Affective bureaucratic relations: File practices in a European deportation unit and criminal court

Indifference has long been acknowledged as a crucial affect to the continuation of bureaucratic practices. Recently, the production of more diverse and layered affective modes in bureaucratic institutions is increasingly highlighted. However, how affects differ within and between sites saturated with ‘paper work’ remains an understudied terrain. In this paper we focus on the relations that are formed in daily file-work within two state institutions: a Deportation Unit and a Criminal Court. We draw on ethnographic fieldwork in order to show that a) affects are locally produced in the relations that are mobilized in file-work b) these affects are unevenly produced within and between different bureaucratic practices. By comparing two different bureaucratic settings yet related in the subjugation they demand of the bureaucratic referent of their practices, we aim to put forward how differences in bureaucratic practice come with their own specific affective modes, showing that bureaucratic practices are saturated in, and thrive on, diverse affects of varying intensity. Bureaucratic action is a deeply affective practice, within which the relationship between caseworker, casefile, and the file’s referent is carefully calibrated. With this intervention we position ourselves within scholarship that complicates perceived dichotomies between rationality, still often associated with bureaucracy, and affect. Developing sensitivities towards such variety of bureaucratic affect offers nuanced perspectives on file-work and what kind of sovereign power the relations that are made through file-work subsequently allow to be reproduced.

File-work: Mediating affect 'We need to be able to take decisions, to act effectively. If the human aspect comes on top of that than it would become much more difficult'. These words of a Deportation Unit official fit the vast body of literature highlighting the dehumanizing effect of bureaucracies on those who populate these settings -the bureaucrats. Such approaches often tend to understand bureaucracies as sites lethal to human spirit and the imagination (Graeber, 2015) or as instruments of dehumanization and rationalization (Bauman, 2000). As sites too, that may elicit fear on the side of the 'clients' of bureaucracies, located in the threat that bureaucrats will 'wrap you in paper' (Gefou-Madianou, 1997: 139), or will 'waste your time' (Carswell et al., 2019). What fuels these understandings of bureaucracy might be what Michael Herzfeld (1992) described as 'the social production of indifference': an indifference of state bureaucracies to their subjects that can only be experienced as humiliating and, at times, violent. 'Why', Herzfeld in his classic work therefor wondered concerning the bureaucrats that keep these bureaucracies going, 'do some people turn into humorless automatons as soon as they are placed behind a desk?' (Herzfeld, 1992: 1).
However, recognizing bureaucratic affects to be mere indifferent is a generalization that does little to shed light on the variety of everyday affective practices of bureaucrats. Crucial to the reproduction of bureaucratic work, these practices deserve a more nuanced recognition. Affects have to be made, and are so in relations between the file, the file referent, and the caseworker. These bureaucratic relations constantly change throughout a procedure, partly due to the division of work that is characteristic of bureaucracies. Comparing bureaucracy to an octopus, Ten Bos (2016) highlights how bureaucracy can never be found in one (central) place but instead is dispersed. As such, no grand overview but only traces of bureaucracy can be found (see also Van Oorschot 2014b, 2020. In a similar vein, we looked into affects mobilized in bureaucratic relations with the modest ambition to catch glimpses of 'bureaucratic affects' rather than offering definite explanations about them. To shed nuanced perspectives on indifference as being but one bureaucratic affect is necessary not to glance over its complications. For example, as Bosworth recently showed, an affect of emotional withdrawal on the side of immigration officers in relation to detainees not simply enables them to continue their daily job. Bosworth shows this affective mode to come with painful consequences, located in the corrosion of morality in the work practices within these institutions, both on the side of the detainees as well as the officers (Bosworth, 2019). Another example comes from Hertoghs, who shows how bureaucrats tasked with the heavy responsibility of decision-making in asylum procedures navigate compassion (Hertoghs, 2019). Indeed, 'bureaucratic decision-making involves not only the expected dry stuff but also emotions and senses' (Kuus, 2019: 621). Empirical insights derived in local settings nuance the idea of the mere indifferent bureaucrat. This gives rise to questions on how and where which affect arises, and what variety -instead of a homogeneity -of affects can subsequently be identified.
To be clear, in this article we do not seek to reject the idea of a production of indifference as a technique to the continuation of bureaucracies. Rather, we recognize indifference to be one of a variety of affects mobilized in practice. We therein understand the production of affects to take place in a relational context. As Heyman alerts us, Lipsky's work on streetlevel bureaucrats already spoke 'to ways in which material and ideological work conditions of state workers affects how they carry out their tasks and treat the public that they encounter' (Heyman, 2012(Heyman, : 1277. The insight that encounters between bureaucrats and their work setting bring forth bureaucratic affect is what we wish to take along. To sensitize ourselves towards the varieties in the mobilization of bureaucratic affects we put forward an empirical exploration of two ethnographic encounters with 'dealings with dossiers': on the one hand practices of bureaucrats in a Deportation Unit, on the other practices of caseworkers in a Criminal Court. 1 In short, rather than focusing on how bureaucratic systems make bureaucrats indifferent, as it were affectively unavailable, we here wish to contribute to approaches highlighting the variety of bureaucratic affects. Doing so, we not only focus on affects concerning the bureaucrat and the referent. Large swaths of bureaucratic work get done in the relations between caseworkers and their casefiles rather than in interactions with their 'clients' (Scheffer, 2010). We therefore stay close to the materiality of the work and focus on filework. In file-work, bureaucrats, bureaucratic material, and the subject of bureaucracy as the file referent, come together. In our settings we witnessed a variety of affects to arise andimportantly -were sensitized about the specificity of their mobilization depending on the relations that are made between casefiles, file-workers, and file-referents in specific moments of the bureaucratic practice that brings them together: file-work.

