and promoting labour migration to the Pacific Rim, Discourses 2 and 3 display a marked scepticism towards these central tenets of the discourse. Both firmly reject the proposition that labour migration would be a favourable adaptation strategy in the interest of Pacific Islanders and instead encourage a critical engagement with current development practice and the political and economic situation of PICs (Discourse 2) and community relocation (Discourse 3), respectively. Third, the difference between Discourse 1 and the alternative discourses becomes most evident when considering the underlying fantasmatic logics. While the discourse on labour migration articulates clear horrific and beatific scenarios to evoke affective hold, I find no employment of fantasmatic images in the alternative discourses, which suggests that in their current form they do not provide attractive anchor points for identification. This might be the case because, as noted, Discourses 2 and 3 seem to be constituted largely in response to the discourse on labour migration, which they seek to challenge, rather than being 'stand-alone' discourses. Given these differences, the remainder of the paper critically examines the discourse on labour migration as adaptation in more detail. It further explores, what gives this discourse its rhetorical power and what might be observed about the actors that engage in it.

Echoes of earlier development discourse, outside-driven agendas and discursive opportunism
The suggestion that labour migration can be an adaptation strategy is neither new nor unique to the Pacific region (Opeskin and MacDermott 2009;Felli 2013;Vlassopoulos 2013;Bettini 2014;Ransan-Cooper et al. 2015;Baldwin and Fornalé 2017;Ober and Sakdapolrak 2017). In the global debate, this idea has been advanced by a range of policy and research actors (Barnett and Webber 2010;Foresight 2011;Black et al. 2011;Asian Development Bank 2012;The World Bank 2016;Melde et al. 2017). In a recent article, Gemenne and Blocher (2017:2), for example, suggest addressing 'managed migration as a new tool for climate change adaptation policy, and one that transforms mobility into a positive exercise'.
The discourse echoes earlier development discourses about the Pacific. By building on the MIRAB model outlined above, it reproduces long-standing ideas that render PICs economically and environmentally fragile. The proposed policy solutions to ameliorate this fragility-managed low-skilled labour migration to Australia and New Zealand-replicate arguments made a decade earlier by the World Bank (2006). While those engaging in Discourse 1 associate a progressive meaning to their call for labour migration as an adaptation strategy, it is really a conservative discourse that (1) assimilates climate change into pre-existing perceptions and understandings about the Pacific and its (lack of) development, and (2) 'couples' adaptation with an existing development discourse so that in this model, adaptation embraces and (re)legitimises established ideas and takes the same pathways as previous suggestions for developing and modernising the Pacific (see also Barnett and Waters 2016). Similar to discursive changes in other regions, existing ideas become re-labelled as climate related (Jinnah 2011;Nagoda 2015;Webber 2016).
Strikingly, most interviewees grouped under Discourse 1 (except for two based in Australia) are to some degree connected to an EU-funded project called the Pacific Climate Change and Migration Project 'Enhancing the Capacity of Pacific Island Countries to Manage the Impacts of Climate Change on Migration' (or PCCM for short). This EUR 2.4 million three-year project ran from 2013 to 2016 and was implemented jointly by the UNESCAP, the ILO and the UNDP. The first of its kind internationally, the project emphasised the need for increased migration by actively 'working to address the issue of human mobility and displacement of Pacific countries impacted by climate change' (UNESCAP 2016). Established as a development and research project, it promoted a very specific perspective, namely, 'that migration is not only a last resort coping mechanism to climate change, but can also be a valuable strategy to help diversify and increase household income and thus improve resilience' (UNESCAP 2014). According to the interviewees directly associated with the project, the idea for it was conceived in response to a funding call issued by the EU (European Commission 2010), and it is evident that the project's core ideas emanate directly from this call. 9 While the EU's engagement with climate and migration through the funding of the PCCM project has not yet led to the implementation of concrete measures or legislation, the PCCM was directly involved in the development of national labour migration policies in Kiribati and Tuvalu, both adopted in 2015 (see also Fornale and Kagan 2017). Furthermore, given the EU's political and economic weight in the region as a major donor, its discursive influence over this discourse should not be underestimated. 10 Given these observations, it seems fair to say that the discourse on labour migration as adaptation was neither initiated nor driven by people from within the region, Pacific Island governments or local organisations. Rather, it was imagined by foreign actors and organisations and entered the regional stage through the channels of international development organisations.
