Documenting the in-between: migrant-centered, multi-scalar epistemologies in irregular migration research

Beyond origin and destination: the case for studying the in-between

A quarter century ago, Wimmer and Glick Schiller (2002) criticized political scientists researching migration for their methodological nationalism, focusing too much on the nation state as the unit of analysis. They called for a move toward socially significant geographies. Since then, political scientists’ approaches have evolved toward comparative, multi-sited, supranational, and local research designs (Fitzgerald, 2006; Mazzucato, 2024; Wisthaler and Paquet, 2024). Yet this rescaling has not fully dislodged the deeper structuring of migration research around the categories and concerns of receiving states in the Global North. Research questions and funding continue to mirror state prerogatives (Sager, 2026), drawing scholars toward prospective migrants in countries of origin, where governments fund research to understand and stem departure (Spaan and Hillmann, 2013; Triandafyllidou and Ricard-Guay, 2019), and toward arrivals in destination countries (Black et al., 2006; Van Liempt and Doomernik, 2006), where institutional access is easier (Mainwaring and Brigden, 2016). Studying aspirations before departure risks oversampling migrants with weak intentions that do not translate into actual movement (Syed Zwick, 2022; Spaan and Hillmann, 2013), while studying successful arrivals systematically excludes those who died, turned back, or remained stranded along the way (Collyer, 2010; Schapendonk, 2012b). Despite a growing body of scholarship on migration journeys, the transit phase is still frequently treated as an intermediate stage rather than a primary site of observation (Crawley and Jones, 2021; Schapendonk and Steel, 2014; Schapendonk, 2012a).

In the imagination of the global north policy makers, the migrant movement toward Europe is a straightforward, linear process, notwithstanding scholarship showing that migration journeys instead are defined by intersecting patterns of mobility and immobility that remain insufficiently theorized in migration studies (Basok et al., 2015; Collyer, 2010; Mainwaring and Brigden, 2016; Schapendonk, 2012a). Migrant decision-making is inferred either prior to departure or after arrival, rather than observed as an ongoing process, in which journeys are interrupted, redirected, or recalibrated. This unwittingly reproduces a Eurocentric bias that disregards fluid, internal mobility systems, such as those within West Africa (Crawley and Jones, 2021; International Centre for Migration Policy Development, 2022; Andersson, 2014; Schapendonk et al. 2012b). It also obscures the temporal dynamics through which migration decisions unfold under conditions of uncertainty and constraint (Schapendonk et al. 2021). Despite scholarships critiquing the conceptualization of migration as linear, research designs have not kept pace with this recognition. This dual focus on origin and destination conceals the multi-layered social, emotional, and economic lives lived in the spaces in between (Crawley and Jones, 2021). The journey itself consequently becomes reduced to a ‘meaningless in-between phase’ Schapendonk, J. (2012a) rather than recognized as the site where migration governance is most intensely contested.

Additionally, too often the legal categories of migration, through which destination states govern migration and define who deserves protection, then in turn define academic research. However, these categories, such as forced versus voluntary, smuggled versus trafficked, victim versus agent, are not neutral analytical tools but reflect the political hierarchies and aims (Abdelaaty and Hamlin, 2022; Crawley and Skleparis, 2018; Hamlin, 2021). Consequently, the journey between these two points, and the migrants who do not complete it, have remained largely out of the analytical frame. Notable exceptions include trajectory ethnography that follows migrants across fragmented routes (Schapendonk, 2012a), research on stranded migrants caught between departure and arrival (Collyer, 2010), ethnographic work that centers the journey itself as the primary site of analysis (Brigden, 2018), and post-deportation surveys that capture migrant experiences at the moment of forced return (Slack and Martínez, 2018). Yet these approaches, innovative as they are, remain constrained: the first three by the access problem they identify, as mobile and criminalized populations actively avoid sustained contact, and the last by its institutional setting, which captures forced return as a bureaucratic moment rather than as a reflective social process.

For political science, uneven attention to transit and return has important implications for how power, governance, and political agency are studied in the context of migration. Contemporary migration policy has deepened the importance of this in-between space. Through the process of border externalization, destination states increasingly shift legal and physical control outward into transit corridors, delegating interdiction to transit country governments and non-state actors (Moreno-Lax and Lemberg-Pedersen, 2019; Zaiotti, 2016; FitzGerald, 2020; Achilli, 2024). Rather than deterring movement, externalization has fragmented and prolonged irregular journeys, producing conditions of protracted transit in which migrants become stuck in the process of movement (Schapendonk, 2012b), experiencing long periods of waiting or settlement due to proliferating migration controls (Mainwaring and Brigden, 2016), and making lives in “transit places’ until conditions become untenable (Crawley and Jones, 2021). The in-between is no longer a phase to be passed through on the way to a destination. It is where migration governance is most actively exercised and where the human consequences of externalization are most acutely felt.

These downstream effects of border externalization are not always legible as formal politics, yet nonetheless shape territorial control, state authority, and political order along routes of movement (Brigden, 2018; Mainwaring and Brigden, 2016). Recent work shows that non-state actors involved in clandestine mobility interact with border externalization efforts in ways that actively shape migration governance (Achilli, 2024; Author, 2026; Cobarrubias et al., 2023). Centering transit experiences addresses an important gap by treating in betweenness as a site of observation where decision making, exposure to coercion, and the effects of policy are most visible (Mainwaring and Brigden, 2016; Siviš et al., 2024). Refocusing on the journey also exposes the label of “transit state” as a political project. Destination countries use it to render complex movements knowable Frowd (2020), and governable from a distance, sidelining migrants’ social, emotional, and economic lives, their participation in local economies, and their formal rights under regional protocols like ECOWAS, in favor of a security narrative that prioritizes removal (Crawley and Jones, 2021; De Blasis and Pitzalis, 2023; Economic Commission for Africa, 2025). I conceptualize that site as a condition of juridical suspension. I which legal authority is formally invoked but rights protections are withdrawn, producing the spatial, temporal, and epistemological features that characterize irregular transit and return (Agamben, 2008; O’Donnell, 2004; De Genova, 2013).

