On militant democracy’s institutional conservatism

Introduction

The history of democracy seems inextricably bound to the perception that democracy is under threat and, thus, in need of defence. One reason for this perception is the idea of a ‘democratic paradox’ which became central in constitutional theory in the advent of World War II. Democracy, the paradox implies, can be abolished from within. Anti-democratic actors can utilize legally and institutionally admissible means to pursue their political goals. As one possible response to this paradox, interwar legal scholars theorized militant democracy, the idea of ‘pre-emptive, prima facie illiberal measures to prevent those aiming at subverting democracy with democratic means from destroying the democratic regime’ (Müller 2012, 536). Originally conceived as a response to fascism, ¹ militant democratic formulations were incorporated into constitutions especially after the end of World War II. Until 2015, these constitutional provisions have informed party bans in 20 European countries alone (Bourne and Casal Bértoa 2017). Paradoxically, however, militant democracy gradually fell out of interest in political theory after its institutionalization (but see especially Sajó 2004).

Despite sustained legal and empirical discussions of militant democracy, Müller’s assertion that ‘there exists no general legal or, for that matter, proper normative theory of militant democracy’ holds well into the 2010s (see Müller 2012, 536; Thiel 2003; Thiel 2016). Yet, current phenomena such as populism, non-liberal political parties and developments towards illiberal democracies in Europe have reinvigorated the search for theoretically viable responses. In this context, proposals for exclusionary measures against purported anti-democrats have re-emerged and resulted in two systematic normative retheorizations of militant democracy. ² Kirshner’s (2014) A Theory of Militant Democracy and Rijpkema’s (2018) Militant Democracy draw on militant democratic thought from the interwar and postwar years. Yet, they are wary of hasty and excessive interventions which might result from arguably normatively underdeveloped early militant democratic theories. Therefore, they offer explicit normative justifications for democratic militancy and develop more restrained, ‘minimal’ defences of democratic self-correction (Rijpkema) and equal participation (Kirshner).

This article traces and critically discusses parallels between the historical militant democratic thought of Schmitt (2004), Loewenstein (1937a, 1937b), and Popper (2020), and Kirshner and Rijpkema’s contemporary theories of militant democracy. Doing so, it offers two arguments, one broadly historical and one normative. With regards to the historical argument, it reconstructs militant democratic theories’ convergence on an ‘institutional conservatism’. ³ The concept of institutional conservatism functions as an umbrella term to capture a theoretical preference for institutions that restrain transformation. Although united by this structural feature, I show that the historical trajectory of militant democratic thought is characterized by a transformation from a ‘substantive’, proto- or anti-fascist, to a ‘procedural’, anti-totalitarian, institutional conservatism. ⁴ A substantive institutional conservatism theorizes institutions that hinder transformation of the existing order, whereas a procedural conservatism encourages transformation but contains and limits it within the boundaries of existing institutions. With regards to the normative argument, I argue that the historically grown, procedural form of institutional conservatism remains democratically problematic because it might serve not only the defence against anti-democratic actors but simultaneously systematically restrain democratic transformation.

I substantiate this argument in four steps. In section two, I reconstruct how institutional conservatism is introduced in Schmitt’s (2004, 2012) Legality and Legitimacy. I show that Schmitt’s essay reflects a preoccupation with varying institutional guarantees for ‘order’. This undergirds the essay’s polemic against the liberal idea of procedural legality and introduces the substantive-conservative idea of an immutable core of constitutional values, rooted in a homogenous people. In section three, I demonstrate how Loewenstein’s essay in the American Political Science Review (1937a, 1937b) and Popper’s (2020) The Open Society begin to transform militant democratic thought towards a procedural conservatism. Loewenstein’s anti-fascist theory is characterized by an ambiguity between an insistence on the defence of democratic institutions and democracy’s ‘absolute values’. Unlike Schmitt, Loewenstein does not assume a fundamentally given order rooted in a homogenous people. Popper’s theory, then, externalizes value commitments to an ‘open society’ and theorizes democracy as an institutional instrument for containing transformation. In section four, I show how the militant democratic theories of Rijpkema (2018) and Kirshner (2014) remain implicated in this procedural institutional conservatism, not only despite but precisely because of their attempts to further restrain and ‘minimalize’ democratic militancy. Finally, in section five, I use resources immanent to this reconstruction to argue that by reproducing the proceduralized institutional conservatism of their forerunners, contemporary militant democrats render sedimented institutional choices difficult to revoke. This produces problems from a democratic perspective, especially under conditions of a non-ideal institutional status quo and potentially democratic challenges to it. Therefore, in concluding the article, I call for theoretical engagements with the idea of defending democracy that move away from institutional conservatism and towards a perspective on democratic practices that emphasizes transformative capacity over institutional restraint.