Why might this group of actors take it upon themselves to lead and feed the discourse about labour migration as adaptation? Why or for what purposes is this discourse mobilised by these organisations? Regarding the EU's involvement, proponents of Discourses 2 and 3 offered three suggestions for why it might be interested in funding projects that promote labour migration. These point to geopolitical motives. First, they suggested that it might be a strategy for splitting up negotiations for Loss and Damage provisions under the transnational Alliance of Small Island States (AOSIS), which has lobbied internationally for the financial compensation for victims of climate change (see also the position of Discourse 3). Second, interviewees suggest that by stepping in to 'assist' PICs the EU makes a claim to political leadership in the Pacific, which might serve to weaken Australia's regional influence. Another, third, explanation advanced by one interviewee is that in the face of a perceived future threat of migration to Europe, the Pacific acts as a testing ground for new policy solutions. This suggestion links to what Farbotko (2010) calls 'island laboratories', where the Pacific functions as an experimental space in which ideas, actions and policy solutions can be tested and trialled. This is evidenced by scholars who-not without optimism-suggest that 'Pacific islanders can serve as a living laboratory for the legal questions the international community will confront in the next few decades' (Burkett 2011:7).
In terms of the specific organisations that lead the PCCM project, Discourse 1 must be interpreted in relation to the broader landscape of regional climate and development politics and the changes therein. Most organisations interviewed compete internationally and regionally for development and climate finance and there is awareness that the flows of funding earmarked for adaptation will increase in the coming years (see also Atteridge and Canales 2017; Denton 2017). By connecting labour migration to climate adaptation, Discourse 1 articulates adaptation as compatible with the core mandates of two of the three PCCM-implementing organisations, the ILO and the UNESCAP (the third organisation is the UNDP, which traditionally has an environmental and climate focus), and, as a result, enables these organisations to move into a position from which they may access climate-related funding. The ILO has long worked on promoting labour mobility schemes, but not on climate change and not in the Pacific. The linking of the organisation's core mandate with climate change and the Pacific comes at a time when increasing funding is likely to flow into the region. Furthermore, the absence of a regional representation of the IOM, which internationally holds a privileged position on climate migration (Hall 2015;Ober and Sakdapolrak 2017), may provide the opportunity for the ILO as another international organisation to occupy this space. Through its involvement in the PCCM project the ILO has become a new player in responses and policy making on climate and migration in the Pacific, which is a significant expansion of their core mandate and increases organisational relevance to the region and to new funding sources (Ober and Sakdapolrak 2017 describe a similar process for the IOM at the international level, while Jinnah 2011 discusses the more general phenomenon of 'climate change bandwagoning'). The second project partner, the UNESCAP, has largely engaged in economic debates. Linking its economic development focus with climate migration also enables this organisation to tap into growing climate aid and to reinvent itself as a climate-relevant organisation. 11

Concluding discussion
In this study, I sought to understand how a set of development actors in the Pacific region conceptualise the relation between climate change, migration and adaptation and to explore what role different actors play in shaping or perpetuating different understandings. Despite their potentially important role for 9 The funding call requested proposals that target environmentally induced migration, foster links between migration and development, and develop tools and build capacities for well-managed labour migration. Furthermore, it specified that projects should help 'most particularly Island States, to develop tools and build capacities that enable them to adequately manage the challenges posed by climate changes and its impact on migratory flows, in particular measures to help these countries to develop tools and build capacities that enable them to adequately plan and manage the migration of their population as sea level gradually rises' (European Commission 2010:7). 10 EU institutions are the largest multilateral donor to the Pacific region and the sixth largest donor overall, following the bilateral donors of Australia, New Zealand, the USA, France and Japan (OECD 2017). development and climate aid across PICs, discourses at this regional scale have not gained attention in the literature thus far.
Whilst actors show competing understandings, signalling that it is a debate still in the making, I identified one main discourse on labour migration and two alternative discourses. It appears that Discourse 1 on labour originated outside the region in international development and donor agencies and has largely been driven by external actors involved with an EU-funded aid project. This discourse's most notable features, which are contested by the other two discourses, include the coupling of climate change with migration and the rendering of pre-emptive labour migration to surrounding Pacific Rim countries as a positive adaptive response. This is particularly interesting in light of other studies from the region that have shown how people do not want to leave their land (McNamara and Gibson 2009), and the dramatic shift of the Kiribati government's priorities away from Tong's initiated path to 'Migrate with Dignity' by the current Prime Minister, Taneti Maamau, who took over in 2016.
In the struggle between competing ideas, not all discourses have equal opportunity to establish themselves in a discursive space. Against the network of EU project-related actors, there is no strong, organised opposition. Proponents of the two alternative discourses are in no way organising themselves as a coalition, nor do they carry any significant weight in terms of funding or relationships with donors and PIC governments. That is, significant economic and political differentials exist between the discourse on labour migration and the alternative discourses: resource-backed UN agencies and the EU as a major donor vs. local CSOs, academics, two smaller IGOs, and a journalist. Accounting for this, it seems unlikely that the alternative discourses will make inroads into the emerging debate not only because of this lack of coordination and resources but also because they fail to invoke strong fantasmatic images, that might 'affectively pull' their audiences.