A note on affective relations in state bureaucracies
If the reader is somewhat uncomfortable with our emphasis on affective investments given the violence that the practices in our sites account for, he or she is certainly not alone. Why do we wish to take these mundane bureaucratic practices seriously at all, if we have larger political fish to fry? We emphasize the affective texture of everyday file-work precisely because it denies these practices a semi-transcendent existence as 'the State' -a problematic move that reifies and totalizes the power attributed to the state to begin with. Such reification of the state as a mystified reality behind political practice is problematic because it fails to do justice to 'the State' as a historically contingent and variable thing (see e.g. Abrams, 1988). Moreover, our emphasis on the affective texture of these practices also complicates their appeal to rationality (cf. Mathur, 2017: 4). whereas it is rationality that is so crucial to their legitimacy (Borrelli and Lindberg, 2019). While complicating bureaucratic affects we thus recognize that the bureaucracies we engage with here are human-made instruments of state power.

On comparing a deportation unit and a criminal court
Our two case studies show, first of all, the need to do justice to differences within and between bureaucratic practices. We highlight their variety by nuancing the ways in which, along a file trajectory, caseworkers are positioned vis-a`-vis the casefile and the subject or 'referent' of the casefile in our different fieldwork sites. Secondly, associating bureaucracy with mere indifference not only seems to obfuscate the social production of affects besides indifference but also, and relatedly, the moment and location wherein which affect is required and generated in what kind of procedure.
Tracing the trajectory of casefiles in two settings, we thus aim to show that file-work is implied in multiple affective investments. These investments may translate into indifference towards the casefile's referent as the opening sentence illustrated. In another part of the procedure however these affectivities might change and can allow for empathy towards that same referent. Moreover, in both our fields we observed how bureaucratic materiality mobilizes affectivities as well. Concern might arise for a casefile's quality, velocity, and completeness. And a conversation between two administrative clerks shows files to be the object of care for them, accusing colleagues of neglect: 'The judges . . . don't get me started. They spill coffee on their files, they lose them on the train on their way home . . . And then come in here asking us where the files are. For all I know she lost them herself!' The clerks' complaint alerts us to the difference in affectivities that can arise between 'file-workers', here the clerks and the judges.
A comparison allows to show this variety in affective relations and how affects are distributed differently and enacted in particular interactions. Bureaucratic affects are mobilized in the changing relations involved in the files' transformation that unfolds as a legalbureaucratic procedure. We thus propose to think of bureaucratic affects as crucially situated. In other words, we see the production of bureaucratic affects as intimately connected with the constantly changing relations build along a file's trajectory between the file, caseworker, and the people 'behind' the file.

Following cases as methodological approach
We draw on fieldwork between 2013 and 2016 wherein one of us immersed in a Deportation Unit and the other in a Criminal Court. These hard-to-access institutional settings make the data that we present here unique, adding to the contribution we make to bureaucratic practice in settings that are, like our field sites, saturated with sovereign power.
To Wissink, full research access to the Deportation Unit was finally approved by the director following a personal meeting in 2015. This meeting was arranged via-via after months of networking, studying accessible sources and gathering knowledge about the organizational culture. The Unit, part of the Immigration Office, is itself divided in a handful of subunits that are each populated by dozens of bureaucrats. Caseworkers' movements are mostly limited to such a subunit, and are responsible only for a specific transformation on a deportation file before it is transferred to a next desk. For this reason, the researcher followed files throughout their trajectory. She complemented these observations with dozens of semi-structured interviews with immigration officers. Van Oorschot received access to the criminal law section of a court in 2013, having requested permission first, through the national Council for the Judiciary, through a formal and centralized procedure. And subsequently with the local Criminal Court in question, through a series of conversation with the head of its criminal Law section. This access, concretely, entailed access to its two criminal law floors where she was stationed in one of the shared offices of the court clerks. The researcher was privy to a host of work practices, and especially to the administrative workers', clerks' and judges' file-work. Tracing cases from their entry, into the court, in the form of an often-incomplete file to the court session, 'shadowed' individual judges were 'shadowed' in their desk work and accompanied to court. The researcher further invested in reading the casefiles herself prior to these largely desk-bound activities.
In these settings, the authors spent hours, days and months among files and bureaucrats. Despite the ethnographic approach we took in our fieldwork, we did not actively participate in the file-work itself. We were not qualified or authorized to do so, nor ethically inclined to do so. However, our presence alone undeniably made us part of the relations of these practices, but we focused on being present, and could witness firsthand the 'backstage' practices that constitute these bureaucratic sites. Both researchers found themselves in a unique position to observe work practice, elicit impressions, and ask further questions about specific cases. Both researchers also dependent heavily on informal conversations on site, not only while behind desks but also with the people present over coffee or lunch. From this rich collection of data, what struck us were caseworker's varying, at times seemingly conflicting affective investments in their jobs -jobs that are heavily concentrated on filework.
In the following, we want to further situate our argument by tracing our way through a body of literature on both casefiles and caseworkers in legal-bureaucratic practices, as well as by delineating scholarship on affects and bureaucratic practices. Having done so, we will elaborate on relevant differences and similarities in our two fieldwork settings. Then we turn, first, to a discussion of a deportation file trajectory, and second, to a discussion of the casefile's travels through the Criminal Court. Concluding, we discuss the contribution of highlighting, through a relational approach to bureaucratic practices, the nuanced specificities in the local production of affects in state practices that are so unambiguous in their violent consequences.