This article responds to that gap by proposing a migrant-centered, multi-scalar research design for studying irregular migration governance in corridors shaped by externalized control. It makes two methodological contributions. First, it advances the returnee interview as an underutilized sample position to document migrant experiences of migration governance. Here, I encourage treating returnees as epistemic actors rather than policy subjects, in alignment with migrant-centered and decolonial methodological traditions (Thambinathan and Kinsella, 2021; Mullings, 1999; Denscombe, 2025). I define returnees as migrants who attempted irregular journeys and came back whether through deportation, voluntary return, or failed crossing. They occupy a vantage point unavailable to prospective migrants or successful arrivals. Returnees have completed a chapter of the journey and can narrate it whole, with enough distance to reflect and enough physical security to prioritize having their experience understood. Second, the multi-scalar research components encourages pairing retrospective interview narration with large-N transit survey data allows researchers to trace migration governance across scales, connecting individual experience to structural patterns that neither data source can reveal alone. Drawing on 32 semi-structured interviews with Gambian returnees conducted through a locally embedded partnership, combined with over 1,000 regional transit surveys collected by the Mixed Migration Centre across West African transit hubs, the article demonstrates what this migrant-centered, multi-scalar approach makes visible that remains opaque to origin and destination-centered designs. From this vantage point, migration governance appears not as coherent policy but as a lived sequence of encounters with fragmented authority, negotiated intermediaries, and enforced waiting, conditions actively produced through externalization rather than incidental to it.

Theory section: migrant-centered, multi-scalar methods

I conceptualize in betweenness not as a set of parallel dimensions but as an interconnected condition in which spatial, temporal, epistemological, and juridical forms of suspension mutually constitute one another. The juridical is central to this configuration. Extending Agamben’s concept of state of exception, I conceptualize border and transit zones as spaces in which law is formally enforced yet in terms of access to rights, substantively suspended (Agamben, 2008).¹ Legal authority is present, through both state governments and international institutions, but the protections normally associated with legal order are absent (O’Donnell 2004). This juridical indeterminacy produces the spatial and temporal features of in-betweenness. Detention, deportation, interception, and immobilization occur through the language and apparatus of law, yet without the stabilizing guarantees of due process or rights (Jegen, 2025). Spatially, juridical suspension produces zones of confinement and forced immobility: detention centers, informal encampments, and interior checkpoints that filter and trap migrants far from territorial borders, creating what Moreno-Lax and Lemberg-Pedersen describe as an interconnected network of sites where migrants are legal subjects without access to the legal protection (Moreno-Lax and Lemberg-Pedersen, 2019).

Drawing on my own fieldwork, Assan Darboe’s (pseudonym) account makes this concrete. While living and working in Morocco, a state that has historically hosted migrants as well as used migrants as a bargaining tool with Europe (Abderrahim, 2024; Stock, 2019), Assan describes this experience of being governed as a legal subject without legal protections. Moroccan authorities, he recalled,

“came to knock my door and found me sleeping, then took me away like I was a criminal. They would count [all black Africans living in the compound, arrest them] and tell the Spanish authorities that they [captured] the migrants in the waters. So, when the Spanish send them money, they would take it and then drop the migrants in the bush.”

Several returnees independently described the same pattern: a staged capture, presumably reported to European authorities, followed by abandonment in the desert rather than formal deportation. I cannot independently verify the financial mechanism Assan describes. Nevertheless, his account reveals how governance is experienced and interpreted by migrants who are subjected to it. There is a clear legal administrative form of governance evident: authorities count migrants, and presumably file a report, process a payment. However, legal protection disappears entirely: no charge, no hearing, no rights, only abandonment in unfamiliar and deadly territory. Independent investigative reporting corroborates the broader pattern Assan describes: European governments have funded and equipped North African security forces implicated in staged captures and desert abandonments, in at least one documented case resulting in the deaths of dozens of migrants (Lighthouse Reports, 2024). Fieldwork in Niger documents the same mechanism: after Niger criminalized migrant smuggling, security escorts abandoned convoys at the sight of patrols, leaving migrants stranded in the desert, sometimes fatally (Uzomah, 2024).

Such governance encounters produce a lived experience of prolonged in between, marked by neither departure nor arrival (Griffiths et al., 2013; Biorklund Belliveau and Ferguson, 2023; Siviš et al., 2024). The border thus emerges as a threshold in which legality and illegality, protection and punishment, and rule and arbitrariness blur (De Genova, 2013). These spatial and temporal conditions, in turn, shape the epistemological dimension of in betweenness. Migrants experience waiting, immobility, or repeated failure, as well as increasing abuse from state forces without access to legal redress or the ability to challenge their treatment before a court (Valdez-Castro, 2026; Angulo-Pasel, 2022). However, these are not simply reflections of institutional malfunction or bureaucratic delay, but a structured form of governance that obscures who bears responsibility for executing the law and normalizes uncertainty (Andersson, 2014; Heyer, 2022). Conceptualizing in betweenness as a lived state of exception therefore clarifies how transit becomes a site where sovereign power is exercised most intensely, not through the absence of law, but through its selective and uneven application (Ates, 2023). Epistemologically, this condition produces systematic obscurity: migrants cannot document their experiences through official channels, states actively erase evidence of what occurs at borders (Alpes, 2025; Moreno-Lax and Lemberg-Pedersen, 2019) and deploy what Stierl (2025, p. 14) calls non-knowledge as a ‘technique of governing.’ Simultaneously, researchers face structural barriers to sustained access to mobile and criminalized populations (Collyer, 2010; Schapendonk, 2012b).

Importantly, (de Genova 2002; Scheel and Tazzioli, 2022) the populations at the center of this study are not pre-given as irregular migrants. ECOWAS guarantees free movement and settlement across the region where most of their journeys pass through (Economic Commission for Africa, 2025). However, scholars document that member states routinely violate this guarantee, expelling and blocking documented and undocumented migrants alike despite ratification (Aniche, 2022; Bisong, 2021). Alassane Conateh’s (pseudonym) experience at the Bamako border shows what that guarantee is worth in practice. He presented his original ID card and medical card, full legal documentation, and Malian border guards told him to pay anyway or turn back: “either pay the money or go back to where I came from.” He described paying at every checkpoint on the road into Mali regardless of what documents he carried. A focus on the in-between reveals that externalized bordering practices practically suspend migrants’ economic, social, and human rights protections even where formal legal entitlement exists (De Blasis and Pitzalis, 2023; Moreno-Lax and Lemberg-Pedersen, 2019).

