Spatial reason of the state: the role of space in protest repression in Turkey

ABSTRACT Between 2007 and 2019 the Turkish regime used protest bans extensively in order to impede collective mobilization. In this paper, drawing on Michel Foucault’s discussion of raison d’état and an original dataset of protest bans, we examine these legal practices as part of the state’s repertoire of protest repression. We point to two limits against the indefinite extension of state regulation that Foucault identifies: an external limit posed by public law and regime of rights, and an internal limit that questions the effectiveness of ‘too much’ government. We argue that authorities use spatial control as a technology to negotiate these two limits. Specifically, authorities deploy the state’s prerogative of regulating public space as a ‘politically neutral’ legal technology to reconcile the banning of protests with the external limit posed by freedom of assembly. Spatial control also works as an effective form of government to negotiate the internal limits of raison d’état. We use illustrative examples to unpack the mechanisms of how spatial technologies neutralize protests to bolster an authoritarian regime. The study contributes to empirical research on protest repression as well as theoretical discussions on the rationalities of government by expanding the geographical scope of existing research to an autocratizing context.


INTRODUCTION
On August 25 2018, just a month after the removal of the two-year-long state of emergency in Turkey, the Beyoğlu District Governorate in Istanbul banned the 'Saturday Mothers' protest at Galatasaray Square on the busy pedestrian street Istiklal Avenue. This protest, organized by mothers of mostly Kurdish citizens who were 'disappeared' and murdered, had been ongoing for 699 weeks. Following the ban, the police forcefully dispersed the protestors, including elderly parents of the disappeared, detaining some in the process. They then encircled the square, not letting anyone enter. Since then, protestors moved their protest to one of the side streets. When asked about the violently enforced ban, the Interior Minister Süleyman Soylu stated: We banned the protest, that's right, because we wanted to end this exploitation and trickery. What should've we done? Let motherhood be exploited for a terrorist organization? … Should we just give up fighting terrorism … let them have the state, the country? We will not let concepts like motherhood, state, nation be used by those who have been their enemy for years. This nation has not let this country into the hands of their aghas hundred years ago, we would not let it into the hands of these beggards, either. ('Soylu'dan 'Cumartesi Anneleri' açıklaması', 2018) Soylu's framing of one of the longest running peaceful protest campaigns in Turkey as a direct national security threat is an extremely vocal manifestation of how the highest echelons of the Turkish state may see and treat peaceful assembly as a danger to the state and public order. Yet neither the ban nor the methods of its enforcement or the language used by Mr Soylu for its justification are an exception in a regime that has turned increasingly authoritarian in the last decade (Esen & Gümüşçü, 2016;Öktem & Akkoyunlu, 2016;Somer, 2016;Sözen, 2020). In fact, in Turkey between 2007 and 2019, the Turkish regime resorted to the use of protest bans in a regular and widespread fashion (Arslanalp & Erkmen, 2020a, 2020b. Initiated mostly by provincial authorities as administrative decisions in the name of protecting public order or security, often prior to a planned protest, these bans were in effect ranging from hours to months. When defied by protestors, they were regularly enforced by police violence. In this paper, we draw on Michel Foucault's discussion of raison d'état (reason of the state) to examine these legal practices of protest repression. In his work on governmentality, Foucault underlines that the problem of how to effectively deal with 'revolts' and 'seditions' to maintain 'good internal order' had historically played an important role in the development of raison d'état as a type of governmental reason (Foucault, 2007, pp. 267-270). We argue that this problem continues to occupy any state to this day, and one of its facets is the problem of how to govern and control peaceful protests that have become part of citizens' repertoire of participation. Nevertheless, as Foucault highlights, practices of raison d'état are commonly confronted by and come into interplay with two distinct limits against the indefinite extension of state regulation: an external limit posed by public law and a regime of rights that questions the legitimacy of government intervention, and an internal limit posed by liberal governmentality that questions the effectiveness of 'too much' government. In the arena of protest control, we claim that the former expresses itself in the domestic and international jurisprudence of freedom of assembly, while the latter is raised by considerations over the effectiveness of high-intensity repression. Governing demands that authorities manage the tensions that arise from confrontations with such limits.
In this paper, we turn to the case of protest bans in Turkey and ask: How does this interplay between raison d'état and its limits play out in the case of protest government in Turkey? What kind of techniques of government are used to navigate them? Our key argument is that space and spatial forms of government play a crucial mediating role in the Turkish state's efforts to negotiate the tensions between raison d'état, freedom of assembly and effective government. We argue that it is often by regulating public space and in the name of regulating public space that authorities curtail protests and render them ineffective. More specifically, authorities regularly deploy the state's prerogative of regulating public space as a 'politically neutral' legal instrument to reconcile the repressive act of banning a particular protest with the external limit posed by the domestic and international jurisprudence on freedom of assembly. Spatial control, however, is more than a legal technology addressing the external limit of raison d'état. We underline how it is also an effective form of government when it comes to protest repression and suggest that it is a way of negotiating the internal limits of raison d'état. A continuously shifting spatial demarcation of protest zones facilitates effective repression by pushing protests outside of visible and symbolically important places, or confining the negative effects of repression on the citizenry into demarcated zones.
These arguments draw on discussions of spatial governmentality but also rethink them in the context of an authoritarian state. Studies of governmentality in general and spatial governmentality specifically often focus on liberal democratic settings where spatial control is deployed against racial and classed others, emphasizing its difference from sovereign repression. In this paper, on the other hand, we illustrate how sovereign acts of repression, relying on legal and practical spatial technologies that are also observable under liberal-democratic states, are being deployed to bolster an authoritarian state.
