Surveillance at sea: The transactional politics of border control in the Aegean

Abstract

The relationship between vision and action is a key element of both practices and conceptualizations of border surveillance in Europe. This article engages with what we call the ‘operative vision’ of surveillance at sea, specifically as performed by the border control apparatus in the Aegean. We analyse the political consequences of this operative vision by elaborating on three examples of fieldwork conducted in the Aegean and on the islands of Chios and Lesbos. One of the main aims is to bring the figure of the migrant back into the study of border technologies. By combining insights from science and technology studies with border, mobility and security studies, the article distinguishes between processes of intervention, mobilization and realization and emphasizes the role of migrants in their encounter with surveillance operations. Two claims are brought forward. First, engaging with recent scholarly work on the visual politics of border surveillance, we circumscribe an ongoing ‘transactional politics’. Second, the dynamic interplay between vision and action brings about a situation of ‘recalcitrance’, in which mobile objects and subjects of various kinds are drawn into securitized relations, for instance in encounters between coast guard boats and migrant boats at sea. Without reducing migrants to epiphenomena of those relations, this recalcitrance typifies the objects of surveillance as both relatable as well as resistant, particularly in the tensions between border control and search and rescue.

Introduction

A key element of the various sea border surveillance policies in Europe is the application of techniques of visualization that are connected by ‘interoperability’ so as to arrive at ‘situational awareness’ (e.g. the European Border Surveillance System [Eurosur] regulations, which became operational at the end of 2013). Yet what kind of seeing is involved in the advanced surveillance systems that operate in the Aegean? How does this seeing affect the migrants themselves? The intensification of border control at sea, in particular in the Aegean (e.g. the launch of the European Border and Coast Guard Agency), the militarization of operations (e.g. NATO support), the criminalization of migration (e.g. hotspot policies and intensified cooperation with Interpol) and the externalization of border policies (e.g. the EU–Turkey deal) require a rethinking of the relationship between policies of visualization and actual interventions such as search and rescue (SAR) operations, in particular now that a growing number of cases are reported in which the loss of life has occurred during and partly through rescue itself (Heller and Pezzani, 2016: 2).

This article engages with the recommendation of the c.a.s.e Collective (2006: 459) that the study of border surveillance ought to pay more attention to the mutual interaction between the subjects and objects that are studied and the possible resistance of the latter. As Galis et al. (2016) argue, ‘thus far, migration studies – even the most critical studies – have emphasized technologies used by those who have sought to block access to Europe’. Instead, technologies need to be studied ‘not only in connection with the rhetoric of those who introduce [them] but also in light of how [they are] materialized in human bodies through clandestine border-crossing practices and material configurations (artifacts)’ (Galis et al., 2016: 8).

To do justice to this plea for symmetry, this article aims to introduce a conceptual perspective that questions the boundary between border surveillance and SAR at sea and allows for a deeper engagement between the subjects and objects of visualization. By elaborating on some recent contributions to border, mobility and security studies a political ontology is suggested that is sensitive to the way the objects of visualization bring themselves into action in surveillance practices at sea. This perspective aligns with Walters’s (2015) emphasis on, first, processes of visible contestation and politicization; second, the processes of exclusion that are at stake; and third, the realization of a more intimate engagement with different styles, forms and materials in surveillance practices. In order to specify this ontology the article engages with proposals that emphasize ontological questions in the study of security practices (Mitchell, 2014; Schouten, 2014; Mezzadra and Neilson, 2013). In general, ontology here is understood as a non-epistemological way of describing the mutual interaction between objects and their relations (Latour, 2016; Salter and Walters, 2016: 4). This concept will be explored in more detail in a later section on the layout of the conceptual architecture.

The first section of the article considers how ways of depicting associated with Eurosur’s promises of ‘situational awareness’ constitute a form of operative vision (Farocki, 2004) in the Aegean. Predominantly engaging with the work of Amoore (2009, 2013, 2014; Amoore and Raley, 2016) and Bigo (2014, 2015), an alternative conception of the kind of politics at stake in such border surveillance unfolds. Three aspects of the relationship between vision and action in border control practices can be distinguished: intervention, mobilization and realization. The analytical potential of these concepts from science and technology studies will be applied to three fieldwork excerpts in the following section. In order to ask, in the final sections, what kind of politics is involved in the operational situations that constitute the material instantiations of Eurosur, it is argued that it is crucial to consider the tensions between surveillance and SAR. These cases reconstitute a political topology that is not wholly structured by a ‘surveillance apparatus’, but which transacts the workings of such an apparatus with migrants, vessels and tactics at sea (cf. Pallister-Wilkins, 2015). Therefore, the article posits a transactional politics, which highlights both the enmeshed character of vision and action, of surveillance and its objects of vision, and the recalcitrance of those objects, which give rise to actual political events.

The fieldwork for this article was conducted in September of 2014. From the spring of 2015 onwards the influx of migrants increased significantly. However, there is continuity between the situation in the Aegean in 2014, the situation in 2015 when the numbers peaked, and after April 2016 when the EU–Turkey deal went into effect. All the ingredients that are central to the fieldwork – such as the large numbers of casualties, intensified surveillance at sea by the coastguard, support by Frontex, the deployment of technologies aimed at visualization, collaboration with military police of member states, the presence of patrol vessels with both security and humanitarian objectives and responsibilities, partnerships with the private sector and the responsibilization of local organizations – continue in the current situation. What has changed since then are the number of migrants, the media attention, the nature and scope of the war in Syria and the conflict in the Middle East, as well as the societal and political pressure on European member states and the EU to take care of the ‘crisis’. These circumstances, however, are less central to the argument and conceptual claims, and the empirical situations that will be described continue to be exemplary of current surveillance practices. In fact, analyses of these continuities in part reveals why it has been so difficult for European authorities to contain the disruptive consequences of their own surveillance practices.

