Containment beyond detention: The hotspot system and disrupted migration movements across Europe

Introduction: Situating containment

Ventimiglia, Italian-French border, 10 August 2016. Four hundred migrants are forcefully removed from the hosting centre run by the Red Cross in Ventimiglia to the hotspot of Taranto, ¹ in Southern Italy. Other forced transfers to the Taranto hotspot took place from Como, the Italian city located at the border with Switzerland. Here, between summer and autumn 2016, migrants ended up being blocked due to the closure of the Swiss border and of push-back operations carried out by Swiss authorities. ²

Two years after the implementation of the European Union (EU) ‘Hotspot System’, hotspots’ facilities partly changed their functions: envisioned as sites for regulating migrant arrivals by sea, hotspots have also become places for migrants’ redistribution on-land, i.e., places where migrants apprehended at the internal borders of Europe are temporarily transferred. Presented in the European Agenda on Migration as frontline infrastructures for incoming migrants’ identification through enhanced mechanisms of intra-governmental control, ³ Italian hotspots have also been used for redistributing migrants already present on the national territory, and particularly for removing their unruly and contested presence from highly visible sites. So far this has happened in the aftermath of migrants’ protests and clashes with the local police in sites such as Ventimiglia, at the Italian-France border. These new evolutions complicate the territoriality and sovereignty of hotspots and call for an investigation of their actually existing geographies, beyond the EU governmental vision and the accounts of the governmental actors involved in their implementation and management.

In Greece, the hotspot-islands were used as ‘dock-and-go’ sites during the first couple of months since their implementation. Migrants were registered relatively quickly and were subsequently let move on to the mainland. With the signing of the EU-Turkey deal in March 2016 ⁴ and the almost simultaneous closure of the Balkan corridor, Greek hotspots have become containment places for blocking migrants on the islands facing Turkey, hence preventing their arrival in Athens.

This article engages hotspots as flexible chokepoints of mobility disruption and looks at the forms of containment and channels of mobility that the ‘hotspot approach’ has generated. In other words, if hotspots serve a containment function as migrants’ temporary detention sites, here we are interested in mapping the geographies of mobility and immobility they enact through and beyond this function.

By speaking of ‘containment beyond detention’, we gesture towards an approach that situates hotspots within broader spatial economies of migration governmentality. Our goal is to move beyond an analytical focus on hotspots as isolated units of border crossing. Second, ‘containment beyond detention’ refers to a focus on containment practices beyond the premises of detention sites. In this regard, we build on Michel Foucault’s methodological displacement, which consists in ‘moving outside the institution […], putting the prison back in a general economy of power' (Foucault, 2007: 116–117). In other words, we look at the proliferation of containment strategies in the Hotspot System without superimposing a preconceived notion of carceral space to the detention mechanism. Presented as a surgical technology for enhanced border functions, hotspots are instead complex and flexible systems for managing migrants’ mobility to, through, and away from the European space. By ‘containment’, we refer to the effects of mobility disruption, spatial fixation and temporal suspension that are generated through measures of confinement that do not coincide with detention. Such practices aim to regain control over migrants’ autonomous geographies.

The notion of ‘containment’ remains partially under-theorised in migration scholarship. When it is employed it tends to be used as a synonymous of ‘confinement’ and in opposition to ‘mobility’ (Mountz et al., 2013). On the contrary, we argue that containment is not only confinement and that there is a need for exploring the nexus between containment and (institutionally forced) mobility. In other words, while containment can also consist in spatial confinement practices (as it is the case of migrants blocked in Libya and prevented from crossing the Mediterranean), it is not only blockage and immobility. Rather, governing migration through containment means troubling both migrants’ geographies and migrants’ presence. For instance, as we will show in this article, containment effects also derive from governmental practices that keep migrants on the move (Gill, 2009; Tazzioli, 2017), not just blocking them in a structure of detention. Finally, the notion of containment allows avoiding the trap of the binary oppositions that underwrite migration theories, i.e., inclusion/exclusion and mobility/immobility. Containment, in fact, is not only about keeping migrants out of a place per se (out of Europe, out of national borders); nor is it only about their (adverse) incorporation into the labour market. Containment is about governing migration by disrupting, decelerating and diverting migrants’ autonomous movements and by hampering migrants’ presence in certain spaces. While containment can take the form of spatial confinement and protracted strandedness, here we point to modes of containment enacted by keeping migrants on the move. Through this take on containment, we map mobility control in its actually existing spatial processes, which go way beyond the governmental phantasy of seamless circuits and channels of mobility control. Far from being smoothly channelled through the asylum system, migrants are subjected to modes of government that frantically try to regain control over their autonomous movements: containment – the physical containment of certain movements, but also, and most importantly, the containment of migrants’ desire to settle in a particular place – is one of these strategies, as the article will show.

Methodologically, the empirical focus of our research is positioned beyond the enclosed space of the hotspot and the bounded time migrants spend there. We instead direct our attention to the control and hosting spaces that are connected to the hotspots but branch beyond the hotspots’ perimeters – the Hotspot System’s impact beyond the space of the hotspot, ‘beyond the camp’ (Garelli and Tazzioli, 2017). In other words, our goal is to follow what happens to some migrants beyond the spaces and times of hotspot detention, and hence to map the outcomes of the Hotspot System on migrants’ routes and lives.