These conceptual claims have direct methodological implications. By centering transit and return as primary sites of knowledge production, this approach treats in-betweenness as a core analytic lens rather than a residual stage. Drawing on my own methodological choices of combining micro-level fieldwork with Gambian returnees and meso-level regional survey data from West African migrants in transit (Author, 2024), I advance a framework of migrant-centered, multi-scalar methods that treat methodology not as a neutral tool but as a site of intervention (Kindon et al., 2007). By centering migrants as self-reflective actors rather than passive subjects, these methods reveal how migrants navigate complex transit contexts and exercise agency within structural constraints (Andersson, 2014; Hamlin, 2021; Siviš et al., 2024). This approach treats migration as, in Desmond’s terms, an unfolding trajectory rather than a fixed status, emphasizing processes over processed people (Desmond, 2014). Grounding analysis in migrants’ perspectives emphasizes their expertise. It resists state-centric representations that reduce lived experiences to reductionist policy categories. This choice aligns with migrant-centered and decolonial research traditions that foreground affected communities as epistemic actors (Crawley and Jones, 2021; Mainwaring and Brigden, 2016; Thambinathan and Kinsella, 2021; Mullings, 1999; Smith, 2022). This study does not claim to realize a fully decolonized research paradigm. Fieldwork with criminalized and mobile populations involves constraints on reciprocity, co-ownership, and long-term collective deliberation. Nevertheless, the design advances these values in practice through trauma-sensitive interviewing, collaboration with Gambian-led organizations, and the treatment of interpreters as epistemic partners rather than technical intermediaries (Mullings, 1999; Kindon et al., 2007; Thambinathan and Kinsella, 2021).

The multi-scalar dimension underscores that migration must be understood across several interrelated levels, ranging from the micro-level of individual decision-making to the meso-level of relations with smugglers and community networks, and to the macro-level of border regimes and global governance (Author, 2024). These dynamics unfold across time and space rather than along a linear trajectory. Returnees reflect on extended periods of transit, and experiences of liminality, waiting, and mobility often continue long after movement has formally stopped, making the journey itself central to analysis (Griffiths et al., 2013; Desmond, 2014). These different scales become visible through distinct forms of data. Interviews allow insight into how migrants interpret risk, constraint, and possibility as they move through transit and return, rendering micro-level decision-making and meso-level relations legible from migrants’ own perspectives. Survey data, by contrast, captures how these dynamics take shape across routes and regions while journeys are still unfolding (Carling, 2024). Read together, these data make it possible to trace how individual decisions emerge within wider structures of uncertainty, coercion, and adaptation, rather than treating mobility as a sequence of isolated choices. This approach aligns with emerging research designs that similarly combine micro-, meso-, and macro-level dynamics of migrant journeys (Crawley and Jones, 2021) and with political science work that analyzes migration governance through multiscalar lenses tracing the interaction of local, national, and international logics (Jegen, 2025).

Research overview and context

This study draws on two complementary datasets. The first consists of 32 semi-structured interviews with Gambian returnees conducted in June 2022 through a partnership with the Global Youth Innovation Network (GYIN) Gambia, a Gambian-led youth organization with established community ties and experience engaging returnees. I gained access through this locally embedded network rather than through international migration management organizations. This choice shaped both who I could reach and what they were willing to say. Interviews were conducted in predominantly English, Wolof, and Mandinka, with trained interpreters present throughout. Another trained interpreter transcribed recorded interviews. I encouraged interpreters to note down discrepancies in translations and cultural meaning. Participants received modest compensation, disclosed only after the interview concluded to avoid inducing participation. The second dataset consists of over 1,000 standardized surveys collected by the Mixed Migration Centre across transit hubs in Burkina Faso, Mali, and Niger, capturing migrants’ experiences while journeys were still unfolding. To better understand the conditions of production behind this data, I visited the MMC’s West Africa offices and engaged with the team responsible for survey design and implementation. I found their approach to representativeness methodologically serious: working without a viable sampling frame, they had invested in PhD-level consultants and careful recruitment strategies to approximate coverage across diverse migrant populations. At the same time, the survey instrument reflects institutional and funding priorities oriented toward policy-legible categories. I therefore read the data critically, sifting for patterns that speak to the micro-level dynamics my interviews reveal rather than accepting the survey’s own framing. Used this way, the MMC data is not neutral ground but a structured account of migration that must be read alongside migrant-centered narratives to capture what large-N instruments cannot portray. Pairing these datasets is a core methodological choice, not a supplement: surveys provide distributional patterns across routes and regions while interviews render the interior logic of those patterns legible from migrants’ own perspectives.

The Gambia serves as the empirical case study. It offers a context where returnee populations are substantial, locally embedded organizations capable of supporting ethical access exist, and border externalization intersects with strong aspirational pressure among youth to migrate (Conrad Suso, 2022). Together, these produce the conditions of protracted transit and (at times forced) return. These features recur across other externalization contexts, and the framework is transferable to other settings where mobile, criminalized populations are structurally hard to reach. This design also responds to calls within externalization scholarship to center dynamics in Global South countries too often treated as passive objects of externalization rather than sites with their own methodological and analytical stakes (Cobarrubias et al., 2023).

The returnee interview opens an epistemic site unavailable to origin- or destination-centered designs. Returnees have completed a chapter of the journey and can narrate it whole, with enough distance to reflect. Rokhaya Diakité (pseudonym), a Gambian entrepreneur who had employed 120 men in the fishing industry before her business was sidelined by Chinese fishing operations and her subsequent bankruptcy instigated her choice to migrate, illustrates what this vantage point makes visible:

If I had known about the sufferings to come, maybe I would have cancelled the journey, or maybe I would have accepted the pain that I faced. Who do I hold responsible? Firstly, I am responsible for my choices, and secondly, the government is equally responsible. Despite its slogan of TEKKI FI (make it in The Gambia) and claiming to discourage young people from embarking on the backway journey, it failed to support me as a business entrepreneur. I was that young person who created employment for 120 men. I think that this young person should be celebrated and supported, but that’s where the government failed. The third person whom I will hold responsible is the smuggler, because he lied to me about [what was to come.]

Rokhaya is not reporting an event. She reconstructs a causal logic, assigning accountability across herself, the state, and an intermediary simultaneously. Her assignment of responsibility to the state tracks a broader pattern documented in comparative research on assisted return and reintegration programs, which persistently promise durable alternatives to migration without delivering the life-chance improvements those alternatives require (Marino et al., 2023). Such meaning making only becomes available after return.