Bureaucratic actors
The concerns touched upon in the above -the production of affects in bureaucratic filework-position us within the ethnography of bureaucratic practices. There, it is particularly instructive to think along with those working on street-level bureaucracies and everyday work practices (e.g. Dubois, 2016;Herzfeld, 1992;Lipsky, 2010;Scheffer, 2007). Also, it has been demonstrated that caseworkers have significant levels of agency and discretion in their dealings with cases (e.g. Heyman, 2009). Rules and regulations do not dictate or determine a practice in any straightforward sense, but are rather used, evoked, interpreted, or challenged, as part and parcel of legal-bureaucratic practice (Dupret et al., 2015). Both on the side of these bureaucrats as of bureaucratic 'clients', too, Gruß shows in a study on border management bureaucracies that '[a]ffective responses and relationships remain . . . integral to the bond between humans and bureaucracy' (Gruß, 2017: 3).
At the same time, Weber's 'bureau-cracy' always already alerted us to the materiality of bureaucratic work -that is, it takes place behind a desk, the site of the production and use of official documentation. Files are crucial mediators of state power -and in that capacity, a crucial matter of concern within everyday bureaucratic practices. Without files, it is sometimes said, there would be no state or legal system to speak of (Vismann, 2008). Studies of the agentic capacities of files have helpfully addressed these salient dimensions of bureaucratic practices. According to Navaro-Yashin (2007: 94), 'state-like structures make themselves evident to the persons who inhabit their domains in the form of materialities'. Casefiles are for instance active in the epistemological sense of the word: they help translate occurrences and persons into legally recognizable events and personae (Van Oorschot and Schinkel, 2015), which 'documentary doubles' become active in a host of case-making practices (Van Oorschot, 2014a, 2020). Hull (2003Hull ( , 2012 addresses the way casefiles in Pakistani state bureaucracy allow for the translation of individual agency into state agency. In other accounts, casefiles do not simply feature as texts to be read; they are also, and in some practices first and foremost, material objects (Riles, 2008;Van Oorschot 2014a, 2020. Objects though with the crucial ability to change form, which enables subsequent action to be mobilized, ensuring the continuation of bureaucratic movement. The transformation of a file is indeed dependent on a host of practices engaged with their materiality. Pushing paper is sometimes quite literally that: rolling files down a hallway, adding piles of ready-made files to mail boxes for the next unit to pick them up, attaching an approved photograph to complete an already existing file, or excluding from a file a document that is considered to have become redundant. Where files come from, how they 'ripen' and grow, what ends up in them and what gets excluded -these are everyday questions making up large swaths of bureaucratic casework.
However, emphasizing files -or bureaucrats -sketches a rather actor-centered picture of bureaucratic work. Instead, our point of departure for analyzing what it is that makes the daily bureaucratic clock tick is no actor but a practice; file-work. In file-work, a relation is mobilized between bureaucrats, bureaucratic referents, and bureaucratic materialities although the exact appearance of these relations -the constellation they form -changes throughout a procedure. We look into the affects that arise from these relations to be able to ask what they matter to and say about bureaucratic practices.