We conducted participatory observations and interviews with migrants both within the premises of hotspots, ⁵ and in border zones such as Ventimiglia and Como in Italy, where migrants went after fleeing hotspots upon identification. We also kept in touch with a group of migrants who fled from the Lampedusa hotspot after having been subjected to identification procedures by the use of force (e.g. forced fingerprinting) and reached the cities of Ventimiglia and Marseille. We also conducted semi-structured interviews with Frontex and Easo ⁶ officers – who are the representatives of the EU inside the hotspots – and with the Italian and Greek authorities in charge of the implementation of the hotspot system and of complying with the EU standardised identification procedures. ⁷ Finally, we interviewed workers of nongovernmental organizations and international organizations, ranging from Doctors without Borders (an actor who refused to work inside the EU hotspots) to the main actors of the Hotspot System like United Nations High Commissioner for Refugees (UNHCR) and the Red Cross. Our choice to interview these different actors was rooted in the consideration that far from being the result of a shared and streamlined political strategy or the manifestation of a supposed effective Europeanisation of migration controls, the Hotspot System is in fact the outcome of several frictions – between the EU on the one hand and Greece and Italy on the other.

In order to account for the outcomes of the hotspot approach on migrants’ journeys, we mobilise an analysis that is receptive to hotspots’ heterogeneous and flexible functions. Such analysis maps the spatial productivity associated with the EU hotspot approach beyond the policy discourse on securitised crossing hotspots. Instead, our study maps the spatial processes that hotspots ignite, from the repurposing of hotspots as ad hoc containment sites, to the proliferation of informal hotspot-like space, to the spatial expansion of detention practices.

In the first section of the article, we trace a genealogy of the use of the term ‘hotspot’ suggesting that the multiplication of hotspots-like designated spaces is related to a governmental re-conceptualisation of the border as a site requiring prompt enforcement interventions. The article moves on by investigating the mechanisms of partitioning, identification and preventative illegalisation ⁸ that are at stake in the hotspots of Lampedusa and Lesbos. Hotspots are not analysed as sites of detention per se: rather, we turn our attention to what we call ‘containment beyond detention’ and to the exclusionary humanitarian channels of the EU Relocation Scheme, which pertain to the Hotspot System functioning. We conclude analysing channels of forced mobility in light of the fight against ‘secondary movements’ that is at the core of the current EU’s political agenda and sketch a research agenda for a genealogy of practices of migration containment.

Our argument about hotspots as mechanisms for containment beyond detention critically engages with three bodies of literature: carceral geography, enforcement archipelagos and migration logistics. Geographers’ engagement with migrant detention importantly brought to the fore the nexus between disciplinary mechanisms and the spaces of control, by illuminating the ‘embodied inscription of incarceration’ (Moran, 2015), unpacking the economy and micro-economies of detention (Belcher and Martin, 2013; Conlon and Hiemstra, 2014) and illustrating the experienced temporality of imprisonment (Wahidin, 2002; Wahidin and Tate, 2005) and the spatialities of containment (Mountz and Loyd, 2014). This body of literature also radically unsettled the boundaries of the prison, investigating forms of containment that take place outside of the fences of detention centres while being strictly connected with them (see, for instance, migrants’ punitive forced transfers from one detention centre to another – Gill, 2016). Some authors challenged the inside/outside opposition in relation to detention, looking at practices of incarceration that take place out of the prison (Moran et al., 2013) or considered the ‘intimate economies’ (Conlon and Hiemstra, 2016) of detention. These are important analyses that refocus the analysis on the different circuits (of people, goods and practices) that constitute the carceral space. In particular, recent contributions have turned attention to ‘the different forms that enforcement can take’ and to the ‘transforming spaces of enforcement’ (Burridge et al., 2017: 241).

Yet, also when detention and detention sites are not taken as the starting point to analyse disciplinary mechanisms of containment and economies of illegality, ⁹ containment is conceived in terms of specific spatialities of physical confinement that proliferate beyond official detention centres. In this article, we take detention geographers’ indication to move away from the infrastructure of reception centres one step further: we point to effects of containment that are partially disjoined from spaces of direct control and confinement and that consist, rather, into forced mobility induced by state policies, migrants temporarily stranded in border zones and in the disrupting and deceleration of migrant geographies. Moreover, we study hotspots from the angle of the government of mobility at large: how does the Hotspot System channel migration movements in relation to the broader economy of capturing, diverting and slowing down migrants’ mobility? Finally, we do not conflate hotspots with carceral spaces; instead, we investigate their spatial productivity. This approach proved particularly important to understand the functioning of the Lampedusa and Lesbos hotspot as chokepoints in which migrants are partitioned, identified and sometimes preventively excluded from the channels of asylum, and to specify the outcomes of migrants’ immobilisation beyond the strict definition of incarceration. In this sense, the hotspot has been rightly defined as a ‘highly flexible informal mechanism for governing diverse migrant population’ (Painter et al., 2017). In other words, the hotspot responds less to a disciplinary and punitive rationale than to a ‘border tactic’ (De Genova, 2017a) for temporarily disrupting and channelling mobility.

