Policing the internal Schengen borders – managing the double bind between free movement and migration control

ABSTRACT Analyses of security practices at borders have focused on the dematerialisation and de-territorialisation of control of individuals’ mobility. This paper explores the nature of the control the state still exercises over individuals’ mobility at national borders. It focuses on a border that is supposed to have been opened, between France and Italy, inside the Schengen Area. It is based on analysis of the practices, representations and organisation of French border police officers, beginning with the legal and organisational transformations due to implementation of the Schengen Convention at the France–Italy border. It then turns to the study of border police officers’ targeting practices, using Heyman's notion of a ‘plausible story’. Finally, it assesses the influence of deportation practices on the territoriality of the control of individuals’ mobility, as well as its effects on targeting practices. These borders are at the core of the interaction between the construction of a new, European political centre, and the affirmation of an older one, the national, political centre. This paper demonstrates that border police officers are in charge of dealing with the tension, a double bind of sorts, emerging from this interaction. National, internal borders are still a site in which the state manages individuals’ mobility.


Introduction
Individual's mobility control has long been considered as one of the main form of the exercise of sovereignty. In recent years, scholars have explored the link between the authority of the state and the power it exercises over its territoryspecifically, over the latter's limitsand whether state authority over the edges of its territory has been increasing or decreasing. Some scholars question the relevance of national borders (Sassen 1996, Mezzadra andNeilson 2013) and of the state's exclusive authority to define a jurisdiction and its limits (Sassen 2013). For example, for some authors, the use of geographical limits of national territory as a 'stage' (Andreas 2001), or the tightening of border controls in general, is a clear sign of loss of state sovereignty (Bosworth 2008, Brown 2010. On the other hand, analyses of security practices at borders have focused on the extension or multiplication of areas in which checks and surveillance are conducted, and thus exercise of State power, beyond and inside the limits of national territory. Because of the dematerialisation (Amoore 2006, Salter 2006) and de-territorialisation (Balibar 2003, Walters 2006, McNevin 2014 of control of individuals' mobility, we would be witnessing a gradual denial of national borders' role as gatekeepers of national territory and as the 'ultimate pillar of statehood' (Chalfin 2007(Chalfin , p. 1625. This reflexion around the link between sovereignty and control of individual's mobility is particularly striking in the case of the internal borders of the Schengen area. The Schengen Agreement signed in 1985 allows nationals of member states to move freely throughout the Schengen Area. The Schengen Convention, adopted in 1990, extends this right to citizens of other states who have a residence permit in one of the Schengen member states. The implementation of this convention led to the dismantling of ports of entry and the end of systematic border checks within the Area. We could conclude that these internal borders have lost their mobility regulation function and consequently the member states a part of their sovereign power. If the external borders of the European territory have been widely studied, with a specific focus on the analysis of the kind of sovereign power exercised (national or European) only a few studies were devoted to the internal borders (Groenendijk 2002).
This article analyses the tension between freedom of movement on European territory and national immigration control policies exercised at geographical borders. Indeed, since the mid-1970s decision to end labour immigration, national borders have been defined as the legitimate locus of control of individuals' mobility. This article explores the nature of the control the state still exercises over individuals' mobility at national borders: that between France and Italy, inside the Schengen Area.
The materiality and territorially embedded nature of the control of individuals' mobility at these edges of the French territory will be investigated. What kind of consequences has the lifting of internal borders inside the Schengen Area had on this exercise of state power at the edges of national territory? This study is based on the analysis of practices, organisation and representations among police officers at this specific kind of national geographical borders.
Based both on archives from the French border police  and on interviews and observation of the daily tasks of the Police Aux Frontières (PAF), the French border police, it will pay specific attention to the organisational evolutions of this police department throughout the years and to the condition of exercise of their activity. Focusing on the controls operated at the France-Italy border, I will examine how tensions are expressed in the border police' work. The implementation of the principle of free movement of people through national borders between two member states of the Schengen Area is a perfect example of a double bind for border police officers.
In Section 2, I will analyse legal and organisational changes due to implementation of the Schengen Convention at the France-Italy border. I will then turn to the study of border police officers' targeting practices, using Heyman's notion of a 'plausible story'. Finally, I will assess the influence of deportation practices on the territoriality of control of individuals' mobility, and their effects on targeting practices.

Theoretical framework
Studies of security practices at borders have highlighted the increasingly 'mobile and fluid character of borders' (Loftus 2013, p. 3). The general idea is that border control and border surveillance actuate 'beyond physical borderlines' (Loftus 2013, p. 3). In this literature, two elements are usually pointed out: the decreasing materiality and territoriality of borders.