The paper makes three crucial empirical contributions. We bring the largely theoretical and historical study of raison d'état into the empirical study of contemporary protest repression, hitherto unexplored. We substantiate our theoretical claims by drawing on illustrative examples from an original dataset of protest bans in Turkey , a case that extends the study of protest repression beyond its common focus on the West, which also largely informs the study of raison d'état. Studying contemporary Turkey, a paradigmatic case for a large number of regimes that have experienced democratic backsliding/autocratization (Bermeo, 2016;Lührmann & Lindberg, 2019;Waldner & Lust, 2018) brings into the light how different modalities of power and rationalities of governing operate in gradually autocratizing settings. We maintain that the tensions of governing we underlined above might be particularly acute in these regimes that attempt to dismantle the opposition while maintaining the pretence and legal framework of a democratic constitution (Levitsky & Way, 2010;Robertson, 2010;Schedler, 2013).
The remainder of the paper is structured as follows. After a theoretical introduction where we discuss the concept of raison d'état as elaborated by Foucault, specifically underlining its two limits, we discuss the role of space in protest regulation and repression. We then turn to Turkey to provide an overview of manifestations of raison d'état in the arena of protest repression and examine the ways protest bans have been deployed in Turkey to illustrate how space plays an important role in the negotiation of raison d'état with its two limits, identifying various mechanisms through which this happens.

FOUCAULT'S RAISON D'ÉTAT: BETWEEN SOVEREIGNTY AND GOVERNMENTALITY
In his lectures 'Security, Territory and Population', Foucault examines the development of raison d'état as a specific rationality of governing that takes the preservation of the state and expansion of its forces as its objectives. For Foucault, the novelty of raison d'état lay in its radical departure from the medieval conception of a political rule by rejecting any moral-transcendental grounds for ruling (Foucault, 2007, pp. 237-244). Instead, as a mentality of governing, raison d'état imagined the state as an autonomous object that had to be preserved and fostered by any means necessary (Foucault, 2008, pp. 4-5). The strength of the state, its integrity, security and wellbeing were supposed to be the primary objectives guiding the practices of government. Any other moral, political or legal concerns were secondary to the fulfilment of these objectives and had to be disregarded when posed as obstacles to the former. Raison d'état, for instance, always preserves the option of circumventing, suspending or violating the law when deemed necessary for the good and salvation of the state (Foucault, 2007, pp. 261-262).
Foucault therefore sees in the advent of raison d'état a major shift in the way political rule is conceived and consequently how power operates. According to Foucault, with raison d'état the question of effective government eclipsed the problem of legitimate sovereignty that have pervaded the prior reflections on political rule (Foucault, 2007, pp. 256-258). In practice, this was linked to the development of an administrative apparatus that not only executes the sovereigns' commands across the territory but functions as an apparatus of knowledge and regulation to ensure the proper circulation of goods and people in order to foster the state and its internal order (pp. 274, 325-328). In contrast to its contemporary connotations, Foucault therefore underlines that raison d'état-based government in its original meaning was not merely or primarily associated with repressive form of power but above all entailed a productive form of power, one that produces a new assemblage of knowledge, rules, and regulations to govern (Dean, 1999, p. 105).

Raison d'état and its two limits
A crucial contribution of Foucault's historical account of raison d'état for understanding the contemporary terrain in which state practices operate is his emphasis on how the raison d'état encountered two major limits in its latter development in European state practice. First of these is the external constraints that were imposed on state/governmental practices by a tradition of public law that sought to constrain the absolutist state on the basis of a new secular juridical regime of rights (Foucault, 2008, pp. 8-9, 39-40). This has culminated in the liberal democratic constitutional frameworks and the international human rights regime that put legal constraints on the exercise of state power on the basis of legitimacy and legality.
It is, however, Foucault's focus on the internal limits that emerged within the art of government that presents the highly original contribution that he makes, which also informs his entire project on governmentality. The rise of what Foucault calls liberal governmentality from the end of the 18th century onwards is an internal critique of the kind of governmental rationality that raison d'état and its unfettered apparatus of police represent. Unlike the external limits imposed by 'liberal' public law, this critique of raison d'état was not concerned with the questions of legitimacy, consent and the rights of subjects. Instead it spoke the language of utility that was already intrinsic to raison d'état, to question the effectiveness of what it considered extensive regulation in raison d'état's approach to the economy (Foucault, 2008, p. 40). This 'political-economic' critique was rather occupied with calculating the right amount of governing necessary for effective government. Liberal governmentality thus worries about 'too much' governing (p. 13), but it also wants to ensure that ensuing liberty is exercised in a 'disciplined and responsible manner' (Dean, 1999, p. 117; also see Lemke, 2019, pp. 188-192). This has led to complex articulations of freedom and security, which recurrently invent newer forms of securitization (Neocleous, 2007, p. 133).

Raison d'état and sovereign power
Foucault is unclear and uninterested about what happened to raison d'état after the advent of liberal governmentality aside from suggesting that these different rationalities are layered onto each other rather than eliminating one another as successive types of government (Dean, 1999, p. 102;Foucault, 2008, p. 107). In this paper we contend that Foucault's analysis about the conflictual interplay between raison d'état and its two limits is still relevant for analysing contemporary governmental practices that authorities enact in the name of protecting the state and its nation. In order to do this, however, we need to point to two criticisms of Foucault's articulation of raison d'état's relationship to sovereignty as a modality of power. First of these concerns the uneasy position raison d'état occupies in Foucault's conceptual schema on modalities of power. Foucault sees in raison d'état the transition to governmental modality of power, which is an imminent, productive and non-juridical form of power that seeks to shape the conduct of individuals as interest subjects (Lemke, 2019, p. 183). This is distinct from sovereignty that he associates with transcendental, juridical and repressive exercise of power over uniform rights-bearing legal subjects.