File-work as an affective bureaucratic practice
Affective modes that contribute to bureaucratic action have to be made in file-work. As casefiles go, travelling through the hands of bureaucrats, changing desks, collecting references to their bureaucratic subject, relations formed around file-work constantly change. It follows that relations wherein affects arise might be interrupted, newly formed, or change in their intensity. Affect thus has no fixed form in the file-work of our settings, rather, '[a]ffect arises in the midst of an in-between-ness: in the capacities to act and be acted upon' (Seigworth and Gregg, 2010: 1). Take for instance Latour's ethnography of the French Conseil d'Etat (2010). There, administrative workers make sure the file 'ripens' and take care of its completeness. The casefile is an object of care and concern: will it ripen well? The counsellors however approach it with weary hesitation and studied caution; if the casefile still incorporates echoes of the human misery the casefiles testify to -a dispute between the state and an individual -counsellors must remain impassive, untouched, as the Law must speak. The documents in Latour's ethnographic work become affective in filework. The relations that are formed along a file's trajectory 'produce and effect affect' (Navaro-Yashin, 95: 2007).
There lies an important distinction to make, as Mercan argues in the context of interpreting a suspects' hearing in a Criminal Court, between 'emotional expression' and 'unconscious affect' (Mercan, 2018). At stake are not different, clear-cut emotions such as rage, sadness, or joy as 'discursive representations' (Fotaki et al., 2017: 4). Instead, we pay attention to 'the movement between bodily states and the intensities that this evokes' (Fotaki et al., 2017: 4). We understand affects then as more everyday and subterranean investments and attachments (Seigworth and Gregg, 2010). As such, affects always incorporate a dimension of fantasy, for instance of conceptions of the good life (Berlant, 2011), or, as we will show, of the good file -which 'good file' may come to resonate with conceptions of the right way to get things done. In that sense, affects are not pre-political but participate in certain economies of value and attention: they imbue certain dimensions of practice with importance while others may be rendered peripheral. But they are also situational and sticky: they adhere to certain bodies and objects more readily than others (Ahmed, 2010). The administrative clerks talking about the carelessness they observe on the side of the judges, a carelessness concerning certain aspects of the file that is judged from the clerks' owndiffering -affective relation to the file, forms one example. Similarly, Reed (2008) shows how prison intake forms attract varying affects. While they are treated with boredom verging on carelessness by prison warders, prisoners in contrast invest these documents with a measure of agency. In a context in which much has been taken away -contact with friends and family, material possessions, part of one's identity -the documents 'stand out by their positive presence, as an object provided, rather than taken away,' (Reed, 2008: 163). These affective modes that are mobilized in file-work are tangible, then, in the dual sense: they speak both of the way these materials are touched -collected, cared for, read, discarded -as well as for the way they touch their users (Navaro-Yashin, 2007). In our fieldwork sites, we develop these accounts of the relationship of caseworkers with casefiles and case referents. Aiming to pay attention to the 'affective nature of the administration of justice', we take their relationality as sites of production for 'institutional affects' (Bosworth, 2019: 543), showing that these relations in different legal-bureaucratic settings -both between and within sites -come with their own affective investments and intensities.
A deportation unit and a criminal court: Sites of affective file-work The 'state': Ethnographies of concrete practice As mentioned, this paper is based on ethnographic encounters with file-work in two settings, a Deportation Unit and a Criminal Court. Both deportation and criminal law are closely associated with the exercise of power and violence. They are both not instances within which the individual seeks the law by claiming rights -e.g. asylum requests, or civil law -but these practices rely on attempts to seek the individual and forcefully bring him or her, as it were, 'under' the law. Over and against approaches that emphasize underlying social or political logics, both research projects were interested in the everyday bureaucratic practices that go into producing 'the State' here. Given the fact that the work in both settings is mediated to a large extent by casefiles, we aimed to understand how these abstractions of the state are made real in and through daily file-work: concrete operations of gathering and disposing documentation, creating files, making decisions, processing cases and people. The first site, the Deportation Unit, is part of the federal state's administration. It is crucial to documenting the recognition -that is, creating bureaucratic evidence -of people as deportable from national territory. As such it is a means of population control that has been recognized as a punishment (Weber, 2015). The exclusionary effect of deportation practice lies not only in the act of deportation itself -often preceded by pre-removal detention -but also in the state of 'deportability' (De Genova, 2002) produced in these bureaucratic practices that remains to haunt the illegalized body even though it stays on the territory. Criminal law practices, our second site, are arguably part and parcel of similar sovereign operations of power. Although the border of the territorial sovereign is not at stake here, criminal law is nevertheless active in ordering bodies, and, through imprisonment or forced labor, confining them 'in order to prevent [them] from engaging in any real movement' (Deleuze and Guattari, 1986: 45). Criminal law and immigration law are thus 'both systems [that] act as gatekeepers of membership in our society, determining whether an individual should be included in or excluded from our society' (Stumpf, 2006: 396-397).

Calls to action: Different futures, different uses
In both practices, file-work forms the point of departure for a majority of the job. In both sites too, casefiles are active in forms of procedural justification -may they be organizational or juridical ones -as they carefully trace their own histories (through stamps, dated testimonies, autographs, authorizations of various kinds), while their standardized appearance is evocative of the promise of equality before the law. They are part and parcel of the State's 'machinery of sameness' (M'charek et al., 2013): while casefiles' content may vary, they must obey the same rules. However, the uses and prospect of the casefiles within our fields differ in important respects.
The Deportation Unit is organizationally speaking an administrative organ charged with the execution of legal decisions: the trajectory of files-for-removal starts with the decision to file an order to leave the national territory, including detention with the prospect of deportation. The Criminal Court in contrast is itself the location of legal judgment. This means that the file-work in the Deportation Unit commences with the legal decision to deport the subject of the file after which the majority of the work consists of processing these files. Deportation files that are contested in court form the exception. The court only comes in between if a file's trajectory is disrupted, for example when a lawyer calls for a hearing. This situation most bureaucrats try to avoid because of the risk, so explains a senior bureaucrat, that a judge comments that the already ordered decision that the Unit aims to effectuate 'might reflect the law' but that one however 'has to take the person into account too'. The Criminal Court, instead, is the site within which a decision still has to be made.
Cases, then, are themselves different things: in the Deportation Unit, it is a legal decision that calls for further courses of action; in the court, the action is geared towards making the decision while the consequences of this decision (e.g. imprisonment) are acted upon by other actors further downstream. The outlook of the file, the futures the file is directed to that mobilize present file practices and the relations formed around them, are different -as we will show below.