The intermingling of vision and action

Without exaggeration, the Aegean can be regarded as a sea under surveillance (Topak, 2014; Triandafyllidou and Dimitriadi, 2013). Although the number of migrants varies and the nature of the interventions and forms of collaboration between member states and between the EU and other countries change, the influx of ‘irregular migrants’, as policy terms have it, presents a situation that the European border authorities consider an emergency of an almost permanent kind. ¹ Surveillance at sea has long had a major role in the events that make up the visuality of clandestine transit, territory, border patrolling, human smuggling, migration monitoring and the spectacle of illegality (Adey et al., 2015; Tazzioli, 2015; Jansen et al., 2015). The tensions between maritime border control and legal obligations to rescue those in peril have had profoundly transformative effects for European politics of unauthorized migration and asylum (Klepp, 2010; Nyers, 2013). While authorities have attempted for years to keep border patrolling and SAR apart (Klepp, 2011; Basaran, 2015), their tensions have since formed the occasion for significant shifts in operations at sea – such as Mare Nostrum and Triton – and the development of a ‘Common European Asylum System’ (Heller and Pezzani 2016).

The character of surveillance has been changing in recent years. As Topak (2014: 823) notes, surveillance systems in the Aegean Sea and on the islands tend to move from ‘patrolling-driven’ to ‘intelligence-driven’ operations in which ‘patrolling the sea with various types of vessels and air units is increasingly combined with “smarter” systems, such as radars, satellites, and coordination centers’ (Topak, 2014: 823).

Surveillance at sea displays the tight and tense relationship between borders and boundaries as empirical phenomena and conceptual categories (e.g. Johnson et al., 2011; Jones, 2008; Paasi, 2009). Mezzadra and Neilson (2013) have used the tensions between empirics and concepts in a productive sense to identify and analyse ‘border struggles’. For them, the border is

not so much a research object as an epistemological viewpoint that allows an acute critical analysis not only of how relations of domination, dispossession, and exploitation are being redefined presently but also of the struggles that take shape around these changing relations. (Mezzadra and Neilson, 2013: 19)

These struggles are conducted with all manner of technological and in particular visual devices (Tazzioli and Walters, 2016).

For a clue as to how these struggles unfold and how surveillance ‘sees’, consider the European border agency Frontex’s official description of the Eurosur system. The description mentions that Eurosur enables member states to ‘see and assess’ on the basis of three layers of information: events, operational information and analysis (Frontex, n.d.). Data on ‘events’ are combined with ‘operational’ information on interventions and ‘analysis’, mainly focused on risk assessments (Jeandesboz, 2008; Hayes and Vermeuelen, 2012; Rijpma and Vermeulen, 2015). It becomes clear that, although it is a self-designated ‘surveillance’ system, ‘seeing’ is only one part of what the system is designed to do. Indeed, the objectives of Eurosur are defined wholly in terms that one might consider strategic, interventionist or political, including a reduction of the number of irregular migrants and the number of deaths at the maritime borders, as well as an increase of internal security of the EU as a whole (Frontex, n.d.). ² In order to do so, Eurosur aims to develop ‘situational pictures’ for EU member states, such as National Situational Pictures and Pre-Frontier Situational Pictures (Jeandesboz and Simon, 2014).

This surveillance system, then, is much more than a system that sees and depicts. It is an operational device that produces images that constitute what Harun Farocki has called ‘operative images’, that is, ‘images that do not represent and object, but rather are part of an operation’ (Farocki, 2004: 17). Accordingly, it is possible to speak of surveillance as operative vision. Operative vision takes place in the registers of what Mirzoeff calls ‘visualization’, which he juxtaposes with ‘seeing’. Whereas ‘seeing’, according to him, is a bodily form of perception, ‘visualizing’ is an originally military form of depiction that is strategic and that schematizes things (such as battlefields) too big to be seen (Mirzoeff, 2015). Operative vision in the surveillance of the Aegean and the Mediterranean tends to be more akin to visualizing than to seeing. This is also how the militarized jargon of ‘situational awareness’ needs to be situated (cf. Suchman, 2015).

A first entrance to analyse this entanglement can be found in the work of Amoore, in particular where she grants the notion of the ‘mosaic’ a central place in border control (Amoore, 2013). As Amoore (2009) has stated, border control is not ‘primarily a way of seeing or surveilling the world, but rather a means of dividing, isolating, annexing in order to visualize what is “unknown”’ (Amoore, 2009: 25). Amoore (2013) emphasizes the idea that the sovereign decision is thick with lively mediations, and the politics of security is shifting from prevention to preemption and calculation (Amoore, 2014; Amoore and Raley, 2016; see also De Goede et al., 2014). The anticipatory logic of preemption does not seek ‘to forestall the future via calculation’; instead, it incorporates ‘the very unknowability and profound uncertainty of the future into imminent decision’ (Amoore, 2013: 9). The capacity to draw the borderline follows this logic by gathering together ‘multiple elements of what is thought to be known about a person – each element, in the singular, a mere possibility … in order to give appearance to an emergent subject’ (Amoore 2013: 81). Both the border, and the subject that aims to cross it, take the shape of a kind of a mosaic, the outcome of a process of ‘piecing together’ (Amoore 2013: 84). The resulting ‘composite image’ of the subject and the borderline is produced ‘via plural dividuated points in play simultaneously’ (Amoore 2013: 101).