The notion of territoriality is used to describe the embedded nature of borders within geographical territory. It is often mistaken for the materiality of border controls. Many factors contributing to the de-territorialisation of borders have been identified. The lifting of borders inside the Schengen Area is widely recognised as a major change in border definition. Foucher (1998, pp. 237-238) sees the creation of the Schengen Area as the end of territorial borders. Analysing European visa policies, Guild (2002, p. 103) shows that building the Schengen Area led to the 'de-linking [of borders] from territory'; borders apply to individuals 'not on the basis of their physical position but on the basis of their nationality and individual characteristics'. The general idea is that any location can become a locus of identity checks. Studying the effects of the implementation of the Schengen Convention, Groenendijk (2002) shows that national government actors took advantage of the adoption of compensatory measures included in the Schengen Convention, to strengthen control over citizens and non-citizens. Groenendjik emphasises the transfer of controls away from geographical borders towards national territory. His study is largely based on the study of legal transformations due to the Schengen Convention. Other scholars have focused their analysis on the shifting of checks and examinations inside national territory, away from national boundaries. The theme of remote controls has been widely studied (Guiraudon 2002a, Boswell 2003, Lavenex 2006, as well as the various technologies used along external EU borders, before migrants reach EU territory. Visa policies are part of this externalisation of border control (Guild 2002, Infantino 2010, as are carrier sanctions (Guiraudon 2002b).
A lot of the observation and analysis produced by this literature describe the recent evolution of border control and surveillance very accurately. However, these studies tend to give an 'ahistorical' (Heyman 2004) image of an immutable, material and linear 'wall-border' corresponding to the geographical, territorial limits of the nation-state. Yet control of individuals' mobility is an entire part of state activity and is exercised in numerous ways and venues. In order to avoid 'an approach framed solely in terms of securitization theory, which often implies presentism by a lack of attention to the space/time structuration that rendered possible an event' (Bigo 2014, p. 211). I will study the border police officers' organisation and practices at the France-Italy border before and after the Schengen Agreement's implementation. Based on the analysis of archive documents produced by the French border police administration, this article will assess the evolution of the materiality of controls at the France-Italy border. Following in that respect an important literature (Torpey 2000, Noiriel 2001, 2007, Walters 2006, Weber and Bowling 2008, I interrogate the forms of continuity and discontinuity concerning the way the state regulates individual's mobility. Furthermore, very few studies have analysed actual, practical controls that are still being conducted at internal borders (Darley 2008, Schwell 2010. I am interested here in the materiality of borders, that is, in the concrete practices and organisation of the services in charge of controlling individuals' mobility at the limits of national territory. The literature on policing borders has focused on the 'dual role' (Heyman 2004, p. 304) of policing actors, who have contradictory objectives: to facilitate the crossing of borders by those described by Aas (2011) as 'bona fide travellers', and at the same time, to target and stop 'crimmigrants'. This tension has been studied through analysis of the targeting practices, determination of the origin of discriminatory practices amongst police officers and identification of the ways in which these criteria are defined and diffused among the profession (Pratt and Thompson 2008, Chan 2011, Quinton 2011. This literature emphasises the 'race' criterion as playing a central role in estimation of the risk associated with an individual (Heyman 2001, 2004, Pratt and Thompson 2008. But there is a consensus in the literature on the fact that even though race is widely used, it does not suffice to determine which individuals to target (Heyman 2001, 2004, Chalfin 2007, Darley 2008Pratt and Thompson 2008). The notion of a 'plausible story', defined by Heyman, is particularly useful in the context of internal borders. It allows us to grasp a multiplicity of criteria and to give an account of the narrative dimension of the representations used by officers checking people crossing the border. This notion is a good way to describe police officers' ability to elaborate and transmit this kind of professional and field-grounded representations.

Methodological note
Scholars have argued in recent years (Côté-Boucher 2010, Loftus 2013, Côté-Boucher et al. 2014) for a 'practice turn' in border studies, 'that emphasizes how actors act and how they give meaning to their actions' (Côté-Boucher et al. 2014, p. 197). Besides, in the study of administrative actors there is a long-standing tradition of giving particular attention to street-level civil servant practices and representations, in order to comprehend the discretion these actors enjoy (Lipsky 2010) when implementing public policies. From the same perspective, public policy studies have shown the significance of analysing the implementation phase in order to better acknowledge the reality of state regulations (Pressman and Wildavsky 1984). In contrast to the approach of these scholars of policy implementation, instead of looking at the gaps between a rule and its enforcement, this paper will analyse the ways in which these civil servants deal with conflicting orders. Particular attention will be devoted to the techniques and formalised representation used by border police officers in order to implement the rule of free movement of people within European territory in a context of restrictive national immigration policies.