The second body of literature, this article, is in conversation with that on island geographies. Alison Mountz talked about an ‘enforcement archipelago’ pointing to the need of studying islands not as isolated exceptional spaces but as part of a strategy of migration management that detaches control from the mainland and ‘shrink[s] spaces of asylum’ (Mountz, 2011: 126; Steinberg, 2005). Certainly, the location of hotspots on Italian and Greek islands speaks to this function and creates a ‘frontline archipelago’ (Garelli and Tazzioli, 2016a) at the Mediterranean frontier of the EU, setting up border enforcement outposts in the central Mediterranean and Aegean routes. Hotspots’ detachment from the mainland is in fact part and parcel of the logistics of managing arrivals and sorting profiles, the so-called ‘management of mixed flows’ at ‘entry points’, as the governmental language puts it Hotspots were in fact envisioned for containing migration away from the mainland (or from central places) as the literature on ‘enforcement archipelagos’ allows us to see. Building on this analysis, our study also looks at their implementation focusing on hotspots’ outcomes on migrant flows on the mainland (despite their stated mandate) and investigating how hotspots interact with other border management technologies. In other words, we are interested in how the remoteness of hotspot-islands becomes imbricated with the structures of migration management on land.

Finally, our work contributes to the discussion about migration and logistics, focusing on the ‘logistic spaces’ (Cowen, 2014) that are generated by and connected with the government of mobility. The main idea behind this approach is that mechanisms for blocking and obstructing migration should be situated within the broader context of the politics of mobility, a politics that strives to channel, control and select movements (Biao and Lindquist, 2014; Mezzadra, 2016; Mezzadra and Neilson, 2013; Popescu, 2015). Our contribution to this conversation focuses on the nexus between logistics and migration containment. We argue that containment is one of the main effects produced by the chokepoints and channels that the infrastructure of migration management – in its hotspot approach – institutes, and that such logistics of containment is a governmental response for obstructing migrants’ autonomous movements and for taming and dividing migrant multiplicities. The infrastructure of containment we refer to is not only limited to a repressive strategy (e.g. mechanisms for stopping and pushing-back ¹⁰ migrants) but refers also to various forms of ‘entrapment’ (legal, existential and spatial) that result from mobility control mechanisms like the hotspots. Thus, we engage with hotspots’ spatial productivity more than simply sign-posting them as markers of a repressive strategy.

A genealogy of border-hotspots

‘Hotspot’ is not a new term in the EU lexicon, and in fact it did not appear for the first time in the European Migration Agenda of 2015. The word ‘hotspot’ has long been used for designating different types of critical border sites, including zones targeted for security, military, and military-humanitarian interventions. Back in the early 2000s, in fact, the EU introduced the term ‘hotspots’ in policy conversations addressing crime and natural disasters, long before its deployment in the field of migration. In this early use, European agencies like Europol, for instance, referred to ‘logistical hotspots’ to indicate critical criminal sites that allowed ‘illicit’ transfers of people and goods. In the field of migration, the nomenclature of smuggling hotspots was introduced in discourses about ‘illegal’ migration and smuggling activities. For instance, the 2009 North Atlantic Treaty Organization Report on migration in the Mediterranean, Sicily and the Strait of Gibraltar talked about ‘hot spots for illegal migration’, referring both to unauthorised migrant crossings and smuggling centres. ¹¹ In this context, hotspots describe both the ‘hot’ places of migrants’ and refugees’ entry into the EU, and the logistics of migrants’ journeys, and smuggling sites inside and outside Europe. ¹² Even some third countries are named smuggling hotspots, e.g. Egypt was recently labelled as the new smuggling hotspot by Frontex. ¹³

As part of this genealogy, it is also important to trace the historical-political background of the Hotspot System, and to highlight its internal frictions (between member states, the EU and European agencies), as well as the multiple political and legal paths that led to the implementation of Italian and Greek hotspots. Keeping both the discursive and historical-political background in mind, three areas emerge as relevant to the hotspot genealogy, we trace here: hotspots should be studied in relation to migration smuggling hubs, hosting systems for asylum seekers and member states’ disobedience against the EU identification standards. ‘Migration smuggling hubs’ is the designation used by Frontex to indicate critical migrant routes that are object of governmental concern. In its 2012 Risk Analysis – so three years before the EU Agenda on Migration – Frontex talked about the Eastern Mediterranean route as ‘the undisputed hotspot for illegal entries to the EU’, ¹⁴ contributing to the securitarian approach to hotspots. The Italian hosting system for asylum seekers represents the second important pillar in the genealogy of the Hotspot System. In particular, Italy’s restructuring of the reception system in 2014 constituted to some extent the material backbone for the implementation of the EU hotspot approach. In 2014, in fact, Italy created new temporary reception centres called ‘CAS’ (Centri di accoglienza straordinaria, Extraordinary Hosting Centres) to respond to migrants’ increasing arrivals by spreading hosting functions across the whole Italian territory. This measure represented the humanitarian landscape against which hotspots infrastructures emerged: hotspots as a securitarian filter that would allow the access of some migrants to the Italian hosting system. ¹⁵ The third important component of this brief historical-political background of the Hotspot approach consists in the repeated frictions between Greece and Italy on the one hand, and the EU on the other, regarding migrant identification procedures. In 2015, Italy and Greece were admonished by the EU ¹⁶ for not having fingerprinted all incoming migrants between 2013 and 2014. The Hotspot System can be seen as the spatialisation of the EU attempt to discipline Italy and Greece to comply with the EU identification procedures and forcing the countries to share the collected fingerprints in the EURODAC ¹⁷ database. As Wendy Brown has eloquently captured, ‘through the hotspots, the Juncker Commission pushes back against assertions of nation-state prerogative with EU technocracy and biopolitics; its agencies regain EU control over immigration to Europe and control over the migrants themselves’ (Brown, 2017: 3). Building on this genealogy of the EU implementation of flexible infrastructures called ‘hotspots’ on both a discursive level (the EU hotspot-lexicon that designates critical sites) and within the field of migration governmentality (the political ground and the ‘antecedents’ that paved the way to the Hotspot System), we can now move to the analysis of the Lampedusa and Lesbos hotspots.