If both border and subject are wholly ‘pieced together’ by the surveillant mosaic, the question becomes: In which political topology does surveillance operate? The analysis seems to suggest that it does not operate within a seamless technological vision because vision is always messily entangled with action. Encounters at sea are not merely occasions for ‘piecing together’ the objects of control, but messy transactions in which migrants often become recalcitrant elements in the political topology of border control, something Amoore and Hall (2014) hint at in an analysis of border camp activism and resistance.

A second entrance to unravel the relationship between vision and action is offered by Bigo (2015, 2014) who has reconstructed how different actors apply different concepts of the border as ‘solid’, ‘liquid’ or ‘cloudy’. This distinction identifies three different ‘fields of action’ (Bigo, 2014). First is the idea of the border as a solid barrier; this is related to a conception of ‘defense’. The second is related to fluid border checks and to practices of ‘policing and surveillance’ that concern processes of identifying, authenticating and filtering. The third is ‘the universe of the transnational database’. It is connected ‘to the digital and the virtual, to data doubles and their cohorts, to categorizations resulting from algorithms, to anticipations of unknown behaviors, to the prevention of future actions’ (Bigo, 2015: 59). As such, the third characterization of the border ‘pervades the second universe (and sometimes the first) by justifying technology and the management of surveillance at a distance in the name of the protection of a group of the population’ (p. 59). From this point of view, ‘borders are pixels. The sea no longer exists’ (p. 60).

A preliminary conclusion to be drawn from Amoore’s (2013) and Bigo’s (2015) analyses is to refrain from presupposing ‘coherence’ in surveillance, in which the patchwork of different techniques and various forms of visualization employed by different actors is presented as a seamless web. Presupposing coherence in surveillance compositions would threaten to adopt the rhetoric of technological advancement with which Eurosur and comparable systems are praised, thereby ignoring the ubiquity of friction (Tsing, 2005) and of seams (Vertesi, 2014), and the practical entanglements to which such friction and seams give rise. To productively attend to this inconsistency of vision without introducing presences that ‘simply’ exist, whether or not they are perceived, the above can engage with concepts from visual theory, most notably Gibson’s (1979) ‘ecological approach’ to visual perception. Some core notions of science and technology studies, in particular laboratory studies – in which the relationship between vision and action and the intimate connections between subjects and objects play a crucial role – can also be employed to maintain the centrality of friction and seams in discussions of surveillance.

Visualization not only requires data-gathering and imaging techniques, but is also intricately bound up with manipulations, the enrolment of devices (Haraway, 1988; Latour, 1999) and locomotion (Ingold, 2011). As Ingold (2011: 46) argues, we do not perceive ‘from a fixed point’, but along what Gibson (1979) calls a ‘path of observation’, a ‘continuous itinerary of movement’ (pp. 195–197). According to him, ‘vision is a whole perceptual system, not a channel of sense. One sees the environment not with the eyes but with the eyes-in-the-head-on-the-body-resting-on-the-ground’ (p. 205). First, vision is thus not just an activity of the eye or the mind, nor is it merely a logic, but it also is embedded in the environment in which visualization operates. And second, objects require responses. Therefore, the vision–action dichotomy and the possibility of a ‘neutral observer’ collapse. Visualizing inhabits an extended and distributed infrastructure. As a result, perception, according to Ingold (2011), is ‘a function of movement … what we perceive must, at least in part, depend on how we move’ (p. 47). More specifically, these lines ‘lay down the conditions of possibility’ (p. 85, emphasis in the original) for future relations that follow along different pathways.

To continue the study of the connection between practices and technologies of visualization we deploy three concepts rooted in science and technology studies and in particular in laboratory studies, namely intervention, mobilization and realization. These notions are sensitive to the idea that facts are situated in historical and practical settings without reducing their facticity to social conditions. They attend to the technologically complex accomplishment of even having objects within security practices, without simplifying their occurrence to an after-effect of securitized gazing. This would be one way to take into account the resistance that can be mobilized by migrants when in security’s sights (cf. c.a.s.e collective, 2006: 456). Resistance is not an out-of-grasp potential, but is actively generated along the mediations of a surveillance apparatus (Amoore, 2009).

According to Stengers (2000), the meaning of the experiment in modern science lies in ‘the invention of the power to confer on things the power of conferring on the experimenter the power to speak in their name’ (p. 89). This interpretation is readily applicable to the kind of programs states deploy in border surveillance policies, as authorities too are continuously engaged in technologically configuring ‘the power to speak’ in the name of things. And in surveillance as well objects made present are both open for relations as well as resistant to reductionism. From an ontological point of view, objects mobilized and intervened upon become both real (they ‘express’ themselves) as well as related (they are connected to other objects and observations) (Latour and Woolgar, 1979). They are thus of a recalcitrant nature (Miller, 2013: 39). Interventions, mobilizations and realizations are thus located in the various translations that take place across centers of calculation and various sites ‘in the field’ where data must be gathered and circumstances must be evened so as to make information storable and comparable.