Analysis of the targeting practices 'require[s] direct attention to the work routine and organizational culture of port officers' (Heyman 2004, p. 306). The methodological approach of this paper can be defined as 'ethnographic sensibility as an epistemological commitment' (Côté-Boucher et al. 2014, p. 204). It relies on the gathering of information about both the materiality and the symbolic aspect of the controls, thanks to interviews and observation of border police officers' concrete tasks. The empirical elements of this paper are part of my Ph.D. fieldwork. The data I have used here have been collected through interviews with street-level and chief PAF officers from the local offices of Nice and Menton (20 interviews), near the France-Italy border, and analysis of Interior Ministry archives. However, this analysis is also informed by the intermediate nature of my fieldwork in general, between interviews and observation, since I gained a particular form of access to the field. Due to my former political commitments, I did not approach PAF at the central level, so as to avoid denial of access. Thanks to academic connections, I reached a superintendent-director of the PAF for the Alpes-Maritimes department. 1 After an initial interview in October 2008, he gave me access to all the units under his authority: at the airport, along the French-Italian border at Menton, and inside the services specialised in criminal networks linked to migration. The fieldwork took the form of two one-week stays, in November 2008 and in March 2009. The interviews were scheduled directly with the superintendent, and I was accompanied from one unit to another by police officers: thus, other police officers perceived me as a 'trainee'. This ambiguity of my status proved helpful, allowing me to observe certain practices and hold informal conversations with various police officers for several hours. The limited total duration of my investigations was somewhat offset by these extensive periods of observation.
The nature of the interviews and the fact that they were not recorded, accentuated this ambiguity. My main questions were focused on the PAF's daily tasks and organisation of its work, and on the representations associated with them. I also asked police officers who had already been working before the lifting of internal borders within the Schengen Area, how they experienced this transformation. However, I did not systematically observe the practices of police officers, so the data I gathered are more centred on actors' discourses about their daily work. Moreover the material condition of the framework did not allow me to conduct an in-depth ethnographical investigation.
The specificity of this study is to cross empirical data on practices and historical data which relates to the legal and organisational framework of the PAF's intervention. I consulted a range of archive documents dated from 1945 to 1992, from two administrations: the PAF itself and the Direction des Libertés Publiques et des Affaires Juridiques (DLPAJ), a branch of the Interior Ministry in charge of the development and enforcement of regulations concerning entry and residence of foreigners in France. This enabled me to develop a long-term analysis of the border police, and to obtain information about the relationship between the police and the authorities in charge of elaborating public policies. Finally, this look at the contemporary period allows me to emphasise the permanence of certain practices, ten years after the actual opening of borders in this region. 2 2. Evolution of the legal and organisational framework of PAF intervention at the France-Italy border

The gradual institutionalisation of the French border police
The police department in charge of border checks at ports of entry emerged within the French police's intelligence department (Direction des Renseignements Généraux). 3 This origin had an impact on staff training, the locations they were posted in, and the tasks and duties of this police service. One of the main effects of the border police' development within the context of the intelligence service was the lack of material and legal means to ensure linear border control along national land borders. The PAF's institutionalisation and gradual autonomisation went hand-in-hand with an increase of its material resources for border surveillance, not only at ports of entry, but also in a zone extending inward into national territory, along secondary roads.
Until 1973, when the border police became independent from intelligence services, the only distinctive element setting border police officers apart from regular intelligence service officers was their posting at the limits of national territory. Border police officers were part of intelligence service staff and worked in intelligence service facilities situated along the land borders and in airports (from 1953). From the 1960s on, the PAF sought to extend its reach beyond ports of entry, both along secondary roads and further into national territory, in a limited zone close to the border. To fulfil that objective, officers from the Compagnie Républicaine de Sécurité (riot police)a much more operational police forcewere put under PAF supervision during the entire decade. In 1969, following the 1968 events, 4 these officers were permanently transferred to the PAF. This represented a turning point in the PAF's professional identity. Until then, as members of the intelligence service, they had principally been entrusted with administrative tasks. A year after the PAF became autonomous; a new kind of brigade was created: the 'Brigades frontalières mobiles' or BFM (Mobile Border Brigades). These brigades were to patrol the French land border between ports of entry. The zone covered by these brigades extended 60 km inward from the border itself. 5 This kind of brigade was also created in the Netherlands in 1976; in Germany, before the Schengen Agreement, the Bundesgrenzschutz carried out this type of border surveillance in a zone extending 30 km into national territory (Groenendijk 2002). The 1970s are usually presented as a turning point for migration policies. Indeed, following the 1973 oil shock and resulting economic difficulties, most Western European countries closed their doors to labour immigration (Laurens 2009). Previously, migration controls were mainly carried out inside the territory, when migrants accessed the labour market. In other words, before the 1970s, the control of people's mobility took place within French territory, via control of labour market access more than through checks at the border. Hence, the PAF's autonomy and the creation of these special brigades along national borders coincided with the tightening of migration policies and migration control becoming more territorially embedded.