Schmitt – substantive institutional conservatism

German legal and political theorist Schmitt is frequently referenced as a theoretical forefather of militant democracy. To be sure, Schmitt’s membership in the National Socialist German Workers’ Party renders his work an unlikely point of departure for defences of democracy. Nevertheless, his 1932 essay Legality and Legitimacy, which is concerned with the defence of a value-laden constitutional core, is widely considered to anticipate theories of militant democracy. But Schmitt’s oeuvre remains an uneasy starting point for contemporary theorists (who either extensively justify or altogether avoid references to Schmitt). Recently, militant democracy has been problematized because of a Schmittian ‘decisionism’, which here refers to the necessity to politically decide who is identified as a potentially repressible anti-democrat. As Invernizzi Accetti and Zuckerman (2017, 186) point out, militant democracy relies on the arbitrary, that is, political decision on what constitutes the constitutional core of democracy and what constitutes threats to it.

Contemporary militant democrats, however, may still justify their project as an attempt to overcome this arbitrary character of Schmitt’s theory. In keeping with a wider tendency in the history of ideas to separate Schmitt’s earlier decisionism from his later, fascist concrete order thinking (recently e.g. De Wilde 2018), ⁵ Schmitt’s pre-1933 work is considered to be open to various political projects. For his theory of a ‘militant constitution’, Rijpkema (2018, 78, 110) acknowledges, ‘the constitution does not necessarily need to be democratic’. ⁶ Legality and Legitimacy, according to this reading, can still be interpreted as an indeterminate defence of a strong, value-laden constitutional core. This baseline theory, then, would need to be complemented with determinate, democratic value justifications to be normatively tenable (Rijpkema 2018, 190; Kirshner 2019, 64–67). ⁷ Retheorizations of militant democracy, then, claim to offer more explicit justifications and to determine conditions under which democratic defence serves democratic ends (Rijpkema 2018, 190; Kirshner 2019, 66).

It is neither the purpose of this article to reiterate basic disagreements on Schmitt’s significance for democratic theory nor to further emphasize the claim that militant democracy is decisionist in nature. It is worth stressing, instead, that a full appreciation of Schmitt’s role in militant democratic thought should not exclusively centre on the theme of decisionism but take into consideration Schmitt’s particular brand of institutional conservatism. Schmitt’s theory constitutes the starting point of the trajectory of institutionally conservative thought that I trace in this article. While characteristic for Schmitt’s wider oeuvre, this institutional conservatism can be reconstructed particularly well starting from a proto-militant democratic fragment of Legality and Legitimacy that contains the essay’s ostensible commitment to a constitutional core:

I agree with Hauriou that every constitution recognizes such fundamental “principles,” that these principles belong to what Carl Bilfinger terms the “constitutional system,” which is, in principle, unalterable. We are also in agreement that it is not the intent of the constitutional provisions concerning revision of the constitution to open a procedure for the elimination of the system of order that should be constituted through the constitution (Schmitt 2004, 58).

The militant democratic interpretation of this fragment centres on Schmitt’s initial emphasis on ‘fundamental principles’, on the defence of constitutional values. This interpretation, however, omits the significance of the second sentence. ⁸ Read in its entirety, the fragment devaluates procedures by exempting a given system of order from the scope of constitutional revision. As Invernizzi Accetti and Zuckerman (2017, 186) point out, this implies ‘a direct rejection of the reduction of a constitutional core to amendment procedures’. This anti-procedural conception of the constitution is central for Schmitt’s theory. Militating against the ‘formal’, ‘legalistic’ character of the Weimar constitution’s second part (and contemporary legal positivists like Kelsen), Schmitt takes the side of those legal scholars who would eventually support National Socialism’s subversion of the Weimar constitution under the auspices of the anti-positivist idea of ‘substantive’ constitutional values (Schulz 2019, 110).