Yet, it is hard to dissociate raison d'état from sovereign modes of power when we reflect upon how raison d'état mandates and justifies an entire array of repressive practices including protest repression in the name of protecting the state, its internal order and future prospects, as Neocleous (2003, p. 45) points out. If deciding on exception is one of the most important sovereign acts à la Schmitt (1985) and Agamben (2005), then as Dean contends raison d'état as governmental rationality is always bound up with such decisions to the extent that it is concerned with identifying and neutralizing the so-called threats to the state from its internal and external enemies (Dean, 2013, p. 60). Moreover, contrary to a Schmittian conception, such sovereign decisions, competencies, and practices are always delegated to multiplicity of state and non-state actors including territorial bodies (Dean, 2013, p. 224; also see Feldman, 2010;Honig, 2009;Scheuerman, 2006), thus sovereign and governmental modalities are very much intertwined.
Second, we claim that Foucault's dissociation of raison d'état from the question of sovereignty leads him to downplay tensions intrinsic to the relationship between raison d'état and the interests of the ruler and/or ruling party. 1 While the notion of raison d'état alludes to the unity, homogeneity and autonomy of the state, neither the state nor its interests are monolithic or independent of socio-political struggles. How the interests of the state are conceived and articulated are always interwoven with hegemonic projects and the narrower interests of the rulers in building and maintaining their own power. Raison d'état in this sense is a discursive space that enables and justifies the connection of competing sovereign actors with the notions of national security and good of the state and the public.

SPATIAL TECHNOLOGIES AND PROTEST REPRESSION
In this paper we argue that space plays a key role in navigating the tension between raison d'état and its two limits in the specific arena of protest control. Our argument builds on a vast literature that uncovered the use of spatial technologies for the government of populations drawing on Foucault's discussion of modern operation of power. Scholars highlight that modern governmentality is often 'spatial governmentality', meaning that governance often works through 'managing the spaces people occupy rather than managing the people themselves' (Merry, 2001, p. 16). Reflecting on developments in urban environments, scholars such as Ewick (1998), Merry (2001) and Sanchez (2001) demonstrate how social control through spatial governmentality is achieved by excluding certain population categories from certain spaces instead of directly punishing them or using disciplinary mechanisms. These newer forms of social control focus on concealing or displacing offensive behaviours rather than eliminating or reforming them, displaying a logic of zoning (Merry, 2001, p. 17; also see Valverde, 2012). They are proactive in nature, attempting to manage risks 'by anticipating problems and preventing them rather than punishing offenders after the incident' (Merry, 2001, p. 16). Via these exclusionary mechanisms, spatial governmentality seeks to produce 'secure' spaces where responsible citizens can exercise their 'freedom' in a way that conforms to expectations of behaviour that is created by that space (see Beckett & Herbert, 2008, for a critical summary).
Scholars of protest mobilization have also explored the role of spatial technologies. They have acknowledged that space can serve to strengthen and camouflage practices of state repression and surveillance (Schwedler, 2012, p. 233) and underlined that states can use a range of place-based strategies (Nicholls et al., 2013, p.11) and the built-in urban environment to disrupt movements (Sydiq, 2020). Yet more could be learned about how contemporary states control and contain mobilizations through a lens that takes spatiality into account. Mitchell (2013) provides an illuminating example in this direction in his discussion of freespeech regulations as spatial regulations in the US case, demonstrating how by regulating where free speech can take place authorities effectively control and silence it: As courts in the United States have over the twentieth century moved away from a regime that penalises what is saidin essence liberalizing free speechit has simultaneously created a means to severely regulate where things may be said, and it has done so in a way that more effectively silences speech than did the older regime of censorship and repression. It could be argued, to put all this another way, a liberal approach to silencing oppositionto keeping it from being heardis to let geography, more than censorship, do the silencing. (p. 57) Mitchell argues that this regime of protest control, which has been referred as 'negotiated management' (McPhail et al., 1998) exercised in the United States and Europe (Della Porta & Reiter, 1998;Salát, 2017), has neutralized the potency of protests without banning them. While recognizing protests as legitimate acts of political participation, this approach seeks to control the boundaries of protests and shape the conduct of protestors through a range of spatial tactics and tools. However, when such tactics fail to prevent transgressive protests, police resort to more coercive tactics of strategic incapacitation, which for instance temporarily declares entire districts as no-protest zones (Gillham & Noakes, 2007, p. 343).
Through spatial governmentality, then, freedom to protest is exercised as long as it is not effective. Below, as we unpack the mechanisms through which protest bans are enacted in Turkey, we illustrate how protest bans also rely on spatial technologies, excluding protests from certain locations in the city, attacking their visibility and accessibility, severing their ties with possible allies; hence effectively limiting protests proactively. But what we observe goes beyond aforementioned accounts of spatial governmentality. Studies on spatial governmentality predominantly reveal the extent to which freedoms in (neo)liberal democratic societies are securitized and consequently tamed through spatial technologies of government without necessarily resorting to sovereign ban and punishment. A critical thrust of that literature also underlines the crucial role non-state actors play in shaping the conduct of individuals towards non-transgressive behaviour.
In this paper, we instead study administrative decisions to ban protests which are quintessential expressions of sovereign power. Bans are prescriptive pronouncements issued by the administration; they aim to be directive not enabling, nor do they leave much space for negotiation with the civil society. When challenged, they are enforced by police violence. They ultimately seek to eliminate the targeted protest rather than contain it.
Nonetheless, such sovereign acts have to grapple with the juridical and governmental limits Foucault identified. They thus make and declare exceptions to the right to protest using space as a legal justification for bans, articulating the repression of protest with the jurisprudence on the regulation of freedom of assembly, while simultaneously drawing on technologies of spatial governmentality. Moreover, as enactments of raison d'état, they are primarily embedded in a public discourse of national security, public order, and well-being of the nation and the state. Thus, what we observe are sovereign acts of repression relying on the legal and practical technologies involving space that are also observable under the liberal-democratic states yet deployed in ways to bolster an autocratic state whose interests are increasingly merged with the interests of the ruling elite/party.