Trajectories of 'files-for-removal'
The first concern that occurs around an incoming file in the Deportation Unit is whether to add an order to leave the territory to it addressed to the individual the file refers to. This order might also state to detain this 'deportee' with the prospect to deportation. Once such a decision is composed, it is printed out on a clean paper that gets folded in between a bright colored carton cover: the file-for-removal. Much still has to be done in the space in-between decision and deportation. Cells in the detention center have to be reserved and equipped to deal with dietary requests of detainees, for example during the Ramadan; embassies of the (suspected) 'return-country' need to be mobilized for cooperation so that travel documentation can be obtained; flight seats need to be reserved. Different desks that the file is transferred to allow for different actions, while the file gathers materials that account for these actions. For each and every gathered document there is a specific location in the file: mugshots are stapled on the left inside, logistical information is stapled on the front, legal decisions stay on the inside -well protected. As the file goes, a high-ranked official in the Deportation Unit explains, a collection of different subunits 'takes care that the file is ripened to actually make the trip'. They 'feed the file', like another bureaucrat explained. These hints at a certain aliveness of the file elaborate on the actions that these files call for, but also point toward a certain 'affectual relation' between bureaucrats and files (Fotaki et al., 2017: 7).

Affective relationships between bureaucrats and files-for-removal
During a training for newly hired employees of the Immigration Office the first power point slide reads: 'All bureaucrats need to pursue one shared goal: To make the right decision based on a file of quality'. But what do caseworkers mean when they speak of quality? It became clear that this depend on the location of the file. For senior bureaucrat Rose, responsible for acquiring travel documents, the quality of files depends on specific documentation: 'Most important are the fingerprints and the photos', referring to the mugshots of the deportation file's referents. 'We need good photos. There should not be a fridge on the background or something like that.' The movement of files is another dimension to the 'quality' of files; movement indicates the progress of the file whereas stagnation might obstruct the process completely. For example, the legal decision to detain a person with the prospect to deportation needs to be signed within 24 hours after the individual was arrested. Also, the completion of the file has to be achieved before the juridical term for detention ends. The pacing of the work in the Deportation Unit is thus quite high: without movement a file risks dropping out of its trajectory.
'Files are coming in, they are put on a desk, and if someone is not there for three days it stays put there. If it stays on that desk for six days you lost them. For 100 files that means 600 days!' 'What exactly is lost?' 'Time! The file lies still. The faster a file goes, the faster it leaves.' Paul, deportation caseworker, shows concern about the loss of precious time. The faster the file goes the better: there is no time to waste because the decision to deport that marked the starting point of files-for-removal comes with a procedural expiration date. But not only do caseworkers take care to meet procedural deadlines, they also prefer files that allow this kind of speediness. William, a rather new bureaucrat, explains how his preference for 'fast files' navigates him in selecting files: 'If it is a fast file I will work on that one first so that it is already done.' 'What is a fast file?' 'When there is not much in the database.

Then you have to open less documents and your motivation [for a decision] is done faster.'
Files are fast in the material sense: working with a small number of documents is simply quicker than opening a file composed of many. But fast files are also met with optimism, relieve even, allowing for a relatively easy satisfaction without too much hesitation or reasoning. Indeed, affects here mobilize action on the side of the affected body (Seigworth and Gregg, 2010: 2), in this case the bureaucrat selecting what file to push for deportation.
Lucy, working on adding travel documentation to files, yet defines the normative feature of 'her' files as follows: 'It is a good file when it is easily treated. When it is a Romanian, we say "ooh pfff, peanuts!".' She refers to the probability to obtain the sought-after documents indicating that the good file is connected to the expectations of bureaucrats to be able to 'feed' the file satisfactory. The chances for the file to get stuck on this procedural requirement are considered as well; without the cooperation of the embassy of deportation destinations, a file trajectory is disrupted and the case might be lost. Bureaucrats anticipating this possible loss does not rarely raise unique, file-technical problems of its own. As some countries demand two identical photos to come along with a request for travel documents, the question may arise how to add these photos to the form in a way that the requested laissez-passer will be issued. 'How am I exactly supposed to continue?' One puzzled bureaucrat comes over to the desk of his colleague. 'Just attach the photos', the colleague waves the question away. A few minutes later, the phone rings. The colleague replies to the other side of the line: 'Just staple it. No, I don't have glue. Just two staples.' Putting down the phone the bureaucrat elaborates to be well-considered: 'I am a stapler, not a gluer'. There lies satisfaction in taking neat care to do it just right, to assure the file's quality. Poor 'quality' of the file can cause delay, stagnation, and even cancellation of the file as a deportation case. Concerns arise when this quality depends on the preferred response to the documents by a third party like the embassy: diligence is put into doing it just right, causing insecurity about one's skills to deliver a 'good' file. Indeed, '[a]ffect permeates organizations profoundly, influencing people's motivation, their political behavior, decision-making and relationships' (Fotaki et al., 2017: 4). Such affective incentives create categories of preferred files crucially depending on the task distributed to a caseworker which is intertwined with the procedural location of the file.