Beyond this opposition to democratic proceduralism, it is striking that the constitution is merely endowed with the passive function of ‘recognizing’ fundamental values of a constitutional system, which express the immutability of the constituted system of order. Schmitt’s emphasis on fundamental values, upon closer inspection, is thus concerned primarily with an order expressed through these values and its immunization against transformation. At least in Legality and Legitimacy, Schmitt’s ‘decisionism’ is not as open-ended as some militant democratic interpretations make it seem, which can be demonstrated with reference to the essay’s conclusion:

A constitution that would not dare to reach a decision on this question; one that forgoes imposing a substantive order, but chooses instead to give warring factions, intellectual circles, and political programs the illusion of gaining satisfaction legally, of achieving their party goals and eliminating their enemies, both by legal means; such a constitution is no longer even possible today as a dilatory formal compromise; and, as a practical matter, it would end by destroying its own legality and legitimacy (Schmitt 2004, 94)

At first glance, this metaphorical call for the constitution to decide would seem to posit the constitution at the centre of political decision. Yet, as becomes clear upon further inspection, the constitution decides between a substantive order and an ‘illusion’. Opting for illusion, then, leads to a loss of the constitution’s legitimacy. Such a decision, thus, would become illegitimate by default – which displaces the liberal idea of legality as legitimacy. The constitutive decision is predetermined as a decision for the imposition of a substantive order, and against democratic procedures.

This order, then, is presumed to express a given substantive content, namely, the ‘substantive contents and forces of the German people’ (Schmitt 2012, 90; my translation). This predetermination shows that Legality and Legitimacy foreshadows Schmitt’s post-1933 concrete order thinking or, as Schmitt previously coined it with reference to French legal scholar Hauriou, ‘institutional thinking’. As Maus (1969, 119) points out, concrete order thinking obscures the irrationality of Schmitt’s decisionism with reference to apparently objective, rational necessities ‘derived’ from the German people. Decisionism and the maintenance of an ostensibly predetermined, given order thus coincide. This early actualization of concrete order thinking complicates the assumption of Schmitt’s post-1933 ‘turn’ and invites an interpretation of Schmitt’s work as continuously concerned with maintaining order rather than with exceptional political decisions. This is the aspect of Schmitt’s essay that is relevant for this article: The constitution is theorized as an institution for the perpetuation of an order that is assumed to be essentially given. Schmitt’s theory of militant democracy ante litteram, thus, is a substantive theory of institutional conservation.

Such an interpretation is supported by recent discussions of Schmitt’s institutional theory and its concern with preserving order which seems to overshadow the alleged ‘turn’ from decisionism to concrete order thinking. Meierhenrich’s (2017, 172; see also Meierhenrich and Simons 2017) reconstruction, for example, highlights that Schmitt’s thought is centrally occupied with an ‘unstinting search for institutional responses to the specter of disorder’. Moreover, Croce and Salvatore (2021, 14) highlight the non-exceptionalist, conservative thrust of Schmitt’s legal thought and its prime concern with ‘determining durable conditions of stability’ and with ‘identifying and eradicating through the legal practice the models of life that menace political stability and homogeneity’. The particular function of Legality and Legitimacy in Schmitt’s institutionalism is carved out by Meierhenrich (2017, 182, 187) who traces a ‘cumulative radicalization’ in Schmitt’s institutional preferences and treats the essay as the starting point of a ‘transition from pragmatist institutionalism to extremist institutionalism’ which left only ‘a short theoretical step to justifying Nazi dictatorship’. ⁹ Legality and Legitimacy begins to advocate unlimited executive power, rudimentarily constrained only by plebiscitary acclamation (McCormick 2004, 208–9).

This interpretation demonstrates Legality and Legitimacy’s concern with the maintenance of order, carried out by the institution of a virtually unconstrained executive which is accountable only in the moment of plebiscite. Its institutional recommendations, including the denunciation of the Weimar constitution’s procedural part in favour of a value-laden constitutional core, must be read in this light. A substantive institutional conservatism thus appears to be the earliest theoretical reference point of militant democratic theory.