Illegalisation on the spot in Lampedusa

We start our journey across the hotspots’ landscape from Lampedusa island, the southernmost tip of Europe, the island-mark of the central Mediterranean route into Europe and the first hotspot to be opened in September 2015 with the idea that it would work as a model for all others. The first year of the Lampedusa hotspot was a time of experimentation with the flexible structure and multiple mandates of the hotspot. The Lampedusa hotspot is located in a hollow in the central part of the island on military premises. The Contrada Imbriacola facility that now hosts the hotspot has been serving several functions over the years – from first aid centre to hosting centre and now hotspot. How do migrants end up at the Lampedusa hotspot? When the hotspot opened migrants tended to land at Lampedusa as part of Operation Triton, the EU border enforcement operation led by Frontex that started in November 2014. Upon disembarkation in the small harbour of Lampedusa, migrants are moved to the hotspot where the Italian police and Frontex officers perform those identification procedures that are the defining feature of hotspots as heightened mechanisms of border surveillance and technologies for filtering mixed flows at entry. How does the hotspot mandate for the full identification of incoming migrants actually play out at Lampedusa? Which geography of mobility does it create for migrants and refugees – which diversions and disruptions of their mobility does this mechanism perform? More than technologies to enhance security, identification procedures perform a preventive illegalisation and often result in forms of punitive entrapment for some migrants on the island.

Let us first focus on preventive illegalisation practices via screening procedures. As part of the Lampedusa hotspot procedure at its outset, all incoming migrants received an Italian Home Office questionnaire bearing the question, ‘What is the reason for your being in Italy?’ with four options to choose from: ‘poverty’, ‘family reunification’, ‘work’ or ‘other reasons’. Until March 2016, ¹⁸ war and persecution were not explicitly listed options and had to fall under the rubric ‘other’ – an odd bureaucratic choice for a form meant to discern refugees from economic migrants (Garelli and Tazzioli, 2016a). Migrants, activists and hotspot workers we interviewed report that police officials often filled in the document for migrants, ticking the box that would decide their destiny (either protected refugees vs. destitute undocumented economic migrants) simply based on migrants’ country of birth, and only allowing Eritreans, Syrians and Iraqis – those eligible for the EU Relocation Programme – to start the asylum procedure. All others, including large numbers of people coming from West African countries, were instead illegalised on the spot: they were excluded from protection mechanisms and classified as undocumented economic migrants.

What happened to these illegalised people during the first months of operation of the Lampedusa hotspot? They were officially asked to leave the country within seven days through an expulsion order called ‘7 days decree’ and asked to leave the national territory, implying that they should pay for their own removal. Thus the hotspot mechanism produced a large population of illegalised people on the national territory, living in destitution in Italy or across Europe. Those whom the Lampedusa hotspot did not classify as eligible for protection ended up remaining on the European territory as migrants, irregularised by the workings of the hotspot system itself. Here we see how the filtering mechanism envisioned as part of the securitarian approach of the hotspot results in internal flows of undocumented migrants, staging an uncanny implementation of the securitarian mandate for full identification. Hence, the hotspot is not only a mechanism to prevent most people from accessing the asylum procedure; it is also an instrument that results in adverse incorporation processes on the national territory, invisibilising migrants’ presence. The few accepted as asylum seekers are either forced to claim asylum in a country where they do not want to be (Italy or a third country not of their choice); those labelled as non-eligible for international protection end up living destitute lives, as undocumented migrants in Italy or elsewhere in Europe.

The Lampedusa hotspot has also been performing a sort of punitive entrapment of migrants on the island: indeed, refusing being fingerprinted for migrants would imply relinquishing the freedom to decide where to claim asylum in Europe. ¹⁹ While this refusal jams the swift screening system of the hotspot frontline (Garelli and Tazzioli, 2016b), the spatial outcome onto migrants’ lives is that those who refused to be fingerprinted ended up stuck on the island, trapped there for their non-compliance with the hotspot regulations. For instance, a group of 250 migrants, mainly Eritreans, who arrived at Lampedusa on 5 November and 4 December 2015, refused to be fingerprinted and organised a protest on December 17 to claim their right to choose where to seek asylum in Europe. Migrants who resist fingerprinting procedures at the Lampedusa hotspot end up in a spatial trap: they do not have access to the protection system and are indefinitely detained on Lampedusa, far away not only from their aspired-to refuge, but also from the mainland. Ten of the 250 migrants who refused to be fingerprinted were still on Lampedusa after three months since their arrival, while the others were transported to Sicily, having finally given their fingerprints out of exhaustion or as a result of the use of force on the part of the police. ²⁰ Few months later, we met few of them in the French city of Marseille, trying to reach Calais passing through Paris.