When the referent enters: Navigating conflicting affectivities
The person subjected to the envisioned deportation has to be present in the file even though in a highly specific, partial form. File-work enables these necessary translations, by which the file referent features more as an absent presence (M'charek et al., 2014), or rather something that haunts the procedure at times. In the words of Paul once again: 'You need to be able to take a distance from what you are doing . . . Many people in here treat it just like a "file". And that is good.' However, there are moments that the encounter with the referent becomes part of the filework given juridical procedural rules; here, face-to-face interaction intrudes on the caseworker's preferred face-to-file interaction (Scheffer, 2004). At the time, one of the main reasons why a file did not pass court was that files lacked a 'hearing' form, indicating the importance a judge in the council for alien law attributes to this act of procedural care for the referent. Caseworker Kevin however feels trapped in a catch-22: the law calls on bureaucrats to deport but they have to take a referent's possible resistance into consideration too. 'We do not take our "duty to hear" very seriously. Someone who really has a serious argument will bring that issue up through their lawyer before that moment [in court]'.
To allow the voice of the referent into the file is not only difficult in a social psychological sense; it also endangers the affective values of swift work and lean files itself. It is at such moments that the relations that are made in the file-work cause conflicting affects. This paradox is heightened in the 'return conversation' with families, yet another procedural requirement. While preparing such a meeting, William expresses his reluctance: 'Such a conversation always ends with tears. Those people get a call from us stating that we will discuss their stay: they do arrive with some hope. But it is always about return . . . after such a day I am completely worn out. It is really terrible.' 'Then why do you schedule these conversations?' 'Because, for families, you have to frame [the decision] even better. Before you can arrest them and remove them you have to be able to prove that they received information about voluntary return.' William hints at his struggle to combine different affective relations, both integrated in the procedure, that occur around these 'return conversations'. To schedule these conversations with the files' referent implies an interruption of the otherwise carefully guarded detachment of bureaucrats from the file referent. But including the obligation to 'hear' also respond to a procedural care, not only to the completeness of the file but also maybe to the referent. The case-worker is torn, becomes 'worn out', because of the conflict that the combination of these affects form: care for the referent and care for the file require detachment from the other. William has to open the file to the referent whereas at the same time he is affected by the call to relate himself to referents of flesh and blood and their hopes to interrupt the process of deportation.
The presence of several affects in itself is not unusual in bureaucratic practice. After all, should such a 'collaboration' between detachment and care not ease the work rather than making it so exhausting? Although this might be the case when these affects are consequently mobilized on the same subject -care for the file enables detachment from its referent -, what is so wearisome about these conversations is that the affects that transpire are ambiguous in their call for action. The care for procedure calls for these conversations in the first place but are themselves installed as procedural care towards the referent. This conflicting call put on bureaucrats working on deportation was also recognized by Bosworth among officers in pre-removal detention centers in a similar way, including the confusion this dual relation causes: 'Institutionally they are exhorted to maintain distance from detainees, while having to work with them intimately' (Bosworth, 2019: 544). Case-workers like William are indeed confronted in these conversations with the impossibility to remain detached from the referent. In other words, the bureaucratic affects mobilized in the different relations that intersect in the 'hearing' interaction appear hard to reconcile: the relation bureaucratreferent and bureaucrat -casefile. What makes these conversations so exhausting is that the affectivities they mobilize are heterogenous: a conflict arises between affective modes when file-work does not allow to circumvent their crossing.