Loewenstein and Popper – from substantive to procedural institutional conservatism

Five years after Legality and Legitimacy, two essays in The American Political Science Review by emigré political scientist Loewenstein introduce the concept of ‘militant democracy’ (Loewenstein 1937a, 1937b). Despite political enmity, continuities between Schmitt and Loewenstein’s theories are plain. Loewenstein (1937a, 424) is writing in fear that, once the existing order is subverted, ‘chaos reigns’. Against such chaos, Schupmann (2015, 224) highlights, Loewenstein’s and Schmitt’s ideas converge on denying ‘democratic change to core areas of the constitution’ for the sake of a ‘stable political order’. Loewenstein’s essay further echoes Schmitt’s polemic against Kelsen’s legal positivism and scolds ‘legalistic self-complacency and suicidal lethargy’, ‘democratic fundamentalism and legalistic blindness’, the ‘exaggerated formalism of the rule of law’ – and instead advocates ‘a better grasp of realities’ and the defence of democracy’s ‘absolute values’ (Loewenstein 1937a, 424, 431–2).

Yet, Loewenstein’s target is a different one. His theory of militant democracy constitutes an attempt of combating fascism on its own playing field, which further implies that it is an anti-fascist practice rather than a general, coherent normative justification for restricting the rights of anti-democrats (Schupmann 2015, 218; Hacke 2018, 251; Fuhrmann 2019, 93). This contextual restriction makes it imperative to briefly turn to Loewenstein’s conception of the fascist threat. Loewenstein holds that the ‘fascist technique’ ‘could be victorious only under the extraordinary conditions offered by democratic institutions’ such as ‘speech, press, assembly, and parliamentary participation’ (Loewenstein 1937a, 423–4, my emphasis). Fascists, on this view, seize power legally, preferably through representative institutions.

But how does fascism seize power, and why is democracy unable to defend itself against such seizure? Loewenstein operates on the basis of the pessimistic assumption that democracy, ‘a rational system’, ‘could not devise emotional formulas capable of competing with the fascist Pied Pipers’ (Loewenstein 1937a, 428). ¹⁰ It lacks the emotional sway that the fascist technique exerts over the masses, making Loewenstein (1937b, 657) fear that ‘liberal democracy […] is beginning to lose the day to the awakened masses’. Loewenstein, in the last analysis, considers democracy ‘too precious to be entrusted to the people’ (Maddox 2019, 492). Therefore, his theory seeks to contain the ‘emotional masses’ in a ‘disciplined’ or ‘authoritarian’ democracy which he defines as ‘the application of disciplined authority, by liberal-minded men, for the ultimate ends of liberal government: human dignity and freedom’ (Loewenstein 1937b, 657–8). While not reproducing the focus of Schmitt’s theory on personalized dictatorship (Hacke 2018, 251), Loewenstein’s (1937a, 432) militant democracy nevertheless ultimately relies on ‘concentration of powers in the hands of the [liberal] government’ and tasks it with restraining legal-procedural transformation via the ‘suspension of fundamental rights’.

But such continuities notwithstanding, Loewenstein deviates from Schmitt in theorizing democratic militancy as an institutional guarantee for democratic institutions. Militant democracy is justifiable only because of democracy’s institutional character. For Loewenstein’s militant democracy, the institutional character of democracy justifies the temporary suspension of any of its basic characteristics at the hands of the executive. Democratic ‘fundamental rights’ first need to be ‘institutionalized’ (Loewenstein 1937a, 432). Democratic ‘values’ merely attach to these institutions a higher dignity – they are emphatically not derived from the people, the ‘emotional masses’. Loewenstein’s institutional conservatism, thus, oscillates between a procedural appreciation of liberal democratic institutions and a substantive defence of democracy’s absolute values.

In a 1946 statement titled ‘Freedom is Unsafe Without Self-Government’, the implications of this ambiguity between a substantive and a procedural institutional conservatism are spelled out. Loewenstein (1946, 47) advocates the constitutional entrenchment of a political right ‘to participate in the government of one’s country’ with the end of securing ‘freedom’. This highlights the tension between a procedural and a substantive conservatism in his theory: The procedural idea of participation in government is explicitly stood on an equal constitutional footing with substantive individual liberties. Procedures are an end in themselves, but only for the reason that they stabilize other, substantive ends. The text specifies that, in practice, this implies the entrenchment of a given institutional form of democracy: Any government that does not satisfy the criteria of a representative organization, electoral institutions, political parties, ‘sufficiently’ general suffrage and secret, honest elections, is illegitimate (Loewenstein 1946, 48). In sum, Loewenstein’s institutional conservatism is thus characterized by the executive perpetuation of a particular, basic institutionalized form of democracy against fascist seizure of power, which is justified with a set of democratic institutions that is identified with substantive, fundamental values. ¹¹ Whereas Schmitt’s straightforwardly substantive institutional conservatism constructs varying institutions for securing a substantive order, Loewenstein’s rather ambiguous institutional conservatism advocates fixed institutional procedures for substantive reasons.