Migrant-centered, grounded interviews as an epistemic strategy

Transit and return remain a scholarly blind spot. People in transit often purposefully hide and avoid formal contact to evade detection by the state (Crawley and Jones, 2021; Schapendonk and Steel, 2014; Basok et al., 2015), and dominant research designs continue to privilege points of departure or arrival. Grounded interviews offer a way to address these epistemic constraints by enabling migrants to reconstruct how they made decisions, anticipated risks, and interpreted responsibilities across time. Rather than inferring migrant decision making solely prior to departure or after arrival, this approach foregrounds how migrants produce meaning during and after movement, as they reflect on interrupted journeys and repeated encounters with state and non-state actors. Interviews thus capture migration not as a single choice or event, but as an unfolding process shaped by uncertainty, recalibration, and relational negotiation. Centering these narratives operationalizes the broader argument that migrants are epistemic actors whose accounts constitute knowledge about how migration governance operates, not supplementary texture to it (Achiume, 2019). As Matsuda (1987) argues, the act of “looking to the bottom” illuminates dimensions of law and governance that frameworks built from above routinely miss (p. 326). In this sense, migrant narratives do not merely add texture to existing models; they instead expose misalignments between abstract analytical categories and the complex, lived realities of transit. They show that migrants in prolonged transit encounter migration governance not as a coherent set of rules but through uncertainty, waiting, and fragmented authority (Schapendonk, 2012b; Andersson, 2014; Siviš et al., 2024).

Scholars across disciplines have demonstrated that accessing the lived realities of migration is possible through engagement with migrants (Eastmond, 2007; Hess, 2012). Fieldwork with people in transit has redefined the journey as a political terrain, revealing that migration governance is not a monolithic barrier but a fragmented system that manages rather than blocks mobility (Hess, 2012). A variety of actors exercise this governance piecemeal: smugglers, intermediaries, and transit state actors like police or immigration officials, all of whom migrants encounter within pervasive prolonged uncertainty, a lived reality that runs parallel to, and often at odds with, the state’s framing of migration governance (Author, 2026; İçduygu and Sert, 2014). Recent research on decision nodal points, the moments when migrants encounter transit-state officials, treats migrant narratives as the analytic object rather than the policy outcome. This choice reveals how these encounters function as concrete drivers of the journey rather than abstract implementation gaps (Moshe et al., 2025). In-depth narrative interviews trace how this governance operates in practice, because they follow migrants’ voices through the key events and the related decision making (Moshe et al., 2025). Ultimately, these narratives provide a “subjective mapping of experience,” revealing migrants as active agents who must constantly navigate this improvised system through complex survival strategies (Eastmond, 2007, p. 250; Bilger and van Liempt, 2009). Trajectory ethnography extends this insight. Following migrants through time and space rather than treating transit as a fixed phase between departure and arrival shows that onward movement is rarely linear or fully planned (Schapendonk and Steel, 2014). Another approach is to research on informal community structures, which for example documents gendered economies that state-centric accounts of border enforcement miss (Arrouche, 2023; see also Vogt, 2018, on intimate labor along the Central American route).

The term grounded is used here not to denote strict adherence to formal grounded theory, but to describe an iterative, relational, and context sensitive approach to qualitative inquiry.² For example, I progressively reworked interview scripts in response to recurring themes, emergent contradictions, and gaps identified through earlier conversations. This allows any study to remain responsive to migrants’ own framings rather than imposing fixed categories. Grounded interviewing in this sense prioritizes attention to how migrants describe their experiences, how meanings shift across temporal moments of transit and return, and how migration governance is understood retrospectively as journeys are reassembled after interruption or return. I treat interviews not as transparent accounts of events but as sites of retrospective sense making, where migrants actively reinterpreted decisions, risks, and responsibilities considering subsequent experiences.

While grounded, migrant-centered interviews correct a real gap in state-centric accounts of transit migration (Brigden, 2018; Mainwaring and Brigden, 2016; Vogt, 2013), there are important methodological considerations when conducting interviews. Prime among these is the positionality of the interviewer. An interview is a relational exercise, in which researcher and migrant co-create the record. The researcher elicits themes through both the interview protocol and the systematic engagement with the data (Braun and Clarke, 2022). The researcher’s interpretation adds a layer of mediation between lived experience and final text (Eastmond, 2007). Thus, at every step, the researcher co-creates the data, and the identities and biases of the researcher are important limitations. Working with local partners as co-creators of the research process, including NGOs, community leaders, and research assistants from the migrant community, can mitigate these limitations (Mackenzie et al., 2007; Bilger and van Liempt, 2009). Collaborating with insiders, or relying on trusted mediators, opens access and builds the trust needed to reveal realities that outsider researchers often cannot see (Bilger and van Liempt, 2009; Moshe et al., 2025). Involving local partners in setting the research agenda, or having them jointly read draft texts, keeps the final record grounded in participants’ lived experience and checks the researcher’s own subjective filters (Eastmond, 2007; Mackenzie et al., 2007). For example, I invited interpreters and transcribers to act as research assistants rather than limiting their role to technical support. Interpreters provided linguistic translation, but they also offered contextual guidance throughout the research process. They advised on sensitive questions, helped navigate local social dynamics, and debriefed with me after each interview to help identify the hidden meanings in silences, local idioms, jokes or body language. I engaged transcribers in a similar way and encouraged them to annotate interviews with alternative wordings or contextual notes when translation risked flattening or losing meaning. This practice treated transcription as an analytic task rather than a mechanical one and helped preserve layers of interpretation that extend beyond the literal transcript. I compensated interpreters and transcribers for their time, intellectual contribution, and emotional labor, recognizing that this work required judgment, cultural fluency, and ethical attentiveness. As Pacheco-Vega and Parizeau (2018) argue, regular check-ins with community partners allow researchers to ensure accurate representation and guard against implicit bias distortion.