Surveillance in the Aegean: three excerpts

Three fieldwork excerpts will help to conceptualize the intricate connection between vision and action in operative vision. These excerpts are taken from a fieldwork project in the Aegean that was conducted throughout September 2014. The fieldwork consisted of some 20 interviews and more informal conversations with a variety of people concerned with migration and border management in Greece, including activists, non-governmental organization (NGO) workers, migrants, coast guard and police personnel, policy makers and administrators. Over the course of September 2014, extended visits to key sites in the governance of migration were made by two of the authors, both in Athens and on the islands of Chios and Lesbos. This resulted in a number of opportunities to conduct participant observation of the practices of members of the coast guard, control center personnel, aid workers and activists. Informants, locales and practices were selected in order to explore the ways in which visual and digital technologies were impacting not only the operational and governmental response to unauthorized transit to and through Greek territories, but also the extent to which those technologies were changing the wider politics of migration.

Interviews and observations concentrated on the work and difficulties of visualizing realities in the Aegean relevant to unauthorized mobility. Fieldwork sites and informants were specifically selected in order to gain an understanding of how visual technologies were imbricated in, and changing, both practices of border control and unauthorized mobility. The multi-sited fieldwork thereby allowed us to reconstruct how vision was done differently at various locales in the field (cf. Marcus, 1995). However, beyond the variety of ways in which vision took place, we aimed to reconstruct how visualization at one site could be entangled with and make possible visualization at another site. At stake was not only the multi-sited character of a shared matter of concern – at the time often called ‘irregular migration’ – but translations and transport between sites. In keeping with our ecological approach to vision, we aimed to understand how vision in one particular locale was compiled from visualizations at other locales. The selection of sites and informants followed these pathways as much as possible. Fieldwork and subsequent analyses of our recordings and notes were not geared towards gaining an overview of the field that would be only accessible to researchers, but rather to understanding how people in a variety of locations worked to gain some sense of what was happening and what others at different locations were able to see (cf. Candea, 2010: 36–37).

An initial selection of sites and informants was made through discussion with experts familiar with local activism and government concerning migration. Subsequent contacts were made on the basis of emerging questions and insights. Interviews were recorded, and observations and informal interviews were jotted down, after which extended field notes were drafted.

One ~ In a large room located somewhere within the labyrinth-like headquarters of the Hellenic Coast Guard (HCG), housed on the docks of Piraeus’ harbor complex just outside the Greek capital, a screen flickers. In front of it sits an HCG official who, ordered by her superior, is toggling check boxes displayed in a list next to a map. The variously colored points, lines and planes display a section of the Aegean Sea. At the edges of the screen, lines indicate coasts, while in its middle the sea is shaded in grey, indicating depth, and interlaced with different, remarkably straight lines, indicating sea-lanes and zones of different kinds. Following the coast lines at some remove is a second line, marking off the ‘territorial waters’. As the HCG official toggles boxes, swarms of differently colored and angled triangles appear and disappear with each click. Her superior, a senior official in the Sea Borders Protection Directorate, explains that most of the triangles represent AIS-signals, which many vessels at sea are required to broadcast. He aims to demonstrate what the ‘system’ – Eurosur – has in store, the display of AIS-signals on a map being one of its more innocuous features. The room is part of the official Greek National Coordination Centre (NCC) of Eurosur and one of the nodes in a network that spans multiple such centers across EU-member states, the European Commission and the Frontex headquarters in Warsaw. ³

Eurosur purports to provide ‘pictures’ of what is happening in border zones, yet the overview on the screen demonstrates how abstract, selective and partial such pictures are: Where are the people housed in the makeshift barracks in the harbor of Chios? Where are the medical personnel of MSF? Where are the burial grounds of unidentified migrants? Where are the mobile radar units? Yet beyond the well-established fact that maps depict realities by thoroughly transfiguring them, often to the point of distortion and deception, it should be noted that the act of seeing this overview – even if distorted – is happening not ‘from a disfiguring distance’ but by way of a vast chain of interventions. It is not merely that the screen map filters out and focuses on certain matters of concern – location, administrative zone, speed, vessel typology, sea border – thereby assembling a type of vision with a specific intensity: who goes there! Depicting those matters only ever happens by way of interferences in the world. One simple click is connected to an entire spectrum of actions: installing AIS transceivers, enforcing compliance, negotiating agreements, designing a channel access method, categorizing vessel types, and so on. More generally, the expansion of surveillance at the external borders, for instance by implementing Eurosur, affects different national member state policies in different ways. European funds, framework programs and policy directives are implemented in national legislation and policy guidelines and accompanied with financial resources, technical support, staff members, training programs and technological and logistical infrastructures. In this context, ‘situational awareness’ promises to equip authorities with a ‘common pre-frontier picture’ that should deliver surveillance of border zones and people on the move as close to real time as possible. Or, as the HCG Head of Border Protection expressed, to ‘patrol through the screen’ (Interview, 16 September 2014). Currently, the main preoccupations of the Hellenic NCC are gathering reports – so-called JORA-files – validating and enhancing their contents and testing the ‘surveillance services’ that are offered and broadcast by Frontex from Warsaw. From the demonstration by the senior official in the Sea Borders Protection Directorate, it becomes clear that, as he puts it, ‘we are not yet capable of having a blinking dot on the screen, indicating a boat or situation, that we can follow live’. This is still ‘in the future’ (Interview, 26 September 2014). The bulk of the data visualized on Eurosur’s main interface is after-the-fact. Yet according to the Sea Borders Protection Directorate official, Eurosur today constitutes ‘a new idea about what we do, a totally new way of working. It is a beginning, a new concept of how to act’ (Interview, 26 September 2014).