METHODOLOGY
In this paper, we primarily rely on an original dataset of official protest bans issued by authorities between 1 January 2007 and 31 December 2019 in Turkey. Starting with the AKP's (Adalet ve Kalkınma Partisi -Justice and Development Party) second term in power, this period covers multiple AKP governments as well as the state of emergency (21 July 2016-18 July 2018) and its aftermath. These data were collected by the authors for a research project involving emergency politics, autocratization and protest repression. We previously examined this dataset to identify aggregate patterns of protest repression in Turkey with respect to issue, type and geographical distribution of bans (Arslanalp & Erkmen, 2020a, 2020b. In this paper, we use the dataset mainly as a repository of examples to study the spatial dynamics of repression. Protest bans are administrative decisions by authorities, generally governors, to restrict a protest or all protests, completely or partially. Our dataset includes only those cases where there is a declaration by a public authority (governor, district heads, rectors, public prosecutors) banning a protest or declaring a protest illegal. These include decisions that (1) ban all protests in a province or a district, (2) ban protests on a specific issue or at a particular location, (3) restrict protest tactics (e.g., forbidding marches but permitting a press statement), (4) restrict the participation of potential protestors by closing off access to potential protest sites (e.g., establishing checkpoints on the major roads into a city), and (5) declare local curfews. We had no separate category for spatial bans as multiple ban types (specifically second, fourth and fifth) referenced or worked through a spatial logic. In this paper, most of our examples are pulled from the second category, which includes bans that target a specific location.
In order to create the dataset, our research team went through the online archives of two newspapers, Hürriyet and Bianet, for every single day for the period, to identify and code protest bans across 28 variables in total. Hürriyet is one of the oldest mainstream papers owned by a major business group (Doğan Holding until April 2018 and since then pro-government the Demirören Group), and Bianet is an independent online news site with a focus on human rights issues and claim to independent journalism. We picked these two newspapers to provide a wide net and eliminate potential biases. With Hürriyet, we tried to maximize our geographical reach. As one of the oldest mainstream newspapers in Turkey, Hürriyet has a wide network of local offices, high circulation and a nationwide coverage. Bianet is known for its attention to human rights and opposition activity. Coupling these sources reduced potential biases and omissions from each source: potential inattentiveness to bans in the case of Hürriyet and comparatively limited coverage in the case of Bianet.
Our dataset covers the bans reported in these two news sources and does not claim to represent the entire array of bans enacted in Turkey in this period. We do not have access to a separate database of bans to cross-check our dataset. Possible alternative sources were either unreliable (such as the websites of governorates) or inaccessible (such as the police archives), especially under the conditions of State of Emergency (2016-18) when we started this research. Newspapers presented a public and relatively reliable source of data. They are also standard data sources when it comes to protest event analysis (Hutter, 2014;Koopmans & Rucht, 2002). However, given the research that has demonstrated how newspapers only partially cover the protest universe in many countries (Uysal, 2017, pp. 44-57), we expect our dataset to underestimate the actual number of bans.
Using these two sources, we identified 694 bans in total. Of these, approximately 20% are reported in both sources, 34% only in Hürriyet and 46% only in Bianet, justifying our decision to use two different sources. This also shows that in spite of much lower volume of daily news, Bianet's attention to human rights issues resulted in Bianet reporting more protest bans overall. When we compare these sources with respect to the issues or the geographical distribution of the bans covered, we do not observe any major differences (for figures on these issues, see Appendix A in the supplemental data online). When we look at the coverage over time by breaking up the data annually, however, we encounter a crucial divergence between the sources in 2019, after Hürriyet was sold to a pro-government business group. While the news sources were comparable in their coverage until then, we see a major slump in the reporting of bans in Hürriyet in 2019. Whereas Bianet reported 105 bans in 2019, Hürriyet reported only seven (52 of the 105 bans only reported by Bianet are the weekly bans on Saturday Mother's protests). This contingency again proves that not relying on a single news source in Turkey's authoritarian conditions was the correct decision.

PROTEST REPRESSION IN TURKEY AND PROTEST BANS AS ENACTMENTS OF RAISON D'ÉTAT
As has been extensively discussed in literature, in the last decade Turkey has become a competitive authoritarian regime as checks and balances that limit executive concentration of power eroded, basic liberties have been increasingly violated, and the incumbent party and its leader Recep Tayyip Erdoğan used the powers of the state to curtail political contestation and fair competition (Esen & Gümüşçü, 2016;Somer, 2016;Öktem & Akkoyunlu, 2016;Sözen, 2020). This construction of a populist competitive authoritarian regime was regularly justified in the name of ensuring the survival of the 'nation' and the 'state'. Yet this process never openly denounced or fully dismantled the formal edifice of liberal democratic constitutionalism.
Increased repression of protests has been, perhaps predictably, a part of this autocratization process. This increased repression, however, builds on a long history of repressive practices in the arena of protest management in Turkey. Rather than being accepted as a legitimate channel of participation, protests have been treated as threats to the state and its conception of normal public order despite being a constitutionally enshrined right since 1961 Gümrükçü, 2014;Uysal, 2017).
The current legal framework regulating protests, largely a product of the military regime of 1980-83, grants a wide array of powers to provincial authorities to securitize the public sphere and eliminate contestation when necessary (Bezci & Öztan, 2016). 2 While European Union (EU) reforms moderated some of its aspects in the early 2000s, legal changes since the Gezi Park protests in 2013 and Kobane protests in 2014 further enhanced the powers of the authorities. More importantly, a securitizing orientation towards protests has been deeply entrenched in the legal-administrative culture of the state  and shaped the attitudes of the police towards protestors, whom they perceived as 'marginals' or 'enemies of the state' (Uysal, 2017, pp. 104, 201). Indeed, after a brief liberalization from the late 1990s to the mid-2000s (the heyday of the EU accession period), protest policing has become increasingly repressive . The police violence that triggered Gezi Park protests represents both the culmination and the most visible manifestation of this increasing repression in the second decade of the 21st century in the context of broader autocratization of the regime.