Another complication to consider is that migrants, who move irregularly, and especially those who remain undocumented, have good reason to either remain hidden, or to present themselves in a particular way, since the account they give can shape their safety, their standing in a community, or their prospects going forward (Bilger and van Liempt, 2009, p. 128). Additionally, timing presents another complication. For example, an encounter’s meaning is not fixed: the accumulated weight of later encounters changes the significance of earlier ones, so that what might be dismissed or rationalized after a single incident takes on a different meaning once it becomes a pattern. The same instability extends to what goes unsaid, not just what is retold. Following Fujii (2010), I treat silences, hesitations, refusals, and narrative gaps not as analytic noise but as meta data that reflect the political landscape of the research encounter, and the risks participants navigate in transit and return. Attending to these absences is particularly important in contexts of migration governance, where fear, surveillance, and prior exposure to coercion shape what can be safely articulated. But reading silence as meaningful, rather than as an evasion to work around, first required ensuring participants had the freedom to withhold. In practice, this meant trauma-informed interviewing: allowing participants to determine the pace, depth, and direction of the conversation, including the ability to pause or redirect discussion when needed and to withhold information without penalty. When arriving at particularly traumatizing events, I reminded interviewees of their agency to pause, deflect, or skip portions of the story, and to be reflective of how their own desire to share might shift as they narrated events. Because this agency was genuinely available to them, what participants chose to say or withhold becomes legible as meaningful choice. I therefore treated ethics not as a procedural requirement but as an epistemic condition that structured how migration governance could be narrated.

This worry about self-presentation applies to migrants who remain in transit or undocumented, for whom the account given still carries stakes: it can shape safety, standing, or prospects going forward (Bilger and van Liempt, 2009). The population centered in my research design changes that calculus. My interviewees were returnees: Gambians who had attempted the journey toward Europe, been stopped somewhere along the way, and returned home by the time we spoke. This does not neutralize every incentive to shape a narrative, reputation within the home community, for instance, remains a concern, but it removes the specific incentive Bilger and van Liempt identify as the source of unreadable inconsistency: the need to manage an account for one’s ongoing safety (Bilger and van Liempt, 2009). Removing safety concerns lets the analysis focus on migrants’ own meaning-making, without filtering the narrative through self-preservation. I return to this sampling choice, and what it makes visible, more fully below.

Returnee sampling as an epistemic strategy

The vantage point sketched above, in which interviewing someone upon return to strip security concerns influencing migrants’ shares, is not incidental to this design. It is a contribution. Existing sampling frames cluster at two points in the migration process. Studies of aspiration and departure interview migrants who are still forming or acting on the decision to leave (Altrogge, 2025; Triandafyllidou, 2019, 2022). Studies of reintegration and post-return outcomes interview returnees, yet they focus on reintegration experiences, rather than on the experiences in transit (Ruben et al., 2009). Neither treats the experience migrants have in transit as central to understanding migration governance. Both dominant sampling approaches also face a practical constraint that returnee sampling does not. Migrants actively planning a journey have strong incentive to stay hidden, since criminalization has pushed departure planning further into secrecy. During my own fieldwork, I identified only one migrant willing to discuss active departure plans (which I defined as having spent money to secure a journey). This difficulty is not particular to my fieldwork. Hess (2012) found that their undocumented interviewees were skeptical and refused being taped because of their need to avoid public attention. Bilger and van Liempt (2009) describe this more broadly: an insecure legal situation, whether before or during migration, discourages research participation, because revealing private details runs contrary to a migrant’s own survival strategy. Returnees face no equivalent exposure. The Gambian government does not criminalize their status. Aid and reintegration funding has facilitated returnee networks and organizations who service them, the GYIN among them. This makes returnees identifiable and reachable in ways prospective migrants structurally are not. Returnee sampling is therefore not a fallback for lacking better access. It is the strategy that makes analytically distinct knowledge available and practically accessible at once.

A small number of studies have sampled returnees to learn about the journey. O'Leary (2009) interviewed repatriated migrant women at a shelter on the Arizona-Sonora border on the crossing itself: the cost of organizing an attempt, extortion by bajadores, payments made to smugglers. Slack and Martínez’s (2018) postdeportation survey asks migrants to evaluate their smuggler relationship “after the fact.” Both studies worked with migrants close in time and place to the crossing. Interviews happened at the border, within a month of deportation. My interviewees were a mix of freshly returned and those who had more time back in The Gambia.

There are some risks to interviewing returnees. Fitzgerald (2006) warns that retrospective accounts can become reified, obscuring the uncertainty that was present at the time of the events. Returnees narrate trusting a smuggler, or accepting the risk of a journey, with full knowledge of how that journey ended. Confidence in hindsight is not evidence that the decision felt confident when it was made. Altrogge’s (2025) longitudinal study of Gambian deportees and AVRR returnees illustrates a related point. Interviewing the same people three times across 2.5 years, she found that the two return pathways, assisted and forced, converged over time into outcomes that looked similar regardless of which one a migrant started from. Only repeated interviews with the same people made that finding visible. A single retrospective account would have caught one frame of a still-moving process, not the process itself. A second limitation sits within the returnee population itself, not within any single account. Returnees do not represent migration outcomes generally. They represent one outcome among several, alongside those who reached their destination and those still in transit. Becoming a returnee is not random. It follows from repatriation, exhaustion, running out of funds, or a voluntary return decision, and these pathways may correlate with the kinds of governance encounters a migrant had (Author, 2024). A sample built entirely from returnees may systematically differ from migrants who succeeded or who remain in transit, in ways the design cannot detect from within.

To address some of these concerns, my study does not rely on retrospective interviews alone. The MMC survey data, collected from migrants while their journeys were still unfolding, asked migrants to describe their smuggler, rate whether the smuggler was helpful, and say whether they felt misled. Almost none chose “criminal” as a descriptor. Half called their smuggler helpful. These are not settled, retrospective judgments narrated from safety. They are assessments made by people still in transit, still exposed to the outcome. The pattern in the survey data mirrors what returnees like Katim Ceesay (pseudonym) describe after the fact: trust, wariness, and a clear-eyed sense that a smuggler is a necessary means to an end rather than a criminal predator. Where retrospective interviews and contemporaneous survey responses converge, reification is a weaker concern. The confidence in Katim’s account is not only a product of hindsight. It matches what migrants report while the outcome is still unknown.