Greece: From space of transit to space of containment

The role of Lesbos Island radically changed since the inception of hotspot and during its first year of operation. Given its proximity to the Turkish coastline, Lesbos has been a landing place for people trying to cross the Mediterranean into Europe via the Aegean Sea. This function persisted also during the first months of the hotspot, when migrants were processed at differential speeds depending on their nationalities and were generally let go to the Greek mainland. But the transit-point function of the Lesbos hotspot rapidly ceased and Lesbos became a hotspot-island with an explicit containment and border enforcement function. This transformation of the Lesbos hotspot into a mechanism of capture that directly affects migrants’ lives is one of the most visible outcomes of the EU-Turkey agreement and of the EU effort to stop migrant movements across the Aegean sea.

Let us first focus on the first months of operation. The Moria hotspot in Lesobs opened in October 2015 and worked as a centre of first identification more than as a space of detention until 18 March 2016, when the EU-Turkey deal was approved. During the Moria hotspot’s five months of operation, identification procedures were a preliminary step that would grant migrants a permit to legally reside in Greece for either six months (Syrian nationals only) or one month (all other nationalities). This permit would allow them to get a ferry ticket to Athens and hence move on with their journey to other European countries along the Balkan corridor. During this initial period, migrants’ fingerprints taken inside the Moria hotspot were not transmitted to the Eurodac database. Yet, even in this initial stage, the hotspot system worked as a selective and flexible mechanism, staging, for instance, a specific treatment for the nationals of North African countries (e.g. Tunisia, Algeria and Morocco) starting from November 2015. Labelled as ‘North African migrants’ by Greek authorities, these people were in fact subjected to immediate detention. During the same time period, Pakistani migrants were also excluded from the free ride to the mainland and from the channels of asylum by a political decision that dates back to the beginning of 2016 (Tazzioli, 2016). Pakistani migrants usually refuse to claim asylum in Greece, aware of the likelihood of a rejection and aware that asking for international protection corresponds to a form of humanitarian containment: ‘the asylum application process, for them, was experienced as immediate detention’ (Spathopoulou, 2016). However, due to the lack of detention capacities in Lesbos, most of them were not identified by the Greek police.

During these first months, a spatial hierarchy was established on Lesbos between those migrants UNHCR considered ‘vulnerable subjects’ and all other migrants. The first group was sent to the Kara Tapé camp in Lesbos forced to live in containers but having the possibility to exit the camp during the day. The second group was held in the Moria hotspot. During a visit to Kara Tapé camp in April 2016, Martina Tazzioli observed that most people hosted there (639 at that time) were Syrian families. As we went back to visit the camp in July 2016, we observed a similar composition of Syrian families. Vulnerability and nationality cannot, in fact, be fully disentangled from one another. It can be argued that the former is always inflected by the latter and changes according to the geographical and political context.

The transformation of Lesbos from a dock-and-go site to a space of differential containment is the outcome of two interventions of border politics restructuring that occurred in the span of just a few months, from November 2015 to March 2016. The first one is the trickle-down effect produced by the progressive border closure along the Balkan route, starting from North (Slovenia) and moving on southward (Macedonia). A first access restriction at the Greek-Macedonian border (Idomeni) was put into place on 19 November 2015, when Macedonian authorities established an exclusionary criterion for accessing the country based on nationality: only migrants coming from Iraq, Syria and Afghanistan could enter Macedonia. The second intervention consists in the implementation of the EU-Turkey Deal signed on 18 March 2016 and preceded by negotiations between the EU and Turkey at the end of February that were re-launched after several months of stalemate (when Turkish President Recep Tayyip Erdogan threatened to send migrants to Europe as ‘human bombs’).

The transformation of Lesbos into a space of containment, we argue, should not be read through the lens of the prison-island or carceral site. Far from being sealed spaces or zones of total detention, the governmental spaces for identifying, containing and selecting migrants (including hotspots) are crisscrossed by multiple movements and formed by differential circuits and channels of mobility. The nationality-based principle for managing migrants’ presence on the island we mentioned above is a case in point. Moreover, some were allowed to leave the island relatively quickly; others remained blocked there for months; and some migrants were forcefully removed and brought back to Turkey. The hotspot, in fact, is not a monolithic space simply established to enclose and stop migrants’ mobility: it multiplies migration profiles and differentiates migrants’ juridical profiles. To put in a nutshell, hotspots are procedures for channelling mobility and dividing migrant multiplicities, more than detaining an indistinct group of migrants in a detention facility. As Doctors Without Borders remarked in an interview, the effective partitioning function of the hotspot started when Lesbos ceased to be a space of semi-autonomous transit to the mainland:

soon after the opening of the hotspot, Lesbos functioned as a space of first landing and as a space of transit for almost all migrants, in a quite homogenous way. With the first restrictions made on a nationality basis – and in a more consistent way with the implementation of the EU-Turkey agreement – Moria became a space where migrants started to be subjected to multiple divisions and bureaucratic steps. ²¹

Algerians, Moroccans, Tunisians and later on also Pakistanis were targeted for arrest and were prevented from leaving the island, even after they obtained the one-month temporary permit. Few people were actually apprehended, and the illegalisation of Pakistani migrants was managed rather through a different type of twofold trap: the trap of asylum and a spatial trap. That is, those who decided to go to the hotspot to be identified in order to be eligible to submit an asylum claim were immediately detained inside the facility. In this respect, the centre started to operate as a closed facility for Pakistanis only. Migrants from other nationalities were de facto trapped there due to the geographical boundaries of the island and the denial of the special permit needed to travel to Athens by ferry.