Protest bans have been one of the major tools and indicators of this increased repression. As we extensively examine elsewhere (Arslanalp & Erkmen, 2020a), our data reveal an escalation and expansion of protest bans since 2012. Between 2007 and 2019, we found 694 bans enacted in 53 provinces including 220 curfew decisions in the south-east. 3 A total of 287 of these bans were enacted before the state of emergency, while 241 was enacted during the state of emergency (21 July 2016-18 July 2018). In the post-emergency era until the end of 2019, we observe 166 bans, of which 71 are weekly bans on Saturday Mothers protests (Figure 1).
Protest bans do not only repress their immediate targets. As public proclamations of state authorities, they also disseminate a particular discourse about protests and protestors that depict them as threats to public order, the state, and the safety and security of citizens among the larger public as the following text from Istanbul governor's ban on the Pride Week in 2016 illustrates: Within Law No: 5442, this request has not been approved as provocative acts and events may take place; and as it may cause a disruption in public order and in people'sincluding the participants of the eventtranquility, security, and welfare, when the terror attacks that have taken place in our country and the sensitivities that have emerged in society are taken into account. ('Valilikten Onur Haftası', 2016) Protest bans, thus, regularly portray protests as fraught with potential risks. Accordingly, protests are occasions where 'terrorist incitements' or 'provocations' can occur; they can disrupt the flow of daily life by creating crowds in already crowded areas; undermine the provision of services or commercial activity; and incite public disturbances by hurting 'public sensibilities', an allusion to the conservative social norms that the AKP regime promotes and claims to defend. As the act of banning-meaning-silencing of protest is presented and legitimized as a 'legal' act, done for the benefit of security of the population and public order, protest and protestors are concurrently constructed as 'illegal'. Indeed, depiction of street protests as violent and undemocratic acts that do not fall within the rule of law, and protestors as 'terrorists' and 'bandits' have been persisting tropes in President Erdoğan's discourse, especially since Gezi protests. 4 As it is always the case, raison d'état comes with a particular knowledge/definition of the friends and enemies of the state, which is also true for its enactment in the protest arena where protests are targeted by repression. For the Turkish state, socialist left, radical student groups, politicized Alevis (these three categories often overlap with contentious labour action), and the Kurdish movement have historically been among the potential or actual enemies of the state with shifting emphasis on one or the other depending on the political context (Atak & Bayram, 2017;Uysal, 2016;Uysal, 2017;Yonucu, 2018b). 5 Yet there have been changes and oscillations in the targets of protest as a new neoliberal Islamist hegemonic project consolidated itself (Arslanalp & Erkmen, 2020a). In addition to the overall designation of protestors by regime discourse as instruments of foreign powers that want to destabilize Turkey since Gezi, we see the framing of Kemalists, environmentalists, urban activists, LGBT and feminists as 'others' of the 'real' Turkish nation, embodied by the ruling party and its leader Erdoğan.

ROLE OF SPACE IN PROTEST REPRESSION IN TURKEY: NEGOTIATING THE LIMITS OF RAISON D'ÉTAT
We argue that space plays an important mediating role in negotiating the interplay between raison d'état and the two limits it encounters in the case of protest bans as forms of sovereign repression in Turkey. In the case of the first limit, space is deployed as a legal-procedural justification of protest bans to circumvent and reconcile the juridical limits of repressive behaviour which are posed by domestic and international human rights regimes. Concerning the second limit, which comes from utility calculations of governing actors, spatial circumscription of repressive practices is deployed for effectiveness.
Even a brief look at the type of bans imposed by authorities demonstrates the significance of spatial control and justification. Figure 2 shows that 312 of the 694 bans in our dataset are bans that target a specific protest or a location. As we discuss below, authorities often (but not always) prohibit specific protests on the basis of their location. Among the remaining we see 121 decisions that ban all protests in a province or district for a specific time period that may range from a day to a month; 19 decisions that ban a type of protest activity, such as marching, while allowing others, such as press statement; 24 bans on the mobility of participants, which bar them from entering a city; and 220 curfew decisions that completely suspend the mobility of citizens in a specific jurisdiction. The latter two are also spatially mediated bans. It is also telling that bans on all protests in a district or province have become commonplace since the state of emergency (76 instances under the emergency rule and 30 since its termination). States of emergency are by definition when basic liberties are legally derogated. In other words, they are moments when the public law's external constraints on raison d'état are weakened (Arslanalp & Erkmen, 2020a, 2020b.

SPACE AS A LEGAL JUSTIFICATION FOR REPRESSION
There are multiple mechanisms in which authorities use space as a legal basis and justification for suppressing protests. First, governors designate multiple areas within cities suitable for largescale collective gatherings (such as electoral rallies and meetings) as legal gathering spots and publicize them (see also Sydiq, 2020, p. 61, for the Iranian case). When groups plan collective action and inform the governorate as required by law, their application is regularly rejected with the argument that 'the gathering is planned in an undesignated spot', and as such is unlawful. While this gives the decision to prevent a protest a technical and administrative justification, the designation of protest spots in itself, as well as the decision to use this justification to prevent protests reflects political considerations.