Interviewing returnees allowed me to document how migrants’ perceptions of smugglers were often conflicted: deep distrust of smugglers as a category, alongside situational trust extended out of necessity, and these can coexist, in some cases, with exceptional protection from danger smugglers provided. Assan Darboe, Ebrima Manneh and Saikou Touray illustrate this pattern. Assan credited the smuggler who guided him through Libya, Bamba, with intervening on his behalf against rebels and absorbing bribes meant for him. He still called smugglers categorically “not good.” A smuggler found Saikou in the desert after he escaped a Libyan torture camp. Saikou was injured and starving. That smuggler sheltered and fed him for six months and organized his onward transport, without payment. Despite this lifesaving act, Saikou still describes smugglers as not good people. Both men separate a specific debt of gratitude from a categorical judgment about the role, without treating one as evidence against the other. MMC survey data confirms this pattern extends beyond returnee accounts. Across regions, only 1 percent of respondents named smugglers as someone who would commit a crime against them. That is far below the share naming criminal gangs, government officials, or armed groups. Migrants’ retrospective accounts also show how migrants evaluate smugglers in relation to other actors encountered along the route. As Ebrima Manneh explained:

“At some point, you do not have a choice [but to trust a smuggler]. When you go to a deep forest full of lions, if you see a human being, how would you feel? You would have no choice but to follow him because he is the only surviving person in there.”

In this context, turning to a smuggler was not a choice among options, but the only visible path through spaces where state authority offered neither protection nor passage. Trust here is not moral or absolute; it is situational and conditional, shaped by comparison with other available actors. Had these returnees been interviewed still in transit, or newly arrived, that same combination, crediting one smuggler while condemning the role categorically, could plausibly be read as protective: hedging an account in case it reached the wrong ears. Narrating from a closed chapter, that explanation loses its force, and what remains is the pattern itself, consistent across several separate accounts.

Because of the sensitive stories that returnees hold, interviewing returnees well is not simply a matter of identifying and locating them. It requires trust, which I built through sustained local partnership, careful translation that preserves rather than flattens meaning, and interview practices attentive to trauma. I detail how these commitments were enacted in practice a little later in the article. Where sustained local partnership makes deep, trust-based interviewing possible with a small number of returnees, MMC survey data extends the study’s reach across a much larger population, collected by an institution positioned to gather it at a scale no individual researcher could replicate. I detail this pairing, and its own limits, more fully below.

Seeing the in-between across scales and temporalities

Migration research drawing on a single level of analysis by necessity entails systematic blind spots, particularly with respect to journeys and transit contexts that fall between origin and destination. Disciplinary divisions contribute. Political science has tended to privilege macro-level data analysis, while anthropological work has more often centered experience, and sociology has often focused on the meso-level (Bonfiglio, 2011; Carling, 2024). Ethnographic interventions have been crucial in bringing the journey into view as a political space and in showing the potential of micro-level methodologies for political science (Brigden, 2018; Mainwaring and Brigden, 2016). But approaches that remain at the level of micro practice cannot show how that practice is shaped by governance decisions made at a scale no interview can reach. This gap has stakes in political science. Authority, coercion, and the exercise of state power are the discipline’s core objects of analysis. Yet in migration governance these operate diffusely, distributed across state officials, non-state intermediaries, and international institutions acting at different scales simultaneously. A method that captures only the micro-level encounter, a migrant’s negotiation with a specific official or smuggler, cannot show whether that encounter reflects a macro-level policy decision, a meso-level institutional practice, or local improvisation. Without tracing the connection, governance itself becomes illegible: visible only in its immediate, individual effects, never in the structure producing them.

Conceptualizing migration through the lens of in-betweenness, a condition of spatial, temporal, epistemological, and juridical suspension, demands a research design capable of connecting individual experience to the structural patterns that produce it. The same demand applies across time. A single-moment interview or survey wave captures a journey at one point. It cannot show how a journey unfolds, stalls, or repeats. Only retrospective narration, paired with survey data drawn from many migrants at many different points, makes that temporal shape visible: how long a journey takes, how often it doubles back, how far lived time diverges from anticipated time. Griffiths et al.’s (2013) account of migrants oscillating between “frenzied” and “suspended” time captures this shape descriptively. It stops short of specifying what evidence could make that shape legible across a population rather than a single account.

A multi-scalar approach that traces governance across micro-, meso-, and macro-levels is one way of meeting that demand, particularly the epistemological task of situating individual meaning-making within the structural patterns that produce it. More recent work has begun to engage scale more explicitly, including efforts to analyze borderwork and migration governance through multi-scalar lenses that trace the interaction of local, national, and international logics (Jegen, 2025), and research designing studies that attend simultaneously to micro-, meso-, and macro-level dynamics of migrant journeys (Crawley and Jones, 2021). A narrower body of work moves in this direction using large-N survey data specifically. This study draws in part on data from the same Mixed Migration Centre’s 4Mi initiative, a standardized survey administered to migrants and refugees while journeys are still unfolding across multiple transit regions. Although 4Mi data are most often disseminated in policy-oriented formats, they have increasingly been taken up in peer-reviewed migration scholarship to analyze journey dynamics and migrant decision-making (Syed Zwick, 2022; Hovil and Gill, 2022). These studies demonstrate the analytical potential of survey data collected in transit contexts, even as they stop short of theorizing scale or governance explicitly. What remains under-theorized, across both bodies of work, is how a research design can trace macro-level governance down into the meso-level dynamics of routes and intermediaries, and from there into the micro-level decisions migrants make.

Institutional data of this kind is never neutral, a limitation already discussed above. Early mixed-methods research treated this by emphasizing triangulation, checking one data source against another for agreement. More recent approaches foreground complementarity instead, using different methods to generate insights that neither source could produce alone (Carling, 2024). This study reads survey and interview data as complementary in that sense: the survey traces distributional patterns across a large population, and the interviews reveal the mechanism underneath. My pairing of MMC data with interviews revealed a gap in a policy premise, for example. EU-funded awareness campaigns invest in educating migrants about the dangers of the journey, on the assumption that better information about risk would change behavior (European Commission, Directorate-General for Migration and Home Affairs, 2021). Both my interview data and the MMC survey raise questions about that assumption. In the MMC survey, 87 percent of migrants said they would still pursue the journey even knowing what they now know about how it has gone so far, and 78 percent said they would recommend it to others. Read alone, the survey leaves room for doubt: migrants were surveyed in transit, before the most dangerous stretches of the route, so their confidence might simply reflect not yet knowing what lies ahead. The interviews close that gap. Migrants describe the decision as a gamble, and a gamble is not disproven by someone else’s loss. If some chance of success remains, another person’s suffering reads as the risk the wager always carried, not as evidence the wager is a bad one. Katim Ceesay heard his uncle’s account of torture in a Libyan detention camp firsthand, before his uncle later died from those injuries, and went anyway. A second interviewee, captured and detained three times in Libya himself, warned his brother specifically against the journey, naming the prison where he had been held. His brother went, was captured in that same prison, and died there. The same logic runs forward, not just past a warning but into active encouragement. Saikou Touray described his standing practice: “if anyone asked me about the journey and he wants to go; I will recommend it to him. I have seen so many who went through the back way and they succeeded.” Risk information here was not missing, misunderstood, or abstract. It was heard directly, weighed against the odds of success, and did not change the calculation. The survey shows that persistence and recommendation are both widespread. The interviews show why: neither one requires believing the risk is small, only that it is worth it. A survey alone risks mistaking a widespread pattern for an informed one. Interviews alone risk mistaking a few accounts for a general one. Together, neither risk holds. A policy built to change minds about risk misunderstands what is being weighed. Awareness does not compete with a gamble. It is already priced in. The same disconnect between campaign messaging and migration decisions has been documented directly among Gambian information-campaign recipients, where local narratives about the opportunities of migration outweigh campaign messaging regardless of message content (Cham et al., 2025).