From a space of transit, the Hellenic peninsula became a site of containment where migrants are partitioned and divided. Similarly, many transit zones in Greece – like, for instance, the Terminal 1 of Pireus port – were transformed into spaces of control, identification and indefinite wait for migrants. The concomitance of these two routes’ closures – i.e., the Balkan corridor and the sea route from Turkey to Greece – produced a situation where relocation became the only possibility for migrants to move away from Greece.

The Lesbos hotspot and the changes that occurred after its opening had a considerable impact on the spatial economy of migrants’ arrivals. Yet, the redefinition of the geographies of migrant sea journeys to Lesbos cannot be entirely grasped if we narrow the focus to the Greek shores only. The spatial economies connected to the hotspot should be analysed together with the strategies of mobility containment put into place as a result of the EU-Turkey Deal and that enacted primarily by the Turkish Coast Guard. The number of migrant arrivals in Lesbos dropped from 1400 to 70 per day in the span of few weeks, between the end of March and early April 2016. In turn, this produced transformations in the geographies and economy of migrant crossing. Due to the constant patrolling of the Turkish Coast Guard, migrants tended to arrive in the Southern part of the island, and no longer in the Northern one, that, while closer to the Turkish shore, was more heavily controlled. Simultaneously, because of increased likelihood of push-backs before migrants would reach Greece, smugglers changed approach. As a volunteer of the Spanish rescue organisation Proactiva told us,

smugglers in Turkey now sell a sort of three-attempts tickets for $1000. These tickets can be used three time by the same person to try to reach Lesbos, as a response to the increased possibility to be intercepted at sea and returned to Turkey. ²³

From hotspots to channels

Conceived as a site of mobility disruption, the hotspot represents a complex set of infrastructures, identification procedures, logistical nodes for managing migrants’ landing and regaining control over migration movements. Up to now, we have shown how the Hotspot System impacts on migrant lives and geographies, producing a population of illegalised migrants on the territory, multiplying migration profiles through the implementation of temporal borders and generating effects of spatial confinement. Yet, as explained in the Introduction section, the modes of containment produced by the Hotspot System are not limited to physical entrapment on the islands; rather, they also entail forcing migrants to undertake convoluted geographies. These are the deferred containment effects performed by the Hotspot System. They take place and become visible in sites that are spatially dislocated from hotspot infrastructures. Simultaneously, the channels of the Relocation Scheme depart from the hotspots. Thus, the hotspot’s bottleneck function should be seen not as a blind alley but as an exclusionary filter for slowing down access to institutional channels of departure.

The Relocation Scheme was launched in simultaneity with the Hotspot Approach in the European Migration Agenda. It was in fact conceived as an EU programme for transferring ‘persons who are in need of international protection’ ²⁴ from Greece and Italy to other member states, hence in direct connection with the hotspot filtering mechanism. However, the ‘need of protection’ clause is not the only condition for entering the channels of the Relocation Scheme: the condition of being ‘at risk’ is in fact fundamentally inflected by the nationality criterion. That is, only migrants whose nationalities register a recognition of their international protection claim above 75% (as part of EU countries’ processing procedures) are eligible for Relocation.²⁴

Channels are usually associated with transportation (of goods and people) and mobility, as well as with the logistics that moves commodities from one place to another. Embracing this terminology to explain the hotspot approach means calling attention to the spatialities that grow out of migration governmentality – i.e., the material logistics and the legal procedures through which states attempt to regain control over migration movements (Tsing, 2000; Neilson and Rossiter, 2010). As Sandro Mezzadra has convincingly argued, the ‘channelling and distribution of migrants through the intervention of both human actors and logistical infrastructures are crucial aspects of the process and dynamics of contemporary migration’ (Mezzadra, 2016: 38). This is the perspective we engage with in this article, as we study the logistical infrastructure that is in place inside and ‘around’ the hotspot (from the moment when migrants land on hotspot-islands to the selective transfer to hosting centres), together with relocation channels instituted by the EU as the hotspot’s twin policy. However, our focus on the channels of migration management is not meant to project the image of smooth spaces of migration control. Nor do we argue that these channels – established on a national level or as part of the EU political agenda – are actually working. On the contrary, our take on migration channels is connected to and built upon an analysis of containment effects: it is around such a nexus between channelling and containing mobility that our perspective develops. By speaking of ‘channels,’ our goal is also to point to an overlooked aspect in analyses about the logistics of trade and human mobility. Channels do not only haul mobility; they do not simply allow mobility to happen in a material sense. Channels also steer and contain movements. Channels contain and decelerate flows; they do not simply serve as the material infrastructure of circulation. More importantly, while the channels of trade circulation are devised for regulating and making efficient the circulation of goods, the institutional channels of migration mobility have been put into place for regaining control over migrant autonomous movements, producing blockages and introducing exclusionary partitions. Migration ‘cannot be stopped, only routed’, Wendy Brown succinctly puts it, as it is glaring from the ‘intricate ad hoc network of spatial and governing technologies for diverting, channelling, policing and governing migrant flows’ (Brown, 2017: 3). Thus, we suggest that future research should further explore the mutual relationships between modes of containment, as techniques of mobility disruption, and the channels of migration governmentality, as ways for regaining control over unruly mobility and introducing exclusionary criteria of access.