For example, in the case of Istanbul, the decision to keep Taksim Square, a space historically significant and symbolic for the labour movement, outside of the list of designated spots for collective gatherings is highly contested. Between 2007 and 2019, this square was banned for May 1 demonstrations except for two years, and the ban was enforced with police violence and city-wide shutdown of transportation network and roads connected to the square. The removal of the ban in 2011 and 2012 and its reinstalment in 2013 demonstrates the changing governmental calculations on the freedom-security matrix. Since 2013, in Istanbul the governorate pushes large meetings to take place in highly isolated spots, which include two gathering zones in Maltepe and Yenikapı. These spots are recently built by land fill in the sea and are enclosed by highways, which is crucial in limiting the visibility and thus effectiveness of protests, as we will explain below.
Relatedly, throughout the period under consideration, there are numerous protests that are banned with the justification that their locationas chosen by protestorsis 'unsuitable for protests' for various reasons that relate to issues of public order or governors' duty to provide public order in a city. The locations that are targeted by the bans as unsuitable are very particularistic yet variable, potentially changing to include any place deemed unfit by the authorities. For example, they include the outside areas or vicinity of certain buildings such as a party or union building or a courthouse, or certain city squares and specific streets. 6 The texts of these bans list multiple arguments as to why these locations are unacceptable as locations for protestssuch as them being close to schools or public buildings, or being highly trafficked areas, and that protests there might disrupt business, tourism etc., as touched upon above. Such justifications are deployed in cases of both small or large protests. For example, Eskişehir governorate banned a protest planned by the Circassian Rights Initiative with the following text, using justifications that relate to the use of space in the city: As our province became an important center for tourism for both domestic and foreign tourism, as there are trips for domestic and foreign tourists who come with tour-busses to Odunpazarı square in this tourist season, especially on Saturday and Sundays, and since the place where the open-air meeting is planned is an area where there are historical places, museums, and spaces attracting tourists, there is no permission given to the demonstration at the economically and culturally significant Odunpazarı square by the decision of Governorship on 13/07/2011 no. 943. ('Eskişehir'de Çerkes Mitingi'ne İzin Yok', 2011) Likewise, in the case of one of the largest labour protests in Turkey after the 1980 coup (Tekel protests of 2010), the Ankara governorate used combinations of these justifications explained above. When workers wanted to conduct a sit-in and a press declaration in front of the union building in Ankara, the governorate declared a ban, arguing that the location where the protest was planned was not listed as a legal gathering spot and as such was unlawful, but also that the spot was very close to business and education centres and heavily used by the population ('Valilik Tekel Eylemine İzin Vermedi', 2010).
Through these case-by-case administrative decisions, then, certain locations in a large number of cities in Turkey are effectively declared protest-free zones for a limited time, though sometimes repeatedly, while allowing the authorities the ability to claim that they remain within the rule of law and respect the right to freedom of assembly. This is different from banning all protests in a city (which as indicated above gets more common under the State of Emergency and is still justified by referring to existing laws), as the bans here are very targeted and specific. Again, in these cases, the locations themselves can be popular protest spots and carry either symbolic weight or practical value, and the decision, while it uses a technical language, is highly political, as we will illustrate below.
Authorities' use of spatiality of protests as the justification of banning them reveals how the Turkish state appropriated the jurisprudence that has developed in the West over the course of the 20th century. Scholars of protest repression have shown that Western governments have institutionalized freedom of assembly as a right, but sought to regulate it with rules and procedures on the time, manner and place of protests (Özenç, 2015;Salát, 2017). In the United States, through the development of 'public forum doctrine', a series of Supreme Court decisions not only recognized protest and assembly as part of freedom of expression, but also tried to draw its legal boundaries in ways that is 'content neutral' (Mitchell & Staeheli, 2005, p. 800) As Salát (2017) shows in a comparative study, courts in Germany, France and the UK have also endorsed permit and notification systems and prior regulations on the time, manner and place of protests as long as they do not discriminate on content. Reconciling competing interests over the use of public space is particularly a crucial argument in such decisions (Salát, 2017, p. 285).
Nevertheless, while this spatial regulation of protest activity restricts and, therefore, as we have underlined above, neutralizes the effectiveness of protests, it often does not ban them (Mitchell & Staeheli, 2005;Salát, 2017). Instead, authorities negotiate and co-produce the spatiality of protests. In the case of protest bans in Turkey, however, unlike the goal of negotiating a constrained space of protest activity, the authorities simply use the spatiality of protest as a legal justification for its complete ban or repression. In the established legal positivism of the Turkish juridical order which has always created a legal foundation for circumventing legal limits on state practice (Parslow, 2016), restrictions on the basis of space serve as a convenient tool that authorities can point to as part of the liberal-democratic jurisprudence. Such reference is regularly observable in the incessant referral to legal frameworks and rule of law in authorities' justification of protest repression.
Competitive authoritarian regimes such as Turkey are particularly concerned with sustaining such democratic appearances for both democratic and international audiences (Levitsky & Way, 2010). One could even claim that for Turkish authorities, keeping the appearance of a state of law has long been part of the raison d'état to maintain a degree of international legitimacy as a longtime member of the Western alliances, which authorities considered central to a state's preservation and strength (e.g., Yılmaz, 1997) . It also aligns with the AKP government's populist discourse that has never entirely renounced democracy as a value.

SPATIAL REGULATION AS EFFECTIVE GOVERNMENT
If space is a legal medium for articulating the repression of protest with the jurisprudence on the regulation of freedom of assembly, it is also an instrument of effective government. In the Turkish case, while bans use space as a justification for restricting protests, they also attempt to decrease its effectiveness via regulating and delimiting spaces of protest in urban areas. They do this by concurrently limiting the visibility and public reach/contact of protests, as well as by severing the ties of protests to protest locations that are symbolically significant. Through these two mechanisms which in practice regularly work together, protest bans directly target the mobilizational capacity of protestors. Even in their most extreme formcurfewsthe bans reveal a possible calculation of effectiveness.