Distributed work and invisible investments
In the Criminal Court things look different. Tasks are distributed here, too, but the legal decision has yet to be made. When files arrive at the Court, they meet the Court's administrative staff, charged with registering the files and 'taking good care' of them. This means that administrative workers check the components of the file -are all present? Are they dated correctly? -and fill out forms about the file's content, which they later attach to the sleeve of the file itself. Whereas the deportation file starts with the legal decision that calls for gathering more documentation, in the Criminal Court the file arrives with documentation included. Casefiles enfold both procedural and evidentiary materials such as witness reports, victim statements, parole service reports, and photographic evidence. However, the files may still be incomplete; in that case, additional documentation may be added to it while the file 'rests' in the Court's file room. The file, in one of the administrative workers' words, is 'alive'; their task is to manage its growth.
Again, highlighting a preoccupation for the file's 'growth', its aliveness, matters. Like the Deportation Unit, the Court faces high caseloads -leading to a rather high pacing of 'through-put' -and like the files in the Deportation Unit, procedural terms and statutes of limitations have to be observed in each individual case. The administrative workers' activities, then, are both 'logistical and legal' (Latour, 2010). Logistical, because these pertain to the management of the file's ability to move forward, from file-room to judges' desk, to courtroom. And legal, as procedures have to be carefully observed. This kind of work practice, as is the administrative workers' perception earlier cited in the introduction, is largely taken-for-granted and 'invisible' to those working with the file's 'textual content', that is clerks and judges (cf. Star and Strauss, 1999).
Invested in providing the judge with a neat file, assisting clerks use a variety of techniques to make the file ready for use by the judge. Not only do they reorder the file's componentsfiles can be a mess -they also code different components of the file using color-coded stickers, for example blue for the police's observations at the crime scene, orange for a criminal record, or red for the defendants' statements. (see also Van Oorschot, 2014a). This work deemed so important clerks will hoard these color-coded stickers to ensure that they may never do without. Clerks also make summaries of the case which judges will use in their own case-summarization practices. For this summary, clerks have to strike a fine balance between the demand for completeness -a good summary includes all relevant information -and brevity -a good summary does not include informational noise. A good file, for them, is one that affords their 'end-users', the judges, easy access to the 'cases' in question. A well-ordered and well-coded file assists them in doing so. Again, much clerical preparation work, caring practices that go into making the file ripe for use, remain just below judicial levels of perception when they do their job well. If however the clerk is new and relatively inexperienced, judges tend to comment on his or her faulty preparation. This caring work, like in the Deportation Unit, leans on a material incentive: of ordering and shuffling documents, of marking and coding.
Contrary to a file's trajectory that starts with a decision in the Unit, judges have to make their decision at the end of the casefile's trajectory in Court. In order to arrive at just decisions, judges try to aim for a judicial kind of objectivity. Understanding their work as situated in possible tension with societal demands for harsh punishments, judges not rarely point to the importance of them 'keeping a cool head' in order to see the facts clearly, without prejudice. This may entail, as Judge Laney suggests, that 'you'll have a night of sleep over it', when you are faced with a difficult decision. Or that you 'follow the steps,', as Judge Beech elucidates: 'you look at the procedural elements of the case, then you look at the specific charges, then, the evidence.' This 'following of steps' is not only a formal requisite. It is also a material practice of spreading out the file on one's desk, carefully highlighting salient phrases, and marking important dimensions of the case on the copied summons.

Allowing oneself to be affected
Especially crucial to 'keeping a cool head', judges emphasize, is to allow oneself to be 'open' to what may transpire in Court.
'You always have an idea of who you're going to meet in Court,' Judge Jameson relates, 'but you have to remain open to it as well. Your impression might be entirely wrong: I had a domestic violence case once. You think, reading the file, what a terrifying brute to hurt his wife like this! And then, the next day, you see this tiny little man, accompanied by this wretch of a wife . . . That changes your perspective!' As part and parcel of conceptions of individualized justice (see Hutton, 2013), the unique personal circumstances of the defendant are important ingredients of their decision-making practices. While these concerns on the part of judges include appraisals of his or her risk of recidivism, they also consider the question of remorse and the foreseeable unintended consequences of a verdict. For example, when a defendant is in dire financial straits a fine may be deemed more punitive than necessary. It is precisely in this sense that the referent of the casefile and the file itself stand in a much more dialogic relation to one another than in a bureaucratic practice like the Deportation Unit. In Criminal Court, while the defendant is present 'in' the file, he or she can still contest the narratives it offers in court; the case is not decided or closed yet.
Judges' preferred affective relation to the file and to the file's referent, then, is one of allowing oneself to be affected, but not prejudiced -a process assisted by not hurrying and by hesitations (see also Latour, 2010 for the crucial role of hesitation in legal practices). 'Keeping a cool head', however, is not the same as affective detachment. According to the judges, weighing the case, turning it over in one's thoughts, is and will always remain 'human work' or a 'craft', not an 'exact science'. Indeed, complete affective detachment from the file's referent -the defendant -is something to avoid, as a proper decision has to be rooted in an appreciation of the severity and the context of the case, as well as the defendants' personal circumstances. 'Some of these people, it partially just happens to them,' Judge Kingsley sighs, when referring to defendants' often difficult upbringing and precarious, marginalized existence. Approaching the case 'with a cool head', then, does not quite mean one is insensitive to these personal circumstances; while revenge may best be served cold, justice, to these judges, requires the cultivation of a prudential affectivity that is sensitive to both the severity of the case and the sometimes dire personal circumstances of the defendant. That is, in order to make just decisions, judges have to allow themselves to be affected by it contrary to the file-workers in the Deportation Unit where a disruption of detachment to the referent confuses the file-work of effectuating an already taken decision. As Judge Emerald reflects: 'We've got an ugly power, you see, to shape people's lives in such a far-reaching manner.' For decisions to be just, they have to be responsive to the specific texture and weight of the events in question. In order for them to appreciate the gravity of the case, as well as possible extenuating circumstances, these criminal law judges tend to value hesitation and reflection on the case.
The relations unfolding around file-work in a Criminal Court, then, are varied and create different affectivities. Affectivities here also seem hierarchically distributed. Where administrative workers and clerks are charged with material care for the file, judges instead may allow themselves to be affected by its contents. Although they want to remain impartial, they cannot remain indifferent -even though indifference may be a 'structural effect' of these legal practices. This exercise of prudential judgment predicated on the capacity to 'arrive at a clear picture' before the court session, and the capacity to think things over, in other words the capacity to hesitate, is in many ways crucially dependent on the diligent work that administrative workers and clerks put in the transformation of a file. A good file, from the perspective of this prudential affectivity, is one that allows easy access to its informational substance, including the personal circumstances of the defendant. If judges may concern themselves with the referent of the casefile, they may however remain quite indifferent to the casefile's materiality: that, after all, is the clerks' job. The paradox that arises between 'hearing' the subject and responding to the call to effectuate an already taken decision, as occurred in the Deportation Unit, then, is here stretched out over various kinds of workers. What the comparisons between our two fields emphasize is that the affects that are mobilized in file-work are shown to be local, unevenly distributed accomplishments.