Indeed, the channelling function of hotspots aims to decrease the intensity – quantitatively in terms of sheer numbers or qualitatively in terms of unruliness – of migrants and refugees’ mobility (to and across Europe). The Dublin Regulation and the Relocation Programme stages forced stops (country of first entry) and forced mobility (relocation destination). Hotspots are sites where the first selection between those who are eligible for relocation and others is made. Thus, while the system actually strengthened the role of Greece and Italy as the buffer zones for the European core, it was officially presented as a mechanism for sharing the ‘burden’ of the refugees across Europe – i.e., moving them around in a regulated and legal way from the places of first arrival to other member states. In reality, the slowness of the Relocation process and the reluctance on the part of many EU countries to take asylum seekers from Italy and Greece, together with the closure of the Balkan corridor and the suspension of Schengen in six member states, ²⁵ have amplified the role of Greece and Italy as frontier spaces of containment, outposts where migrants’ are blocked at the Mediterranean frontline and circulation across the European space is contained (Kasparek, 2016).

It is important to mention that many migrants who were considered eligible for the Relocation procedure finally refused it. ²⁶ What the EU promoted as a safe route towards Northern European countries actually resulted in the impossibility for migrants to choose where to claim asylum. In other words, the ‘safety’ of the (exclusionary) channels of relocation represented a material obstruction to migrants’ right to choice. Hence ‘protection’ started to appear as a synonymous of capture for some migrants, i.e., in their being moved and channelled by state policies. At the European Asylum Support Office Headquarters in Athens officers confirmed that, indeed,

when the relocation started, it was presented to migrants as a way to continue their journey safely; yet, the majority preferred to do that in an autonomous way. Only after realizing that they were stranded in Greece with no alternatives, Syrians opted in mass for relocation. ²⁷

In order to show how the Hotspot System disciplines migrants’ unruly geographies, by producing effects of containment beyond detention – the perspective we introduced in the first part of the article – we conclude on the multiplication of hotspot-like spaces and on the forced internal transfers from Northern Italy back to the hotspots. Since France’s suspension of Schengen in May 2015, the Italian city of Ventimiglia has become an unofficial chokepoint for migrants heading to France and the United Kingdom. Ventimiglia is the closest Italian city to the French border, and together with Calais, it is one of the main border zones that mark and slow down the way to the United Kingdom for migrants arrived in Italy. It is not by chance, in fact, that the Prefect of the Department of Civil Liberties and Immigration of the Italian Home Office cautioned against the possibility of reproducing ‘Calais in Ventimiglia’. ³⁰ Ministry of the Interior Morcone put it even more crudely when he described Italian authorities’ interventions at Ventimiglia as aimed at ‘lightening the frontier up’ and redistributing people across the territory to avoid organised concentrations of migrants. Yet, unlike Calais, which has become primarily a site of indefinite strandedness and wait, Ventimiglia can be considered a space of transit – not in a linear South to North sense, but as a space where migrants’ geographies become fragmented and where forced inverse routes take place. The emergence of Ventimiglia as a critical migration border zone should be traced back to 2011, when France firstly suspended Schengen due to the sudden increase of Tunisian migrants who landed in Italy soon after the outbreak of the Tunisian revolution.

Rail station of Ventimiglia, July 2016. Migrants arrive in Ventimiglia, at the French-Italian border, daily. They come Milan, from Rome, or other Italian cities. In the city of Ventimiglia, they can either go to the ‘temporary hosting centre’ run by the Italian Red Cross, where will be able to stay only one week and will be identified (although not fingerprinted); or, they can opt to remain in the street, a choice often motivated by fears that the temporary hosting system could turn out to be a ‘humanitarian trap’. Even if there is not an official hotspot, nor a detention centre there, the urban space of Ventimiglia has been working as a hotspot-like place, although with important differences with respect to hotspots in Southern Italy. For instance, Frontex and Easo officers are not there: identifications, preventive illegalisation and racialised partitions are made by the Italian police. Migrants who refuse to go to temporary hosting centres are subjected to police identification procedures. Actually, the fact of being ‘hosted’ in the temporary hosting centre is not a guarantee of not being apprehended in the streets. Most of the time, the partition between those who are tolerated on the territory and those who are transferred to detention centres or deported is made on the basis of the nationality: migrants from ‘North African’ countries are left outside the hosting centre, and in a way they are excluded from the channels of the asylum.