Popper’s contribution to the canon of militant democracy, his 1945 book The Open Society and Its Enemies, continues the process of ‘proceduralizing’ militant democracy’s institutional conservatism. Like Schmitt’s theory, the book does not explicitly advocate ‘militant democracy’, nor does it advocate repressive intervention in any detail. It is nevertheless received in militant democratic thought because it advances a ‘methodological’ defence of democracy. Democracy is conceptualized as an instrument for securing an ‘open society’. This open society is posited as the antonym of any predetermined, immutable ‘ultimate political aim’, ‘absolute and unchanging ideal’ or idealized ‘blueprint for society as a whole’ (Popper 2020, 148–9, 151) – a polemic directed at Plato, Hegel and Marx rather than at fascists agitators. Different from Loewenstein’s militant democracy, Popper’s open society is thus no longer restricted to anti-fascist interventions but characterized by an anti-totalitarian claim against closure in political thought.

The reason for the ‘public discrediting of the theory of historical inevitability or destiny’ (Agassi 2008, 219) in The Open Society is a reading of the utopian tradition as prescriptive, deterministic, irrational, therefore methodologically unsound and amenable to authoritarian abuse. Because of the utopian end point’s immutability, Popper argues, utopianism has historically justified oppressive, authoritarian, tyrannical practices (Popper 2020, 151). The open society as a ‘form of social life’, on the other hand is characterized by ‘values […] such as freedom, tolerance, justice, the citizen’s free pursuit of knowledge, his right to disseminate knowledge, his free choice of values and beliefs, and his pursuit of happiness’ (Popper 1963, 206).

The Open Society, then, introduces democracy as an instrumental category to support these aims by securing the minimal, negative precondition of avoiding tyranny (Popper 2020, 118). Popper conceives of democracy in exclusively instrumental, institutional terms. Democracy is ‘a set of institutions’, variously including ‘institutions of government, including those by which it is elected, the courts of justice, the civil service, public health, defense, and so on’ or, ‘a constitution, civil and criminal law, legislative and executive organs, such as the government and the rules by which it is elected’ (Popper 1963, 206; Popper cited in Rijpkema 2018, 75). With this institutional definition of democracy, Popper’s book advocates ‘the introduction of scientific method into politics’, and ‘the application of the critical and rational methods of science to the problems of an open society’ (Popper 2020, 153, xliv). This is not metaphorical methodological jargon: Popper’s rationalist, ‘epistemological defense of liberalism’ extrapolates the ‘open society’ and its rationality from the philosophy of science (Paden 2000, 409, 423; Ryan in Popper 2020, xix). Totalizing critical rationalism’s conception of progress and conceiving of democracy as its institutional form, Popper thus proceeds to theorize social transformation as incremental and reformist, and democracy as a passive exercise of rejecting hypotheses.

This is captured, firstly, by the concept of ‘piecemeal social engineering’, an institutionalized fallibilism utilizing ‘social experiments’ and enabling a gradual – as opposed to radical – ‘remodelling [of] society’, which Popper’s (2020, xliv, 152) theory presents as the only epistemologically sound form of transformation. That is to say that it is the only form of transformation that corresponds to Popper’s theory of knowledge. This implies a transformed institutional conservatism: The institutions themselves are to be constructed in such a way that gradual transformation is enabled but radical transformation becomes unlikely. As Müller has it, Popper’s is a ‘procedural conservatism’ with the intent to ‘carefully manage change’ (Müller 2008, 57). ¹² Biebricher (2020, 159–60), similarly, coins the concept of procedural conservatism as a model of politics characterized by ‘steady problem solving and ensuring that things do not run into the ground in a seemingly unhinged world’, a conservative vision that not only accepts change as unavoidable but in fact encourages piecemeal transformation – as opposed to the ‘substantive intuition of preservation and – if possible – locking in the status quo’.

Secondly, Popper’s (1963, 209) theory restricts democracy to a method of controlling government, explicated in a 1962 address to students at San Francisco State College:

the fact is that the people do not rule. […] The government rules. All the people can do is vote, they can vote a government in to power and, much more important, they can vote a government – or a party – out of power.