However, the legal grounds for the intervention of these mobile border brigades were unclear. First, since their creation, there had been a gap between the legal justification and the objectives of the checks they carried out. In the case of border checks, entering the country constituted sufficient grounds to carry out identity checks on individuals in order to verify the legality of their administrative status. However, the BFM were actually conducting identity checks in reference to an 'attempt to commit a crime or offense', 6 whereas they were supposed to be border checks. This lack of substantial juridical basis can be understood as a form of legal lawlessness (Brodeur 2010)which results in the gap between the formal legal framework and policy objectives. The second aspect of this ambiguous legal situation was linked to the French policing system, in which two security agencies function in parallel: the National Police and the Gendarmerie. The Gendarmerie's jurisdiction is limited to rural areas, and the National Police's to urban areas. The 'green' border 7 is located in the Gendarmerie's jurisdiction. The PAF's jurisdiction is formally limited to border ports of entry. Yet the Brigades Frontalières Mobiles were created and their staffs have increased over the years, without any legal clarification of their intervention. Over the years the material and institutional resources of border police have been reinforced. This progressive change granted the possibility to exercise a more systematic control of individual's mobility at the edge of the territory.

Implementation of the Schengen convention at the national level and creation of the Schengen area
This lack of legal ground for action has regularly been pointed out by the PAF hierarchy over the years. But paradoxically, it has been resolved thanks to the implementation of the Schengen Convention and the lifting of internal borders.
At the European level, the building of an area of freedom of movement has been accompanied by 'compensatory measures' defined in the Schengen Convention. These measures (mainly pertaining to police and judicial cooperation) were meant to counterbalance the 'security deficit' that the lifting of internal borders was expected to produce (Bigo 1996). The same kind of rationale was used by political actors at the national level. Following the adoption of the Schengen Convention in 1990, French Interior Minister Charles Pasqua proposed a new measure to 'compensate' for the lifting of internal borders inside the Schengen Area: the creation of a 'Schengen zone' 8 in which identity checks would be facilitated. A line was drawn inside French territory, 20 km away from the border; inside this zone, police officers would be allowed to carry out checks without any justification. The legislative amendment essentially resulted in adding an exception to the Code of Criminal Procedure, which otherwise specifies precise cases in which identity checks may be conducted. The checks carried out inside this zone, on the other hand, are exceptional and legitimised by the border's proximity. 9 The legitimate locus of control of individuals' mobility has been extended from a 'border line' to a 'border zone'. But more importantly, since the 'Schengen zone' law gives border police the legal basis necessary to carry out border controls outside border checkpoints, this measure legalised a practice that had been in use since the mid-1970s. Paradoxically, a measure meant to implement the principle of free movement of people within European territory actually reinforces the police's legal capacity to control internal borders. 10 Since the enforcement of the Schengen Convention at internal borders, PAF officers have no longer been allowed to check individuals' identities in ports of entry. 11 Indeed, most checkpoint structures there have been dismantled. All border identity checks are now done by members of the Brigades frontalières mobiles. BFM staff expanded very significantly in the mid-1980s, and the end of systematic border checks in ports of entry did not represent a total disruption. Indeed, in 1991, 44% of PAF staff posted to internal borders were members of one of these brigades. There is a form of continuity in the controls being held. The same brigades are holding identity checks in the same area, before and after the lifting of internal borders inside the Schengen Area. The definition of national borders as the legitimate locus of control of individuals' mobility has not been challenged, even though systematic checks are not allowed anymore.
Since the 1960s the PAFor the police department under its supervision-has been performing border checks inside national territory, in a zone whose limits were first justified by the proximity of the border. Since the implementation of the Schengen Convention, identity checks have been facilitated in this Schengen zone, and not the entire French territory. 12 Control of people's movement is still entrenched at the edges of the territory. Furthermore, the creation of the Schengen Area appears to be a way to deal with the double bind of open borders at the European level and migration control at the national level. Border checks have not been eliminated, but they have become invisible to the majority, producing an apparent framework of free movement of individuals. The ban on systematic border checks has led to the holding of supposedly 'random' checks by mobile brigades. In reality these checks are not 'random' at all; on the contrary, they are 'targeted'. Consequently, in the Schengen Area the 'hierarchizing and stratifying capacity of borders' (Mezzadra and Neilson 2013, p. 7) only relies on border police officers' work. It is a legal framework which transfers total responsibility to these actors for dealing with the double bind of free movement and immigration control at national borders. They manage this through targeting practices.

Contemporary border control practices at the French-Italian border
Targeting is a central dimension of border policing. Identity checks at the border imply the targeting of individuals and a very quick judgement from border police officers. As Heyman noted, during the first inspection, officers have a very short amount of time to build 'an interpretative construct about the entrant, her rights, her relations with other people, and her past and future lines of action ' (2004, p. 310). An officer working near the Italian border declared 'when we are checking the train it only stops for two minutes in the station. We have to target [people]. If the train is late, the company loses money'. 13 This study will characterise targeting practices using the notion of 'plausible stories'. Four components of them have been identified. The first involves immediately perceptible characteristics of the individuals crossing the border, such as assumed ethnicity, nationality, gender and/or age, clothes or luggage; the second, a modus operandi, a specific mode of transportation and way of crossing the border irregularly, such as hiring a smuggler or carrying forged documents. The third is the person's itinerary: a point of departure, one or several stop-overs and a destination. The final component is the police's legal and material ability to deport the migrants once identified. Each of the plausible stories used by PAF officers are characterised by these four dimensions. But only one of the four elements can suffice to associate an individual to a particular plausible story. Plausible stories are central to border policingboth the targeting and the decision of what kind of actions PAF officers are to take.