Two ~ On the patrol boat a small screen, no larger than 10 by 10 centimeters, shows a grainy composition of black and white swaths pulsing, alternating and shifting. The commanding officer on board of the HCG patrol vessel – a Lambro 57 Pb Class III – had explained that the image is produced by a thermal camera and is calibrated so as to make cooler parts white while making warmer parts black. When set to the right parameters, aimed in the right direction and focused to the right width, this produces a greyscaled screen image recognizable to any accustomed television viewer as a flowing sea, in white, and a stationary, blackened coast rising in the distance. The captain had brought his patrol boat to a halt in order to scan a specific section of the Turkish coastline. As he explained, the operator of the thermal camera was scanning the shore and saw something. The operator zooms in at an extent of blackness hovering just below the massive black of the coast and against the white of the waves. A boat.

What is this boat, still firmly located in Turkish waters at the moment, about to do? Will it approach the line displayed on the captain’s console and tacitly known by a crew who spend their lives in these waters? And if it will, what can be done about it? One can find out, of course, by going closer and asking the boat’s occupants: ‘What are your plans for this evening?’ Not only would that be a breach of sovereign territory in the midst of carefully normalized diplomatic relations, but it would also further implicate the patrol crew in the fate of this boat. Its fate would become theirs and this, in a sense, is precisely what the ‘control of the border’ is aimed at preventing. The crew discusses, comments go back and forth as the boat and two figures standing on its rear are now clearly visible on the screen. The commanding officer makes a call on a mobile phone, contacting a command center at the HQ in Piraeus, talks for a while, hangs up, and speaks to the captain who readily engages the engines, directing the patrol vessel back to its original course. ‘It turned out to be a registered vessel.’ ⁴

Patrolling the perimeter of the Greek sea border with Turkey, somewhere halfway along the narrow strait between Chios’s eastern coast and the shores of a peninsula extending westward beyond the Turkish city of İzmir, involves a miscellany of tasks. Not only does the coast guard engage with smuggling, fishing regulations, the narcotics trade and environmental protection, the work also takes place in the midst of an ongoing territorial dispute between the two countries. In this narrow strait the coasts are so close to each other that the territorial waters of both parties are conjoined. The one ends where the other begins. In recent years, particularly after the construction of the ‘Evros fence’ on the land border in the north, these waters have become a major transit area for people who believe, not without reason, that they would be stopped and sent back if they sought to gain administrative entry to Greek soil. One of the most traveled routes by migrants from Syria and Afghanistan, as well as from African countries such as Eritrea and Somalia, runs through any one of the larger cities on the western edge of Turkey, such as Bodrum or İzmir, and crosses the Aegean to reach one of the Greek islands from there. In the late summer of 2014, with massive displacement taking place in and around war-torn Syria, coast guard patrols were almost completely preoccupied with boats carrying such people, particularly at night, seeking to repel or at least arrest them as part of EU-wide operations to ‘control the border’. Such preoccupation has only intensified since.

This was why the captain stopped and angled the patrol vessel as he did: by directing the thermal camera at the black/white division between a cooler coast and warmer waters it becomes possible to notice objects as they left the shore, creating a black protrusion into the white plane below. Not only was it clear that he and the camera operator had learned to see such protrusions more readily than a novice, but this ocular tactic was related to extensive experience with the so-called ‘modus operandi of illegal entry’. Although variable and constantly changing, tactics of clandestine travelers are often aimed at reaching the territorial waters of Greece and, ipso facto, the EU. Once across this line, drawn as part of diplomatic negotiations over the ‘Aegean dispute’, a boat and the people on it will have become Greek responsibility.

What these two excerpts of fieldwork and reflection offer up for questioning, particularly when placed next to each other, is how visual perception happens and, moreover, what else also happens when it does. Objects are not simply captured by the eye of the beholder; they are also made present by hands (intervention) and feet (mobilization). Vision and acting are intrinsically related. Additionally, the difference between direct and indirect ways of visualizing and acting diminishes once the efforts of collecting and standardizing data outside control rooms are taken into account. Conversely, direct encounters at sea between patrol boats and migrant vessels take place in the shadow of legislative hierarchies and chains of command that affect the current situation. ⁵ Abstract risk categories that professionals work with in control rooms and coordination centers have real life, if not real time, consequences whilst present situations are interlaced with visualizations, such as graphic displays, radar images and computer interfaces. While ‘in the future’ a more comprehensive vision is expected or hoped for, as the first excerpt makes clear, as of yet this does not exist (cf. Jeandesboz and Simon, 2014). It is also clear that even if it were to exist, it would involve a form of operative vision, a way of visualizing wholly imbricated in ways of acting.