One strength of returnee interviews is that they reveal how prolonged exposure to danger shifts a migrant’s own risk calculus: remaining in place comes to feel as dangerous as moving on. As Ebrima Manneh (pseudonym) reflected on the compounding effect of enduring exposure to conditions of precarity, and his choice to embark on the risky Mediterranean journey: “Me, I said, if I must, let me die today [on the boat]. I am tired of these situations. Every day [on land] is running for your life.” Survey data from the Mixed Migration Centre indicate that such assessments are not idiosyncratic, with respondents frequently identifying criminal gangs, government officials, and armed groups or rebels as sources of fear and harm during transit. The same accumulation reshapes the journey itself, not just how migrants judge it. Returnees described months-long periods of waiting in transit hubs, surviving in the meantime through casual labor. MMC data confirm that such patterns are widespread rather than exceptional: 45 percent of respondents reported stopping to work to finance onward travel. During a group interview, one returnee explained, “The [smuggler] told us that we will spend two days, but we ended up spending eight months there.” What was imagined as brief passage becomes extended suspension, as departure timelines are displaced by enforcement, bottlenecks, and smuggler-controlled sequencing. Rather than progressing steadily, journeys often take the form of recursive cycles of movement and capture. As Boy Kandeh (pseudonym) recounted, “I had lots of difficulties in the journey… I had 10 years in the journey… When I go the beach side the navy would capture us… 7 times.” Together, these accounts show that waiting, repetition, and enforced immobility are not anomalies or temporary setbacks but systemic features of contemporary migration trajectories produced within regimes of externalized control.

Practicing ethical field access through local partnerships

This section details how the methodological commitments outlined above were enacted during fieldwork. I describe how I established access, identified research partnerships and how in turn these research relationships developed over time in practice. Prior work on community-engaged research emphasizes that collaboration depends on negotiated trust, organizational capacity, and a bidirectional flow of knowledge rather than formal affiliation alone (Ross et al., 2010). Building such trust requires engagement preceding formal data collection. Partnerships often emerge through mutual assessment rather than through predefined project structures. Rather than offering a fixed model, this section outlines a sequence of decisions and relational practices that may be adapted to other contexts where research access is shaped by criminalization, mistrust, and uneven institutional capacity.

Before traveling to The Gambia, I spoke with researchers who had previously conducted migration-related fieldwork in the country. These conversations helped clarify local research dynamics, ethical risks, and the reputational landscape surrounding work with migrants, smugglers and returnees. They informed decisions about research timing, initial points of entry, and the feasibility of different access pathways. In the Gambian context, social proximity plays a central role in establishing credibility and trust. Prior work on West African migration highlights how legitimacy often emerges through relational networks rather than institutional affiliation alone (Maher, 2018). Developing such proximity requires in person presence. As Brigden (2018) notes, researchers develop a situated sense of place by moving through the same social settings as participants, even when they occupy different positional roles.

I subsequently undertook an initial pilot trip prior to formal data collection. The purpose of this trip was to build relationships, identify potential partners, and assess research dynamics in person. Physical presence created space for an initial, provisional assessment of my intentions and positioning. During the pilot trip, I pursued a layered approach to access. I met with staff from international and donor-funded organizations operating locally to better understand service provision, institutional constraints, and gaps in support for returnees. These organizations did not serve as recruitment intermediaries. Instead, they provided contextual insight into the institutional landscape and its limits. Because The Gambia requires a formal research permit, I also met with government officials responsible for research authorization and migration-related programming. These meetings enabled compliance with regulatory requirements and situational awareness of ongoing policy discussions. In parallel, I attended migration-focused workshops organized by donor organizations and NGO-led forums addressing services for returnees. These spaces offered insight into how different actors framed migration governance and return.

Alongside these engagements, I reached out to a range of local organizations, including some founded by returnees themselves. These organizations varied in their capacity to engage in research partnerships. Some expressed interest but lacked institutional resources, while others had greater capacity but limited interest in research collaboration. Partnership therefore depended not only on trust but also on organizational capacity and willingness to engage (Ross et al., 2010). Sustained engagement with locally embedded organizations ultimately formed the basis of ethical access and recruitment. Partnerships developed through existing relationships and shared commitments. I collaborated with organizations that were accountable to local communities, engaged with returnees outside formal deportation or reintegration frameworks, and open to participating in the research process as partners rather than intermediaries. This approach grounded access in locally meaningful social networks rather than in any single institutional pathway.

In practice, these commitments shaped my decision to work with a Gambian-led youth organization with longstanding community ties and experience engaging returnees. Collaboration with a locally embedded organization enhanced the credibility of the research and facilitated access in ways a foreign researcher could not achieve independently. The partnership supported recruitment and interpretation during fieldwork and enabled engagement beyond urban centers, grounding the research in locally meaningful social networks rather than formal institutional channels. I provided institutional support to the partner organization to acknowledge the labor involved in facilitating access, coordination, and interpretation. At the same time, I did not rely on international organizations that dominate migration governance in the region as research intermediaries. While such actors play a central role in service provision, affiliation with donor-driven institutions risked shaping participant perceptions of the project and conflating the research with external policy agendas. Working instead with Gambian NGOs situated the research within locally defined priorities and signaled accountability toward affected communities rather than international institutions.