Plausible stories are elaborated by PAF officers for both their clientele, that is, the people they suspect to cross the border irregularly but also for people associated with 'bona fide travellers' (Aas 2011). Each individual crossing the border is associated with a plausible story. Each border can be characterised by a specific geographic and migratory context. This context is defined both by plausible stories associated with 'bona fide travellers' and with the PAF's clientele. These plausible stories are flexible and can be modified. PAF officers target individual on the basis of already known plausible stories but also when they detect anomalies with the plausible stories of 'bona fide travellers'. Once an anomaly is detected; it can lead to the elaboration of a new plausible story concerning the PAF's clientele. 14 This notion takes into account the ability of border guards to modify and to elaborate new stories.
In the Schengen Area, specific rules apply to identity checks. As checks are supposed to be random, targeting becomes much more obvious. Police officers then have to deal with how 'bona fide travellers' regard their targeting practices. T. from the Menton PAF noted: 'In the train, some people react. Once a lady asked me 'and me why don't you check me?' 15 The officers' justification for these practices is linked to the necessary rapidity of border checks aboard the train. This practice is one of the officers' responses to the 'double bind' which characterises their work at the border.
At the time of the fieldwork, the PAF officers at the France-Italy internal border seemed to be highly preoccupied by a specific kind of travellers: young men, sometimes teenagers, coming from Afghanistan, Iraq or Iran and heading to the UK, often carrying no luggage at all. The plausible story elaborated to describe these travellers gives some account about the way PAF officers manage to deal with the double bind specific to the internal borders of the Schengen Area. In October 2008, T., a police officer at the Nice PAF headquarters described the characteristics of this plausible story: They leave their country to flee the war, their journey goes through Turkey, Greece, and they take the ferryboat and arrive in Italy, where they take a train to Ventimiglia [the Italian side of the border]. They do not ask for asylum, they want to go to Calais [the French coastline which is closest to the UK]. 16 The plausible story also contains specific information about the modus operandi for crossing the French border: The most practical way to enter French territory is by train. Their technique is: they arrive at Ventimiglia; they buy a ticket to Cannes, in Cannes they change to Paris. They don't stop in Nice because they know the PAF is more present there. 17 PAF officers also consider the smugglers' practices: We know that they ride the train from Rome to Genoa and then from Genoa to Ventimiglia. We know that in Rome there are smugglers because the first tickets that they buy are paid for by credit card. But for the smuggler, there is no point in getting involved, in going all the way to Nice, it isn't useful for the smugglers, they [the migrants] are on the right track! 18 This plausible story takes European territory into account as a reference point, as the relevant territory for mobility. At the Menton PAF office, G. stated: There are some 'brakes', in Turkey for example, when they arrive there, they can cross the border the same day or they might have to wait for three months, it is the same in Greece, then when they arrive in Italy, they go to those [detention] centres. Once they get there, [the authorities] release them little by little. We have no clue how many will arrive [at our border]. They arrive here with deportation orders. In the Italian detention centres they can be held for two weeks or three months. 19 This characteristic is specific to European internal borders and implies judgement of their European counterparts' work at the external border. Very often, Greece is defined as a scapegoat. 'The Turkey-Greece border is a sieve, and once they are in Greece, they are in the Schengen Area.' 20 The recurrence of this statement has revealed a shared view among PAF officers about the work carried out by their Greek colleagues.