Migrants’ forced transfers by bus from Ventimiglia to the hotspot of Taranto (a city located in the Italian southern region of Puglia and 1200 km away) became a kind of weekly routine starting from July 2016. Actually, these forced internal transfers do not always result into linear North-to-South patterns: as a Parliamentary Report shows that the 48 Sudanese migrants deported to Sudan on 24 August 2016 were transferred to the hotspot of Taranto (August 21) after being apprehended in Ventimiglia (August 19). From Taranto they were taken by bus to the airport of Turin, therefore back to Northern Italy, where they were put on a plane and repatriated to Khartoum. ³¹

Forced transfers to Southern Italy, however, are not permanent moves back. On the contrary, migrants usually ‘show up’ again in Ventimiglia after few days, after escaping the hotspot and returning to the French-Italian border by train. Thus, more than detaining them in Southern Italy, Italy’s measures ended up protracting migrants’ journeys and forcing them to restart their route to France. In fact, one of the ‘hotspot effects’ consists precisely in disrupting the autonomy of migration movements. The same forced counter-route from Northern Italy to the hotspot of Taranto has been used by the Italian police for removing migrants from the Swiss-Italian border, more precisely from the city of Como and from Milan; ³² in some cases, this happened also to migrants who had already claimed asylum, preventing their being brought to a detention centre or to a hotspot. ³³ Como has become a critical border zone quite recently, in summer 2015, when Switzerland enforced border controls and push-back operations. Milan is one of the cities, together with Bologna, that started to function as a hotspot-like space: at the Police Headquarters (Questura) of Milan, migrants’ nationality is the criterion for deciding who is allowed to claim asylum and who is illegalised on the spot, with a decree of expulsion, in a similar way to what happens inside the hotspot of Lampedusa (Sciurba, 2016).

Looking at the Hotspot System through migrants’ scattered spatialities and governmentally mandated mobility channels enables grasping two simultaneous spatial transformations. On the one hand, hotspot-like spaces spread across the territory (Ventimiglia, Como Milan), not as multiplication of carceral spaces, but rather as chokepoints and sites of mobility disruption. This relates to effects of containment beyond detention that often do not generate immobility but rather forces migrants to move away or to undertake convoluted geographies. On the other hand, a series of channels of forced mobility opened up, in connection with the hotspot, in order to move migrants back and refugees around, as ways for obstructing and regaining control over autonomous migrant erratic geographies.

By speaking of EU’s fight against migrant secondary movements what we do not want to suggest that the EU is stopping unruly mobility, keeping migration movements across Europe to the minimum. Rather, our argument is that migrants are kept on the move and are forced to undertake more and more erratic and diverted journeys, as a result of the many internal transfers they are targeted for. Moving migrants back and moving refugees around appear to be the two main modalities through which the EU is coping with migration movements which partly dodge or remain out of the institutional channels, like the Relocation Scheme.

As scholars working on the EU internal borders and on the effects of the Dublin Regulation have demonstrated, migrant unruly geographies are not governed by stopping migrants but through their own mobility, that is to say by keeping them in transit (Fontanari, 2016; Hess, 2012; Schuster, 2005). Yet, our point consists less in bringing attention to the hypermobility of migrants produced by tactics for disciplining and taming secondary movements than in showing the effects of containment generated by moving migrants back and around. The internal deportations that took place in Italy (from the border with France and with Switzerland back to the Italian hotspots) show that a state of incessant mobility and transit is enforced as a migration management strategy. This forces migrants to figure out how to partially restart their journey and generates their temporary containment.

Conclusion

The Hotspot System can be analysed from the point of view of its institutional function, i.e., as an enclosed site for temporarily detaining migrants while they are identified, labelled and partitioned upon landing. Based on ethnographic engagement in the Italian and Greek hotspots, this article has scrutinised the exclusionary criteria and the governmental practices through which migrants are managed and divided through the hotspot system. Instead of entirely focusing the analysis on the institutional spaces of migration governmentality defined by states, we have moved beyond the fences of the hotspot and mapped the forms of containment beyond detention that are connected to the establishment of the EU hotspot approach to the migration crisis. We have investigated the mechanisms of preventative illegalisation of migrants and the heterogeneous channels of mobility (the institutional channels of relocation and the channels of internal forced transfers) that are connected to and enforced by the Hotspot System. Conceived as sites of mobility disruption, hotspots are flexible chokepoints that are actually used both for regulating migrant arrivals and slowing down their mobility, as well as for regaining control over migrants’ secondary movements. In closing this article, we would like to sketch the trajectory of a theoretical engagement with the notion of ‘containment’, a trajectory that – as we have enacted it in this article – consists in tracing the genealogy of heterogeneous forms and practices of containment that are linked to the hotspot approach.

Nicholas De Genova has convincingly pointed out that ‘detention, like deportation, is a term that has no distinguished pedigree in the history of political ideas and legal concepts’ (De Genova, 2017b: 6). ³⁴ Building on this indication, we suggest that ‘containment’ – like deportation and detention – remains a fundamentally under theorised and blurred notion in migration literature as well as in political theory. De Genova importantly points to the ‘economy of detainability’ and the ‘economy of deportability’ of migration (De Genova, 2017b: 1) as areas where more theoretical engagement is needed. Looking at the spatialities triggered by and connected to the hotspot approach way beyond the fences of detention centres, we have sketched the economy of containment that characterises this approach. While carceral geography scholarship has persuasively drawn attention to the transformations of the spaces of confinement, shifting in part away from an exclusionary focus on detention centres, we call for a research agenda that focuses on the economy of containment, grasping this latter beyond forms of direct control and physical entrapment. This involves, we suggest, remapping the European geographies of migration governmentally by looking at the institutional channels and chockepoints established for regaining control over autonomous migration movements. Yet, far from fostering mobility and producing frictionless logistic spaces, the proliferation of hotspots-like sites and of selective channels of protection shore up a whole economy of containment that produce both temporary strandedness and forced convoluted mobility.