This commitment to grounding research in local priorities also shaped the everyday conduct of fieldwork. Compensating interpreters and transcribers for their time, intellectual contribution, and emotional labor registered that interpretation and transcription in this context demanded judgment, cultural fluency, and ethical attentiveness, not simply language conversion. Taken together, these practices shifted the foreground of the research toward participants and local partners and redistributed some analytical authority: interpreters and transcribers shaped what counted as meaningful data, and GYIN’s priorities shaped who was reached and how. Authority over research design, analysis, and the resulting argument remained mine. This is what makes it possible to examine migration governance from the in-between, where its effects are most acutely felt and least visible in policy-centered analyzes.

These commitments were tested directly during fieldwork, not only described in the research design. Near the end of my interview with Djibril Sanyang (pseudonym), he turned the exchange around:

Me: Do you have any questions for me?

Djibril: Yes, I have a question for you.

Me: What is the question?

Djibril: So why the European use to deport someone without no evident, without no problem, without no case, without nothing absolutely?

Me: I do not know.

Djibril: You do not know?

Me: Yes, unfortunately, I do not know.

Djibril: So why are you interviewing the migrants? is it going to be benefit for the migrants or you or the European Union (EU)?

Me: I am trying to understand the experience you have and how the policies affect people who travelled the backway and I am trying to push for better migration policies. So eventually, I hope there will be benefits to other people who want to go.

Djibril: Blessing for that.

The interviewee’s first question named a grievance about arbitrary state power he had no way to resolve himself, a deportation he described as carried out without evidence, without a case, without cause. I answered honestly that I did not know, rather than offering reassurance I could not support. His second question tested the research relationship directly: whether the interview served him, me, or the EU. My answer named a genuine hope, that the work might eventually inform better migration policy, but left out something equally true and closer to hand: this was my dissertation research, its future still uncertain, and part of what I hoped for was that doing it well would let me become the advocate I wanted to be. That omission is worth naming here rather than passing over. The exchange closed warmly, “blessing for that,” but a fuller answer would have named my own stake in the work alongside the hope I offered him.

Each of these strategies solves a problem the others cannot. Grounded, trust-based interviewing gets access to accounts that a less relational approach would never surface. Sampling returnees removes the need to manage an account for ongoing safety, so what migrants say can be read as meaning making rather than self-protection. Pairing that testimony with large-N survey data shows whether an account is idiosyncratic or widespread, and reveals patterns in timing, in the compounding weight of danger, in the logic behind persistence, that neither source could show alone. No single strategy, used on its own, would have produced this picture of migration governance. Together, they make the in-between legible as a site of knowledge production, not a gap between origin and destination.

Conclusion

This study centers transit and return to analyze migration from the vantage point of the in-between. It advances the in-between as a multi-scalar analytical framework that links migrants’ situated decision-making to intermediary practices and to the broader architectures of border externalization. Placing transit and return at the center changes how migration governance is experienced. Surveys capture the prevalence of waiting, extortion, and reliance on smugglers at scale, while interviews reveal how migrants make retrospective sense of these experiences once they are safely returned. Methodologically, this pairing demonstrates that the “in-between” is not an empty corridor but a dense archive of power, precarity, and migrant knowledge production (Carling and Schewel, 2018). This connects micro-level experiences and strategies to meso-level configurations of smuggling and enforcement and to macro-level regimes of deterrence and externalization, making it possible to trace how decisions, risks, and forms of control co-evolve across scales.

The goal here has not been to document every governance mechanism migrants encounter in transit, or to offer a comprehensive account of externalized bordering across West Africa. The narrower argument is methodological: migrant-centered, multi-scalar methods, pairing returnees’ retrospective accounts with distributional survey data, give researchers access to an epistemic vantage point unavailable to prospective migrants or those still in transit. This allows researchers to examine how individual decisions and structural patterns co-produce each other. This argument is not exclusive to Gambian migration or to irregular journeys specifically. The same logic applies wherever a population is mobile, criminalized, or otherwise difficult to reach in the moment its most consequential experiences occur.

The interviews alone draw on a single national-origin population. The MMC survey data, spanning over 1,000 respondents across West African transit hubs, is what tests whether the patterns these interviews reveal are idiosyncratic to Gambian returnees or shared across the region. Beyond West Africa, applying these patterns elsewhere rests on structural analogy, the same basis on which any single case travels to populations it did not sample. Migrant-centered, multi-scalar methods can travel this way to other origin populations, other criminalized or hard-to-reach groups, and other externalized border regimes, testing whether what is visible from the in-between here holds elsewhere or takes different forms. More broadly, this framework contributes to debates on qualitative inference and mixed-method integration by showing how retrospective narration and distributional patterning, read together, can make legible governance processes that neither source captures alone.

Generative AI statement

The author(s) declared that Generative AI was used in the creation of this manuscript. The author used ChatGPT (GPT-5.2, OpenAI) and Claude Sonnet 5 (Anthropic) at different stages of manuscript preparation to assist with language editing, structural organization, and checks for repetition. NotebookLM (Google) was used specifically to verify citations and references. All empirical data, interview material, ideas, arguments, and findings presented in the manuscript originate from the author. The author critically reviewed all AI-assisted suggestions, revised them substantially, and incorporated only those judged accurate and appropriate. The author retains full responsibility for the manuscript's content.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Stephen Ruszczyk, Montclair State University, United States

Reviewed by: Shyam Sriram, Canisius University, United States

Grażyna Baranowska, University of Erlangen Nuremberg, Germany

Footnotes

Agamben’s state of exception describes a zone where sovereign power suspends the normal legal order but keeps its outward form: rights exist on paper but not in practice. For a fuller treatment of the concept across Schmitt, Agamben, and Butler, see Ates (2023). The concept extends to Agamben’s (1998)homo sacer, life included in the political order only through exclusion from it. Migrants navigating juridical suspension fit this pattern: counted and processed by the state, but without the protections that recognition should bring.

As O’Reilly (2025) argues, objectivist versions of grounded theory that presume access to “independent facts” or a researcher’s “blank mind” are often unrealistic; qualitative analysis instead proceeds through an iterative oscillation between empirical material and conceptual frameworks. In a similar vein, Brigden’s (2018) work on returnees emphasizes that clandestine journeys are not simply reported but retrospectively reconstructed, as routes are “imagined and improvised” through memory and narration.