Targeting visibility and accessibility through spatial regulation Visibility and accessibility of protests affects the degree of participation as protests attempt to mobilize not only committed activists but also sympathizers, and aim to create new sympathizers from the disinterested. Public marches or protests in central locations such as city squares call to, or encourage passers-by to, join the action. An individual's decision to participate in collective action is strongly affected by others, thus making a protest less visible decreases its ability to inspire (Ketchley, 2017, p. 134) and inhibits any potential for 'positive feedback' (Biggs, 2003). This following quotation by a lawyer in response to the Ağrı governorate publishing a list that enumerated all the locations in the city where one cannot hold protests in 2010 demonstrates that these effects are clearly perceived and feared by the activists: They have effectively banned all press statements. Should we go to the plains and make a statement to the birds? You make a statement to the people, why would we do it somewhere where there are no crowds? They wouldn't even let us do a press statement in front of our union. ('Ağrı Valisi Kenti', 2010) Moreover, the centrality of a protest decreases the cost of participation for individuals as they are easier to reach and easier to leave, which becomes especially important if they are attacked by the police. This was illustrated clearly in the case of the Gezi protests of 2013 in Istanbul when protestors not only used side streets to run away from the police but also were assisted by the neighbourhood's residents who let them into apartments to protect them.
Lastly, protests are also a means of communication as they deliver the protestors' messages and concerns to the larger public. This is particularly relevant in the context of a country such as Turkey, where the mainstream media is now largely controlled by the regime (Yeşil, 2016;Waldman & Çalışkan, 2017). Unlike government rallies that receive a high level of media attention even when they take place outside of the main arteries of urban areas, oppositional actors have limited access to the public, and oppositional protests are cut off from media attention as well as their interface with the public when they are pushed outside the central areas of the city.
The effects of the imposed bans in the case of the Saturday Mothers as mentioned above is instructive. While protesting in Taksim Square has literally been impossible since the Gezi protests, the ban has extended to Istiklal Avenue, which is the main pedestrian avenue attached to Taksim Square, a usual path for marches. Through repeated bans that were enforced via police violence, protests were pushed away from the busy sections of the avenue towards its less populated end and towards side streets where they cannot be encountered by pedestrians. Among the targets of such bans, the most high-profile ones have been the Pride Marches since 2015, women's marches since 2018, environmental protests against the Artvin Cerattepe Mining Project and the case of Saturday Mothers. These tactics, moreover, are not only evident in the case of protest bans but in police interventions to protests as well, as encountered on many occasions where the police either separate protestors stationed in a location from the rest of the space by installing fences around them or encircle them with their bodies so as to cut their interaction with the public.
Beyond decreasing visibility and accessibility, spatial restrictions work to break the link between various oppositional groups or tactics. For example, governors regularly ban protests in front of courthouses when they know there would be protests in connection with crucial court cases involving the repression of opposition. This breaks the link between street protests and oppositional politics organized within the judicial arena, in effect working to isolate and decrease the visibility of the struggle that happens within the courthouse.
Targeting symbolic protest spaces through spatial regulation Protests also have a strong symbolic element that affects their emotional resonance and reach, and thus mobilizational capacity. Protests are regularly framed and get their meaning in their relationship to historical and symbolic contexts. They recall these historical and symbolic frames not only through the use of certain slogans, images and narratives, but also by their direct link to symbolic spaces. In other words, certain spaces carry symbolic significance and meaning for protestors, and it only makes sense to organize protests in those spaces that signify or recall specific struggles which increases their mobilizational capacity.
In the case of protest bans, one can perceive an attempt to break that link between protests and symbolically resonant spaces as in the case of bans on holding the May 1 demonstrations in Taksim Square. Another telling example is the commemoration of Madımak events, where in 1993 a total of 35 mostly Alevi intellectuals were murdered by an Islamist mob when they set the hotel the intellectuals were in on fire. In 2011, Sivas governor's ban specifically prohibited protestors from walking to the hotel and making a press statement in front of the building.
Such logic of repression is also evident in the case of Gezi Park in Istanbul, and the way it is closed to public access. Gezi Park was not on the radar of authorities before 2013 as a protest spot. However, following the Gezi Park protests in 2013, the park became a new symbol of resistance to authoritarianism and, subsequently, it is closed to the public every year on the dates of the 2013 protests: the police force everyone off the site and encircle the park with barriers to prevent access. In fact, the police confessed that the park is controlled because of its symbolic significance. The police declared that it will not allow a press declaration in the park, 'because it recalls historical events' when the family of Ali İsmail Korkmaz, a young men killed during the Gezi protests in Eskişehir, wanted to hold a gathering there in 2014 ('Gezi Parkı Ali İsmail'in', 2014).
Prohibiting protests in front of courthouses (or prisons), as recounted above, can also be seen in this category. For example, in 2017, the Edirne Governorate banned 'all protests within a 4 km radius of the F-type prison in Edirne' 'for public order and security' when parliamentarians from the HDP (Halkların Demokrasi Partisi -Peoples' Democratic Party) decided to hold a protest in front of the prison where Selahattin Demirtaş, the imprisoned ex-leader of the party, is located ('Edirne Valiliği F Tipi', 2017).

Spatial regulation as political calculation?
Lastly, curfews that were used in Kurdish towns from 2014 onwards demonstrate how spatial demarcation might work as a tool of sovereign suppression, while at the same time reflecting a possible calculus of effectiveness that might take into account electoral concerns of the governing party as well as international reactions. Within the context of the ending of the peace process between the state and PKK (Partîya Karkerên Kurdistanê -Kurdistan Workers' Party) and the onset of military operations into the Kurdish towns, in 2015 and 2016 curfews strongly enacted the raison d'état of the Turkish state, projecting the sovereignty of the state over towns, in order to establish territorial control in reaction to the declaration of autonomy by the Kurdish movement, in this case the sovereignty being unhinged from any limits set to it by human rights.