What does the migration as adaptation discourse open up in terms of broader implications for future migration governance in the region? First, it is important to reiterate the irony that the discourse, promoted by some IGOs and the EU, calls on the neighbourliness of Australia and New Zealand to develop suitable labour migration schemes, while not offering any suggestions for migration schemes elsewhere, for instance to the European region. Existing voluntary migration schemes, recently extended in Australia to include the aged care sector, are unlikely to provide opportunities for the most vulnerable Pacific islanders. Second, most Pacific islanders are more likely to move within their island nations to urban centres, a process that Discourse 1 pays scant attention to which raises the potential for maladaptive outcomes. This discourse effectively directs attention away from the challenges that come with increased urbanisation, with people living in the informal settlements of Suva and other urban centres in some cases already in more vulnerable positions than people in rural areas. Third, the high social costs of temporary labour migration schemes were not an issue that came up in Discourse 1. This implies the emotional, psychological and social costs that come with migrating for work, for the workers themselves and their home communities, are ignored.
In addition to offering empirical insight into how CMD are emerging in the regional development landscape and being articulated and influenced by different development actors, this research adds two new findings to the growing literature on the politics of CMD. First, the analysis illustrates that international discourses have not been passively received by actors and imported wholesale into the Pacific context. Rather, ideas have been dynamically appropriated and transformed in ways that align with the PCCM project organisations' earlier mandates and that match the regional social and political climate, for instance, by abstaining from unpopular notions of climate refugees. Second, the analysis of fantasmatic logics challenges previous research which proposes that discourse of labour migration as a form of adaptation marks a 'shift' or 'evolution' from an alarmist/ sensationalist perspective to a more optimistic view on climate migration (Faist and Schade 2013; Baldwin 2014; Bettini et al. 2016;Ober and Sakdapolrak 2017). By contrast, my analysis shows the emergence of a distinct discourse that draws on both images simultaneously. Accordingly, alarmist imaginaries are an integral and vital part of the emerging labour migration discourse.
This study has some limitations. First, discourses are not static, and so this analysis presents a snapshot in time (the discourses in late 2015 to be specific). Of course, the discourses described here are subject to change and, in their current form, will likely have evolved both in content and in the kind of actors who partake in them. Second, the small number of interview partners reflects this study's exploratory nature but presents another limitation. While it provides critical insight into how the relation between climate change, migration and adaptation is understood among an important set of regionally operating development actors, it is naturally unable to do justice to the diversity of perspectives across the Pacific development landscape and there is scope for a broader study. Specifically, capturing perspectives from national governments and NGOs, additional regional intergovernmental organisations such as the PIDF, and acquiring more interviewees from the PIFS would have immediate relevance. Future work may address some of these points and also examine how the discourse on labour migration as adaptation resonates specifically with people outside the atoll nations of Kiribati, Tuvalu and the Marshall Islands.
It is too early to tell whether Discourse 1 will gain credibility and support across the region to have a bearing on policy development and institutional practice, with the PCCM project having ended in 2016. Undoubtedly, more analysis needs to be conducted of the emerging discourse and the rhetorical change that it seeks to bring about by linking climate change, adaptation and labour migration.
Such work is relevant, as the EU is likely to invest more attention and financial resources on this issue in the coming years. More research is also needed on alternative or nonconventional discourses to migration, including those around internal relocation/resettlement. An aid project financed by the German Federal Ministry for Economic Cooperation and Development (BMZ), launched in 2017, to address 'climate change related migration and planned relocation', seems to initiate a different debate that seems more aligned with Discourse 3. In addition, the increasing resources for adaptation are likely to trigger a competition of ideas regarding what to invest in and what solutions are legitimate. The import of external ideas, such as those regarding labour migration discussed here, raises the prospect of localised adaptation choices and priorities being defined by actors from outside the region. At the time of fieldwork, concerns raised by the other two discourses were not considered legitimate by the proponents of the main discourse. If this is the case, it needs closer scrutiny. A deeper understanding of how different climate-related discourses emerge and are perpetuated is one valuable means of giving greater agency to PICs to create their own narratives about Pacific futures in the face of climate change and the choices that they wish to prioritise rather than donors and development agencies chasing funds for their own activities. For those concerned with developing and supporting sustainable and just responses to climate change and supporting Pacific self-determination, this task is both challenging and urgent.