The two excerpts of fieldwork presented so far illustrate the way interventions and mobilizations take place. Less attention, however, is given to the third notion, realization. What drew our attention during our fieldwork in the Aegean was how the very act of perception in the politics and governance of ‘irregular migration’ was dispersed and how this dispersal proffered the possibility of a politically productive recalcitrance. We recognized the importance of how people and their movements were never merely caught in the mosaic framing of surveillance, risk assessment and preemption, but instead could interfere in and agitate the chains of transaction through which perception was organized. A telling example of the coming into being of recalcitrant objects is present in the precarious work of patrol guards. With the help of this third excerpt we argue that the conceptualization of surveillance ought to engage more carefully with the double effect of objects in the dispersed forms of vision in border control, namely as both relatable as well as recalcitrant. In doing so, it becomes possible to identify the specific interplay between making things visible and invisible in current surveillance policies.

Three ~ As a variety of informants working with the HCG on the Aegean islands of Chios and Lesbos explained and demonstrated, immigrants have long since found ways – tactics – to capitalize on being seen. Far from a matter of hide and seek, immigrants and their ‘facilitators’ frequently turn the means of surveillance back against efforts to ‘control the border’. Being noticed opens up possibilities that can hardly be subdued precisely because perception requires action. So when a migrant boat is noticed and sufficiently close to a patrol boat, people on those boats are able to become subjects of an emergency of some kind, redrawing the lines of sight. A variety of informants doing patrol work expressed dismay, often vehemently, over how migrants were situated to trigger emergencies. They tended to emphasize the agency of migrants in the situational dynamics of encounters at sea: emergencies could be provoked by slashing a rubber boat with knives, jumping in the water, capsizing a vessel. In any event, no longer does the patrol boat survey ‘illegal migrants’ and ‘control a border’; it now witnesses an emergency and, as such, is legally obligated, according to the same agreements that recognize sovereign territoriality at sea, to initiate a SAR operation, drawing the migrants into EU-territory and Greek custody. Surveillant vision cannot help but agitate alternate registers of seeing and witnessing. Patrol guards become implicated in legal consequences and moral narrations through surveillant vision. As one informant, expressing frustration about inadequate orders coming from superiors and, ultimately, EU governance, told us:

Protect us. Make our job right. What do you want to do? To stop them. But they are in danger. I have to stop this boat, but if I stop this boat, it’s an inflatable boat that fits 15 people and there are 55 people in it, so if you say to me to stop it on the border, it’s 99% … these people, maybe they are going to drown, so I put their lives in danger, because they are already in danger, 55 people in that boat. If you are in the laws of the sea, these people are in danger. They don’t need to cut the boat to see that they are in danger. So I have to stop them, but then I am already… I’m illegal as well, because there is the law of the sea that says this boat is for 15 people, so what do I have to do when I see it? If I read the law, I have to stop this boat, control the sea, “eeehhhh [breathes in in frustration], you are 55 in it, you are going to drown!” So, you are illegal for this, first. But my orders are to stop this boat. (Interview, 9 September 2014)

It is for this reason that it is not the migrants who, in many cases, try to remain out of sight when crossing the sea border, but the patrol guards. The translation of international law, national jurisdiction and policy guidelines through a chain of command and control into ‘rules of engagement’ at the scale of the patrol boat itself allows for various responses (Klepp, 2011). If the border is a ‘method’ and an epistemological viewpoint, as Mezzadra and Neilson (2013) see it, it becomes clear that this viewpoint is mobile, intricately connected to action, and in no way to be reduced to the sole workings of a surveillance apparatus. Rules are deeply underdetermined, meaning that they do not dictate behavior but are ambiguous and are to be interpreted with regard to specific cases under specific circumstances. As our informants explained, patrol guards have quite limited means to actually prevent a boat from entering Greek territory, particularly when territorial waters are conjoined. Bright lights may be shone on a boat carrying migrants, warnings and threats may be uttered through a loudspeaker, shots may even be fired in the air or at a boat’s engine, and Turkish coast guards may be informed. Yet as soon as a boat enters territorial waters any attempt to push the boat back constitutes a breach of international law, residing under the principle of non-refoulement (see also Klepp, 2010). Any efforts to actually force migrants to not only stop but also recede is therefore almost by definition illegal. Attempts to ‘push back’ – by dangerously circling a boat, creating threatening if not life-threatening waves, by dragging or pushing boats back into international or Turkish waters, by taking migrants back to Turkish shores, by handing them over some third party – are nonetheless frequently if not regularly made by patrol crews all over the Aegean, at great risk to migrants and to themselves (International Federation for Human Rights (FIDH) et al., 2014; Amnesty International, 2014b). Yet in these encounters it is incumbent that such actions not be seen, registered, noted or reported (cf. Cabot and Lenz, 2012). A crucial part of the EU–Turkey deal has been the regularization of ‘push back’ through eventual deportation to Turkey of those intercepted, indicating how its high-stakes diplomacy is inflected by a tense history of surveillance and rescue in the Aegean.