Situated affectivities and violent effects
In the preceding pages, we have explored the production of various institutional affects in bureaucratic practices. We particularly looked into affective investments that we observed in file-work, more precisely in the relations that are formed on a files trajectory between caseworker, casefile and file referent. Asking empirically what different kind of affects are mobilized in the relations that become possible through file-work, and how and where, is a move away from understandings of bureaucracies as merely governed by plans, reasons, commitments or procedures. What appear as merely two exemplars of a larger bureaucratic logic of cutting up tasks and distributing responsibilities, are two sites of bureaucratic practices saturated with differently distributed and varying affectivities.
While indifference may certainly appear as an effect of these practices especially to those who are brought 'under' the judgment of the Deportation Unit or Criminal Court, these practices are saturated in affects such as diligence, concern, and exhaustion, too. In the Deportation Unit, procedural demands that 'open' the file to the voice of the potential deportee are treated with weariness because it messes with the detachment of the bureaucrat to the 'human aspect' whereas the concern for procedure -the same procedure that does require this care to the referent -does not allow for too much of consideration towards these subjects. Here, an uneasy conflict of affects appears making it difficult to negotiate their dual affective call. In the Criminal Court, material concern for the files is delegated to 'lower' levels in the Court's hierarchy. These practices set the stage for the mobilization of a judicial prudence, within which an affective relation with the file referent is desired -if always something to be kept prudentially in check. And while administrative personnel and court clerks may remain somewhat detached from the individual 'behind' the file while they are invested in its material shape and content, judges instead explicitly aim for the casefile to move them, to be affected.
The affective modalities in both cases differ when paying close attention to where and when they occur; in which set of relations. These affects change throughout a procedure and their variety -at times conflicting, other times collaborating -shapes these practices. Affects are not isolated to a singular bureaucratic actor. They have to be engineered, mobilized, in a relational setting. By treating affects as mediated and local accomplishments in the relations mobilized in file-work, our comparative approach is also an attempt to avoid a partial obfuscation of these practices caused by a dominant focus on an omnipresent affect like indifference as a mere effect of bureaucratic practice. Indeed, the necessity to 'remain' distant, or rather to aim for detachment, hints at the familiar affect of indifference in these practices as Herzfeld recognized it. However, to understand this demand for 'detachment' simply as a psychological, adaptive technique (as in: they have to think this way, otherwise they wouldn't be able to do this job) does not teach us much about the specifics of this detachment. The file-work we attended to put forward bureaucrats to be affectively invested in doing a good job, which requires different modes from them depending on the dense knotting of relations they get caught up in through file-work. Their job does not merely require an 'indifferent' or 'detached' affective mode but a variety of affects depending on the relations involved in the file-work at a specific moment. Differences occur within and between bureaucratic practices, also depending on what affect is required where and when. The comparative character of this paper emphasizes the importance to pay attention to how affects offer insights into the specificities of bureaucratic practice. To learn more about bureaucracies and their violent yet normalized effects, we need to be sensitive towards the particularities of where which affect becomes productive, and how these affects can be mobilized in the relations that are built in the file-work.
In conclusion we therefore wish to share hopes with Berlant that: 'while one can't intend an affect, one can become attentive to the nimbus of affects whose dynamics move along and make worlds, situations, and environments' (Berlant and Greenwald, 2012: 88). If our goal is to inquire into the way bureaucracies shape our lives, indeed how bureaucracies distort and compromise our worlds -and in our cases violently so -reflective questions need to be raised about the situated presence of affects in bureaucratic relations. Studying not its violent effects but the affective relations made at the level of everyday practice is a way of relating ourselves politically to these bureaucratic practices. In this, this article is not 'critical', in that we do not seek a position outside of the state, distancing ourselves from it in order to criticize it (and hence running the risk of reifying its power); rather, we sought a relational and engaged position in recognition of our shared and compromised social reality.