Yet, even in this case, the government, instead of declaring a state of emergency in the whole region or putting whole cities under curfew such as governments did in the past, chose to declare, extend and end curfews on certain districts, neighbourhoods and streets in a reiterative fashion, where the bounds of curfews were extending and retracting continuously as the decision added or dropped neighbourhoods or streets. For example, in the case of the city of Diyarbakır, the heart of the Kurdish region, which was targeted heavily by curfews (96 curfew decisions in Diyarbakır in the dataset), the curfews started in 2014 during the Kobane protests, which entailed violent clashes, with a decision to put the whole city under curfew ('Altı Kentte Sokağa Çıkma Yasağı', 2014). The curfews became more targeted in 2015 and 2016, targeting districts, neighbourhoods or even streets, with decisions declaring curfews in new neighbourhoods as they were lifting curfews in others ( 'Sur'da 5 Mahalle', 2016). In effect, while this meant that certain neighbourhoods remained under curfew for extended periods of timefor example, parts of Sur district remained under curfew for more than three years and were effectively depopulated ('Diyarbakır Sur'da Sokağa', 2017)there was a fluid and complicated geography of violations of human rights created by the curfew decisions which included or excluded targets by demarcating zones of sovereign suppression. One possible interpretation of this could be that the government was seeking an effective calculation of how to exert control without estranging the whole population of the city, combining sovereign, governmental and political rationalities.

CONCLUSIONS
In this article we examined the Turkish state's repression and government of protests in light of Michel Foucault's discussion of raison d'état as a specific form of governmental rationality that puts the preservation and strengthening of the state as its central justification for government practices. Furthermore, and more specifically, drawing on Foucault's discussion on the external and internal limits of raison d'état, we underlined the mediating role space plays in the interplay among the objective of securing state's interests, external limits posed by public law and human rights, and the internal governmental limits grounded on the question of effective governing in the arena of protest control.
Our close study of protest bans in Turkey illustrated how authorities use spatial regulations as a legal technology to justify their bans as well as a technique to increase the effectiveness of protest control. Spatial regulation appears as a politically neutral legal mechanism that helps authorities to formally reconcile their particularistic targeting of specific protests with the formal universalism of the established jurisprudence. While the neutralization of effectiveness of protest activity through spatial control has also been identified and criticized in the Western liberal democratic contexts, this study suggests that in Turkey authorities appropriate this jurisprudence not only to limit the spatiality of protests via negotiated management, thus to contain its effectiveness, but in many cases to unilaterally ban the protests. As such, we observe a form of protest government that is much more intertwined with sovereign repression than existing studies on spatial governmentality in liberal democratic settings highlight. This is manifested in the very act of banning protests and also in the kind of national security discourse that encapsulates these bans.
These spatial bans are effective tools because protests gain power from public visibility, accessibility and emotionally charged symbolic embeddedness. By isolating protests to secluded and symbolically irrelevant places, they precisely inhibit such characteristics. Furthermore, as the case of curfews in Kurdish areas suggests, spatial demarcation also creates distinctions between repressed populations and non-repressed ones, enabling the containment of the most negative consequences of repression that may expand the popular backlash to those in power. Hence, the spatial production of repressed zones and non-repressed ones is part of authorities' calculations on the most effective way of governing populations that negotiate the elimination or neutralization of 'threats' with the maintenance of the state's legitimacy and the government's popular support in an electoral regime (see also Harcourt, 2018;Yonucu, 2018b).
We have maintained that Turkey as a case of a populist competitive authoritarian regime may be emblematic of state strategies of managing the tensions of protest government under such regimes and the gradual autocratization processes that culminate in them. These regimes are defined by their everyday authoritarian practices backed with a nationalist-statist discourse of security amidst their continuous claims to be committed to democratic norms in domestic and international politics. Nevertheless, we also do not want to reify the difference between regime types when it comes to the techniques that are used to contain protests which are seen as 'threats' as demonstrated by recent trends in mass detention of activists prior to international events, declaration of no-protest zones, and militarization of police forces in liberal democracies such as the United States, France or Germany. Indeed, authorities in Turkey frequently latch on to such contradictions/antinomies that pervade liberal democratic rule and weaponize the discourse of rule of law to argue the democratic compatibility of their practices. The case of Turkey therefore can also serve as a warning sign, pointing to the protracted and under the radar erosion of democratic rights through the everyday practices of administrative bodies.
Although this paper examined how raison d'état has been enacted by state actors in the arena of protest repression in Turkey, it is important that future research explores more closely the concrete effects of these practices, as well as how these practices are contested, resisted and negotiated on the ground. In fact, activists actively test and push the limits that authorities try to enforce, aware that the authorities weigh various electoral and political calculations against each other. We have observed this most recently in the cases of women's marches in Turkey, which the authorities in recent years repeatedly securitized and criminalized. When the Beyoglu district authorities in Istanbul banned the planned women's march on 25 November 2019, subsequent and persistent demands on the part of the organizations led to the withdrawal of the ban.
Yet the administration allowed the march with a new spatial limit: the march and meeting could happen only insofar as protestors do not pass beyond a specific spot on Istiklal Avenue, restricting the protest to one-fourth of the avenue. When some protestors wanted to march beyond, the police used teargas and plastic bullets against women, in effect publicizing the violent image that the regime rather avoided when it comes to women.
As such, a protest ban is a contested process as much as a decision and its outcome is produced through the interactions between state and citizens. In Turkey, in spite of repeated efforts to confine protests into controlled areas and the massive use of force in the past five to six years, the struggle over who indeed controls public space continues, as activists look for ways to respond to the limits placed upon them. If the measure of effective government is the successful shaping of the subject's conduct, Turkey's authoritarian government could not yet fully produce the docile subjectivity it has long desired.