The continuation of the migrants' journey is also known to these police officers: In Paris, there are a lot of Afghans in the park and in the Calais region also, it's going to explode! There are like 4000 to 5000 people arriving per day, the NGOs can only handle shelter for 1100 persons a day, it's a fight, these persons are hungry, and they are exhausted. 21 As this statement shows, the plausible story can even predict the future actions of the migrants checked at the France-Italy border. Plausible stories are elaborated through various interviews of people checked at the border, but also through the 'traces' border police officers find: train tickets, documents delivered by the authorities in the states migrants have passed through. The officers' representations have revealed a very rich and detailed understanding of the journey, motivations and objectives of irregular migrants. However, in practice, one single element corresponding to the plausible story elaborated by the PAF officers is sufficient for them to associate an individual to a group of targeted individuals. The last element of this plausible story involves border policing: the deportation capacity of the PAF officers. These migrants are specific in the sense that they cannot be deported. Individuals may only be deported to their country of origin or to the country they have passed through directly before they are stopped. PAF officers anticipate their ability to deport individuals, when targeting those they want to check, and sort them based on the possibility of deporting them. A PAF officer at the Nice train station stated: We are sorting the persons we check according to the criteria defined by the procedure towards the detention centre: we are filtering the persons carrying their identity documents, the ones carrying forged documents, the adults carrying train tickets. 22 If they find Italian train tickets on a person, they can deport him/her to Italy via the readmission 23 procedure. If the PAF can prove the individual is carrying a forged document, a criminal investigation is opened, but they must determine his/her nationality. But once this is done, the PAF's ability to deport the person depends on the procurement of a laissez-passer from the authorities of his/her country of origin. Its goodwill in delivering these passes depends on many variables (El Qadim 2014). One of the police chief's key tasks is to maintain good working relationships with the local official representatives of these countries of origin. 24 Therefore, in assigning a plausible story to an individual they anticipate their ability to obtain a laissez-passer, as the same Nice train station officer put it, 'Tunisians are easier to deport, we put them in the detention centre but they don't stay long.' 25 Furthermore, the deportation procedure is subject to a time constraint. Detention cannot be extended beyond 45 days. So police officers have to identify the person's country of origin and obtain a laissez-passer from it within this period. They prefer to avoid putting an individual in a detention centre if they already know it will not be possible to deport him/her. T. from the Nice PAF asserted that for Afghans, 'there are no official representatives'. So individuals recognised as Afghans by the PAF are not put in detention centres. 26 The criterion used to decide whether the person will be placed in a detention centre or '"readmitted" [deported to Italy] is the possibility of deporting them to their home country '. 27 In this border region, the readmission procedure is used for a population the PAF cannot deport back to their home countries. 28 The rules constraining border police officers' decisions when they wish to deport an individual lead them to maximise the efficiency of checks and arrests. They define criteria to make sure that they will be able to deport the individuals they arrest to their country of origin or to the last country they passed through before France (in this case, Italy).
This plausible story is one among the various plausible stories used by the PAF officers at the French-Italian border. But this specific plausible story particularly illustrates the ambiguity of the Schengen internal borders. The control of these migrants is a clear example of the fact that PAF officers have to face European and French stakes at the same time. It results from the interaction between these migrants crossing European territory and the PAF officers enacting national sovereignty at the geographical borders. The rather exceptional nature of the trajectories of these migrants only reveals the specificity of this border, which is one of the many hazards of their journey. Deportation practices reveal the tensions PAF officers must deal with on a daily basis.

Continuity of the deportation practices at the internal border
Before the Schengen Convention entered into force, when the Italy-France border was an external border, if an individual checked at a port of entry lacked the proper documents to enter the territory, he/she was simply denied access to French territory. This 'denial of admission' procedure was administrative. If the check occurred at the 'green' border, the French border police would arrest and deport the person, handing him/her over to their Italian counterparts. The exchange was done at the main border crossing point: the Pont Saint-Louis port of entry. After the Schengen Convention was implemented, ports of entry were closed. The denial of admission procedure was abolished. If a person is arrested in the Schengen zone, the deportation procedure is now the same as elsewhere in French territory. First the individual is taken into custody, then placed in a detention centre (a Centre de Rétention Administrative -CRA) via a decision of the Préfet, the local representative of the executive authorities.
However, during the Schengen Convention's implementation phase, most member states signed cross-border bilateral agreements with each other, including measures for police co-operation and readmission agreements. As regards our case study, these readmission agreements set out the conditions for the expulsion of third-country nationals residing legally in Italy but not in France, and for third-country nationals who entered EU territory irregularly and passed through Italy. Obviously these measures are reciprocal, and readmission agreements apply to the whole territory. However, in the bordering département a specific, simplified procedure applies: police officers do not need the Préfet's authorisation to implement the agreement. They return the individuals arrested to the Italian police on the other side of the border only with the permission of the latter's. Expulsions take place at Pont Saint-Louis, the same point where they used to be carried out before the implementation of the Schengen Convention. Thus, there is a form of continuity of deportation practices: they still involve the same kind of actors and the same venue. Similarly to checks in the border zone, the implementation of the Schengen Convention has not radically altered the border police's practices at the internal border, despite the elimination of systematic checks at ports of entry.

Facing the uselessness of controls
The striking point about this procedure is its apparent uselessness for all the protagonists involved. An episode observed during my fieldwork was very representative of both police officers' and migrants' perceptions of this procedure.
I was interviewing two PAF officers at the former port of entry of Pont St-Louis. It is situated on the old mountain road between Menton and Ventimiglia. Several years earlier, a new highway skirting the mountain had been constructed; the last active port of entry had been on that highway. There was barely any traffic on the old mountain road, on the other hand, which was almost deserted. The identification department of the Menton PAF had been posted at the former port of entry of Pont Saint-Louis, but no controls were held on this road. The national border was just a few metres higher up the road.