The Farmakonisi case of 20 January 2014 is an important example: a widely reported ‘push back’ case in which 11 refugees, of whom 8 were children, lost their lives when their boat capsized due to it being towed through volatile waters, with many people caught below deck (Amnesty International, 2014a). When lawyers for the group requested information about the coordinates at which the patrol boat had been registered at different times during the incident, no such information was said to exist. Yet it is clear that such information is registered through coastal positioning systems and radar technology. All that was initially handed over by the HCG were handwritten logs by the patrol crew, supposedly read off from the on-board navigation system. Video cameras, Siemens products provided by Frontex – with which many patrol vessels are equipped to aid ground personnel in monitoring situations and to provide evidence at a later date – were said to have been switched off, as was the prerogative of the commanding officer. We are confronted here with specific interventions of ‘de-monitoring’ and ‘re-mapping’ that play a role in attempts by Greek authorities to ‘re-value’ the Farmakonisi case and others like it into unfortunate accidents. In response to such suppression of data, activists and legal aids to migrants urge people at sea to use mobile phones and other devices to track their whereabouts during crossings as proof of where they were at which times. ⁶

Recalcitrant objects and transactional politics

The question that remains to be answered is which notion of politics is at stake here. The case studies demonstrate how this notion of politics ought to be sensitive to the intimate relations between vision and action and the enmeshed character of surveillance and its objects of vision, and the recalcitrance of those objects, giving rise to actual political events. In order to arrive at a notion of politics that meets these conditions, this section will engage with the issues related to ontology and visualization brought forth by the fieldwork.

The excerpts presented earlier emphasize the role of non-human agency in border control practices (Frowd, 2014; Walters, 2014) and the transformative role of material objects in surveillance practices (Aradau, 2010; Dijstelbloem and Broeders, 2014; Andersson, 2015). They also demonstrate that differentiating between surveillance and SAR is connected to divisions of labor between different organizations and different kinds of professionals. Unlike the Italian government when launching the Mare Nostrum operation, official Greek policy maintains a sharp division between surveillance and SAR, which leads to a division of labor between, on the one hand, government agencies and officials working ‘from the screen’, and the agencies and officials who physically encounter migrants when protecting the border on land or at sea on the other. The result is a functional differentiation between ‘white’ collar and ‘blue’ and ‘green’ collar border control authorities.

In addition, the fieldwork confirms many of the conclusions that can be drawn from the analysis made by Heller et al. (2012) of the so-called ‘left-to-die-boat’, a tragic case in which 63 migrants lost their lives while drifting for 14 days within the NATO maritime surveillance area. By reconstructing the chain of events, their study aimed to visualize the information infrastructure that was present in the Mediterranean at that time in order to reallocate agency and accountability to various actors that had crucial information at their disposal. Whereas state agents in all their manifestations and associations connect and relate various sorts of information from highly different technologies so as to visualize risks and to arrive at interventions, Heller et al.’s (2012) reconstruction gathers these scattered images together by reuniting them in different ways so as to reattribute institutional responsibility.

To further elaborate on the relationship between strategies of visualization and ontological issues it is helpful to emphasize the notion of ‘recalcitrant objects’ introduced earlier. Sørensen et al. (2001), building on Latour (2000) and Stengers (1997), describe recalcitrance as the lack of susceptibility to control or authority:

Recalcitrance is an attribute of things being studied and refers to the extent to which they are uninterested in the ‘questions being asked of them’. Recalcitrant objects provide answers on their own terms, rather than those of the authorities studying them; they can object. (Sørensen et al., 2001: 301)

Following on from this, it seems appropriate to grant objects – living bodies, moving boats – a much stronger role (cf. Walters, 2015). The twofold realization of objects as both relatable entities and resistant particulars applies conceptually to the events that take place in actual encounters between coast guard patrols and migrants’ boats. In order to take recalcitrance seriously, it is important to move beyond the seamlessness implied in many accounts of technological border control. For instance, the messiness of encounters in the Aegean necessitates a conception of Eurosur that problematizes the sovereign grasp and vision of such systems. To do so, encounters such as the ones discussed previously will be conceptualized in terms of transactions.

In Knowing and the Known, Dewey and Bentley (1949) intended to provide a concept of transaction that, amongst other things, did not stick to subject/object distinctions and did not separate observation and the observer from the observed. The crucial distinction they make is among ‘self-action’, which involves things acting out of intrinsic powers, ‘inter-action’, in which causal connections mediate the relation between things, and ‘trans-action’, which they describe as follows:

Trans-action: where systems of description and naming are employed to deal with aspects and phases of action, without final attribution to ‘elements’ or other presumptively detachable or independent ‘entities’, ‘essences’, or ‘realities’, and without isolation of presumptively detachable ‘relations’ from such detachable ‘elements’. (Dewey and Bentley, 1949: 108)

There is, in transactions, ‘no radical separation … between that which is observed and the observer’ (Dewey and Bentley, 1949: 103). The concept of transaction is likewise intended not to separate a thing from its environment, but to consider the entire constellation of thing-and-environment as the relevant locus of action. When what happens in the Aegean is considered a transactional encounter of a dispersed surveillance apparatus with bodies on the move, the inability of either party to the transaction to ‘control’ the entire situation becomes more readily understandable. The ‘situational awareness’ that ought to provide modes of intervention in a meeting of Hellenic Coast Guard boats with boats carrying migrants is confronted with recalcitrant objects which can prove hard to tow away, as in the Farmakonisi case. Or they can prove recalcitrant to the extent that they enact a change of the definition of the surveillance situation into that of an SAR mission. Although the implementation of new technologies often promises to subdue such recalcitrance, this analysis showed how recalcitrance is generated by the interplay of visual technology and the being seen (by migrants) of mobile forms of surveillance technology. Moving between border protection and SAR is a way of dealing with the very resistance that objects perform when being rendered relatable.

A focus on transactions extends the perspective of the process in both space and in time (Amoore, 2014; De Goede et al., 2014; Jeandesboz and Simon, 2014). Well before visual perception takes place on Eurosur monitors, transactions are ongoing and attention is becoming distributed. Well before a camera operator moves a joystick so as to scan the divide between black coast/white water, a ‘facilitator’ has decided and arranged to take ‘clients’ across at night and not in the light of day. Well before the official at the NCC has clicked her mouse button, sovereign states have signed agreements over AIS usage at The International Convention for the Safety of Life at Sea. The point is not that perceptions happen insofar as they take into account the preceding transactions and, thus, the always problematic decisions that have already been taken. Instead, above all, perceptions happen insofar as a cascade of transactions is transacted further still by going along with and going against the decisions that are at stake in them.

Transactions carry with them a multitude of decisions taken that, as perception happens, will or will not hold (Latour, 2000; Amoore, 2013). It remains to be seen whether and how AIS signals, displayed on a grid of longitudes and latitudes, come to contribute to the ‘situational awareness’ that the Eurosur system is poised to bring forth. It remains to be seen whether and how the differentials of temperature, allowing a crew to notice a boat, come to enact the ‘control of the border’ that the HCG has been tasked with performing. It therefore seems pertinent not to assume that the entire political topology of the Aegean in particular, or the Mediterranean in general, is organized by the surveillance apparatus. The topology of surveillance-and-migration-at-sea would then appear as the mere enactment of surveillance unfolding. To take a transactional view is to focus on the enmeshed unfolding of heterogeneous but constantly mutually attuned transactive movements. These are movements of both coast guard boats belonging to a larger and more widely distributed surveillance apparatus, as well as movements of migrants. In the transaction, these movements appear as the coordinated action of two types of boats, completely asymmetrically placed in the political diagram of the Aegean yet both finding expression in the topology of transactions. To observe ethnographically is also to remain at a further remove from the overly sovereign, technological rhetoric of visualizing systems, and to not accord a privileged status to the type of vision over which the surveillance apparatus claims ownership, or, better yet, to not locate ‘vision’ exclusively within the surveillance apparatus, as its vision is both distributed and enmeshed in action. Moreover, the action such vision is enmeshed in is a transaction in which objects of vision intermingle with practices of visualization. Developments since September 2014 have demonstrated how surveillance practices entail consequences that in turn become recalcitrant to the securitization of the Aegean, prompting European authorities to effectively forgo non-refoulement altogether (Heller and Pezzani, 2016). As we argue, this is already the case in the scale of encounters at sea, as there can be no vision without transaction.

At this point it becomes possible to understand the recalcitrance of the ‘objects of surveillance’ in a political way – and, keeping in mind the efforts of coast guard personnel to keep out of sight, also of those enacting surveillance. Because surveillant vision is an operative vision it does not exist separate from that recalcitrance, and therefore it is at times subject to political moments during which recalcitrance structures transactions. Such transactional politics exists for instance in the transitions from ‘border control’ to ‘SAR’. The visualization that is involved with these configurations always works on specific paths of observation and is embedded in material circumstances. Moreover, an analysis of dramatic events such as the Farmakonisi case ought to grant the actors themselves the task of enacting transactional politics. Recalcitrant objects aim, to a certain extent, to make themselves part of the reconstruction of a critical event by participating in the completion of the picture of the disastrous operation. The analysis presented here makes clear that these operations can be understood in an ontological way even more consistently by explicitly stepping away from approaches that situate ‘vision’ in the minds of actors or regard it as embodied in technologies. Instead, vision and action are meticulously entangled and call for both ontological and political theorizations as a result thereof.

Conclusions

Surveillance at sea aims to relate vision to action, arriving at ‘situational awareness’ and associated interventions. The connections between various techniques aimed at visualization through ‘interoperability’ have resulted in a patchwork of distributed surveillance practices. Far from providing a seamless web, the monitoring of the Aegean manifests a seascape full of anxiety and risk. Instead of leading to border ‘control’, border surveillance opens up a particular kind of politics we have termed ‘transactional’. This was already the case when the fieldwork for this article was conducted in 2014 and has only increased since then.

The relationship between vision and action in surveillance practices at sea has proven to be harmonious with three characterizations of how connections are made and new phenomena are created in laboratory research, namely by way of interventions, mobilizations and realizations. Surveillance of the Aegean tends to display a logic that engages with the visible and the invisible and is driven by a continuous repair of its own consequences. Visualization leads to the realization of subject-objects that allow for resistance as they have at their disposal the possibility of turning against surveillance policies and thereby subverting the very practices of visualization that enacted their subjectification.

The thesis we have brought forward holds that the conceptualization of surveillance at sea ought to engage more carefully with the notion of recalcitrance and the twofold nature of objects, namely as both relatable as well as resistant. In doing so, it becomes possible to specify the interplay between vision and action. Drawing upon the fieldwork excerpts, a conception of transactional politics seems capable of keeping open the possibility of action. As such, the notion breaks with both the dichotomy between border surveillance and SAR operations as well as with a sharp distinction between the subjects and objects of visual surveillance. Conceptually these dichotomies are insufficiently equipped to take the coming-into-being of political entities, political interactions and political moments into account. Presupposing subjects, objects and points of view that remain outside the specter of politics, either as a constitutive element or as a peripheral entity, a priori excludes all three from the processes of intervention and mobilization that allow for moments of realization and, perhaps most politically pertinent, of recalcitrance.