During the interview, we caught sight of a person walking down, partly hidden by the parked cars along the road. The officers came out on the road to bar his way. One of them shouted at him: 'No, you don't go this way, go back to Italy!' The young man came over to the middle of the road, smiling at the police officers, trying to negotiate. Other police officers came out of the building, shouting the same order at him. He finally went back up the road, very peacefully. The PAF officer told me he would manage to cross the border anyway, either later on by this same road, or else by the other road, by sea or by train. The man trying to cross the border did not seem frightened by the police in any way; however, he eventually did follow their orders. In fact, police officers actually carried out an illegal non-admission procedure on him on the basis of his presumed nationality, which was never checkedand at an internal border. 29 The nature of this border is constantly being negotiated. Its existence was denied by the migrant's gesture in the case above. In contrast, police officers assert its continued existence as a national border, though they have no legal basis for this claim in such cases as this. Their main concern in the case above was reaffirmation of the national border, even if it meant implementing an illegal order in my presence. This insistence on prohibiting this man's crossing appeared to be a symbolic act. A police officer from the Nice region admitted: 'Readmission doesn't work, it's useless, Italian border police officers only tackle the administrative tasks, and then they release the migrant.' 30 In fact, one of the reasons why the readmission procedure at the Menton-Ventimiglia border reinforces this uselessness impression is that readmission appears to be absolutely contrary to Italian authorities' interests. In cases when French border police cannot deport migrants to their home countries, neither can the Italian state police. Furthermore, there are many more French requests for readmission to Italy than the reverse; the ratio is '1 to 10'. 31 Therefore, implementation of the readmission procedure is difficult. As one police officer confirmed, talking about his Italian counterparts: 'That's for sure, they're not delighted! We have to explain it to them.' 32 Italian state police forces are supposed to come and search migrants being readmitted, but as a PAF officer from Menton told me: They don't do it every time.
[ … ] This morning for example, we stopped 35 Afghans on the train. They didn't come to take them, so we turned them back 25 metres away: 'RLT : raccompagnement aux limites du territoire' [escorted back to the limits of French territory]. 33 Even when Italian police come to detain the migrants, they usually release them a few kilometres away. 34 One of the officers working at the co-operation centre stated: 'In reality we are keeping score'. 35 Police officers tend to develop a very fatalistic discourse. G. from the Menton PAF service told me: We can do what we want, as long as the Turkey-Greece border is not secure … When the person decides to leave, it's already too late. They don't want to go back, they sell everything, they have nothing left. You can set the double objective, they will still be there. 36 T. added: The external borders are porous, once they get onto Schengen territory it's over! (repeated three times) in particular with the nationalities we cannot deport, we know that sooner or later, these persons will get to where they want to go. 37 However, a change in the focus of observation could provide some explanation for this procedure. The final aim seems to be postponing the arrival of 'undesirable' migrants in the northern area of France. This role is acknowledged by some officers: 'In Menton, we can't do much; we gather information on irregulars, so it's more like a slowing-down of the flow.' 38 Actually, the 'migratory' situation did not evolve much since 2010. The Vintimiglia/Menton border is still on the trajectory of migrants reaching the European shores and heading towards UK. In 2011, the arrival of numerous Tunisian citizens after the revolution at the Vintimiglia/Menton border led to a political crisis between the two Member states. The politicisation of this issue at the domestic and European level had several consequences. The French government reintroduced temporarily the border checks in Menton in April. At the European level, France and Italy asked for a revision of the Schengen Treaty which ended with the adoption of the 1051/2013 Regulation. Today, the tragic situation of Syria has had consequences on the origin of the migrants but the trajectory and the problematic (no possibility to deport the irregular migrants) remains the same. Between January and November 2014, 22,000 migrants have crossed the Italy-France border. The question has been once again under the attention of the media as the Interior ministry visited the PAF services in Menton. The staffs of the PAF have been reinforced on this occasion. 39 This situation was clearly put by one of the PAF officers I interviewed: Now, most of the immigration flux comes from Iraq and Afghanistan. We are barometer of geopolitics, when it goes wrong somewhere, it comes here! There are also climate refugees from the Indian subcontinent; we are the European barometer for the problems in the world. A lot of foreigners cross in Ventimiglia, a lot of things happen in Ventimiglia at the train station and in the area 40 . The current situation of this border confirms the assertion of this PAF officer.
The representations linked to border controls are often described as a 'masculinist and chivalrous logic' which identifies the nation as a 'proverbial "damsel in distress" and the border as a thin blue line in need of constant vigilance against risky foreigners' (Pratt and Thompson 2008, p. 622). In the case of internal borders, this is a distinct form of management of the population moving through and within the territory. The border is not perceived as a real 'fortress', but rather as an obstacle on these migrants' very long road. A constant dialectic of affirmation and challenging of the border can be observed among migrants and police officers themselves. Furthermore, this population is not associated with criminality or terrorism, but with misery. Police officers are not faced with an influx of aggressive and criminal foreigners, but an unstoppable flow of desperate individuals. The fatalistic discourse and the apparent uselessness of their daily work can be interpreted as the effect of the distance between the traditional role they embody at external borders and their representations of it, on the one hand, and their actual role at the France-Italy border, on the other.

Conclusion
This paper explores the resilient materiality of controls over individuals' mobility and how these controls are anchored to the edges of the national territory. This analysis is based on observation at borders where this control is supposed to have vanished: Schengen Area internal borders. Historical assessment of the border police's organisation, representations and practices throughout the second half of the twentieth century shows that the linearity of border checks has been a central stake in their institutionalisation. However, the legal and material resources available to border police officers have long been insufficient. Not until the mid-1970s did border areas became (more or less) legitimate zones in which to control individuals' mobility. The Schengen Convention and the lifting of internal borders within the EU have had paradoxical effects on border policing and the nature of borders. The state still exercises its power to manage individuals' mobility in a specific legitimate way inside the 'Schengen zone' at internal borders. Border checks do not take place in ports of entry on the borders any more, but national territorial delimitation has not lost all significance. Controls at national borders inside the Schengen Area have not disappeared; the modality of state control over individuals' mobility has been redefined. And this redefinition has to do with the specific process of European integration. These borders are at the core of the interaction between the construction of a new, European political centre, and the affirmation of an old one, the national political centre. Border police officers are in charge of dealing with the tension, or double bind of sorts, emerging from this interaction.
In the 1990s, the national legislative implementation of the Schengen Convention led to the creation of the Schengen Area; in France, this transferred the responsibility of dealing with the tension between the European principle of free movement and national migration control policy, to PAF officers. One of the practices used by border police to deal with this tension is targeting. The precise analysis of the content of plausible stories used by border police officers gives an account of the nature of the hierarchisation (Mezzadra and Neilson 2013) process that takes place at the border. But this hierarchisation is highly linked to the kind of agency that PAF officers enjoy over the 'crimmigrants'' bodies.
Furthermore, PAF officers develop a very precise representation of the migrants' itineraries and are very aware of the weak impact of their own work on migrants' journeys. Thus, border policing practices are both very locally embedded, and associated with a much larger perspective which overtakes this local nature.
National borders are still a setting in which the state manages individuals' mobility. This management can take the form of very tight and systematic checks, or of more targeted checks; in any case, borders are used as a legitimate location in which to regulate individuals' mobility. 1051/2013 of the European Parliament and of the Council of 22 October 2013 amending Regulation (EC) No 562/ 2006 in order to provide for common rules on the temporary reintroduction of border control at internal borders in exceptional circumstances). 12. Nonetheless, the National Police force has showed a growing tendency to conduct identity checks connected to immigration control objectives, on French territory, far away from its borders. The two tendencies are concurrent. The latter checks are held in accordance with the Code of Criminal Procedure, which allows a prosecutor to define a geographical zone and time frame in which identity checks can be carried out without usual justifications. 13. PAF officer from the Pont Saint-Louis office, November 2008. 14. The fieldwork at the Nice airport provides a good example of this characterisation of the border port-of-entry by plausible stories. As put by G. PAF officer at the Nice airport, 'it is an atypical airport, apart from the residents, the travellers are filthy rich, and the smugglers fear that [irregular migrants] would be too obvious for the police. So the airport isn't used a lot at the arrival'. 23. The 'readmission' procedure, in this case, is a deportation procedure between two member states of the Schengen Area. 24. I overheard a very friendly telephone conversation between the head of the PAF and the Moroccan consulate in Nice. 25. PAF officer at the Nice train station, November 2008. 26. There are some exceptions: the efficiency of PAF services is evaluated by their superior officers in terms of deportation statistics. The deportations carried out in the framework of the bilateral readmission agreement with Italy are not taken into account if the simplified procedure is used. So when a particular service's total figure for deportation cases is low, it might send individuals to detention centres even if it already knows it will be unable to deport them to their home countries. The migrants will be deported to Italy via the regular readmission procedure that applies to the entire territory of France, and implies detention. 27. G., a PAF officer from the Menton office, November 2008; he had been working in this service since the early 1990s. 28. At the time fieldwork was carried out, no deportations to Iraq or Afghanistan were possible because of the conflict situations there. 29. If my presence influenced the police officers' actions, I suppose they preferred to reaffirm the national border than to respect the law. 30. PAF officer at the Nice office, October 2008. 31. PAF officer from Menton, November 2008. 32. PAF officer from the Police and Customs Cooperation Centre, March 2009. The PCCC is situated in the former Italian port of entry at the Pont Saint-Louis; it is used (among other functions) for the exchange of migrants in the framework of the readmission procedure. 33. PAF officer from Menton, November 2008. 34. One reason for this is that the closest detention centre is in Rome, very far from the border in Rome. 35