Because someone needs to conduct the social engineering, power is exercised by government officials (see Dryzek 1990, 6); the people periodically accept or reject their government much like a ‘hypothesis’. ¹³ Consistent with this, Popper emphasizes ‘especially’ general elections – they are ‘most important’ because they allow the bloodless ‘dismissal’ of government (Popper 2020, 118–9, 360, 368). This negative function is consciously separated from the idea of popular sovereignty: Democratic participation is conceived primarily as a passive anti-tyrannical instrument, but not as a means of popular self-determination. The people do not govern; they only provide checks on their (assumedly) rational, reformist government via the institution of election; or, in Popper’s (2020, 118) own words: ‘although “the people” may influence the actions of their rulers by the threat of dismissal, they never rule themselves in any concrete, practical sense’.

The trajectory of militant democracy after Schmitt, thus, is characterized at least partly by a transformation of institutional conservatism from a substantive to a procedural form. Schmitt’s theory devises the institutions of a value-based constitution and a virtually unrestrained executive for the defence of an ostensibly substantively given order. Loewenstein’s militant democracy picks up on the idea of an executive that restrains legal revolution but identifies a set of democratic institutions with the ‘values’ to be defended. Popper’s theory, finally, externalizes value commitments to an ‘open society’ and instrumentally advocates democracy as an incremental, anti-tyrannical methodology of government-centred, piecemeal transformation. In this actualization of institutional conservatism, democracy itself, as an institutional method, becomes a restraining force.

Kirshner and Rijpkema – procedural institutional conservatism

‘Neo-militant democracy’ (Malkopoulou and Norman 2018, 96) is characterized by an explicit opposition to the far-reaching implications that the lack of constraining justifications for envisioned interventions in traditional theories of militant democracy might have for democratic practice. Kirshner’s (2014) A Theory of Militant Democracy and Rijpkema’s (2018) Militant Democracy, the two most systematic proposals for neo-militant democracy, thus present themselves as more restrained than their forerunners, as ‘self-limiting’ (Kirshner 2014), ‘minimal’ (Ellian and Rijpkema 2018, 18) and geared to prevent ‘premature intervention’ (Rijpkema 2018, 169). Yet, they argue that minimal militant interventions which leave the democratic character of the defended entity intact are advisable given the conviction that the demise of democracy cannot be met with inaction by democrats.

Despite converging on restraint and minimalism, Kirshner and Rijpkema develop two distinct justifications for militant democracy. Rijpkema’s theory justifies militant interventions on the basis of a theory of democracy as, in the most minimal sense, characterized by self-correction. Kirshner’s ‘self-limiting’ theory of militant democracy, on the other hand, is based on the idea of a universal liberal democratic right to participation and thus justifies militant interventions in relational terms. Their strategies, in different ways, mirror and continue what I have called a transformation towards a procedural institutional conservatism. I will discuss them in turn.

Rijpkema’s ‘minimalist’ theory advocates for militant defences of ‘self-correction’ which is conceived as an inalienable democratic ‘essence’, a minimal baseline that can be defended in democratically justifiable ways. Drawing not only on Popper but also on his ‘popularizer’, philosopher and Labour politician Magee and on Dutch legal scholar van den Bergh, Rijpkema’s ideal conception of democracy closely resembles the methodological definition found in The Open Society:

A properly functioning democracy has a good deal in common with science (in its ideal form): policies are tested like hypotheses on reality, and correction can subsequently take place (Rijpkema 2018, 134–35)

This rational process of trial-and-error, according to Rijpkema, initiates progressive learning processes while ‘authoritarian thinking’, which lacks such recalibrations, is considered ‘irrational and unscientific’ (Rijpkema 2018, 72). Democracy, as is the case in The Open Society, thus figures as a rational mode of producing fallible results en route to social progress in Rijpkema’s theory. ¹⁴

This rationalist conception of democracy underpins the idea of self-correction, which is conceptualized as the continuous possibility to revoke – or ‘falsify’ – previously made decisions. Rijpkema (2018: 134) situates this approach ‘somewhere between procedural and substantive democracy’ insofar as self-correction remains tied to the procedural idea of majority decision but attains substantive value as an essential function of democracy. This essentiality justifies militant intervention once self-correction is threatened:

The justification for action therefore lies in erosion of the democracy’s self-corrective capacity. Only if parties threaten to damage this, can we legitimately call them antidemocratic parties and ban them. A democracy is not acting undemocratically or inconsistently in doing this: its essence, after all, lies in the permanent revocability of its decisions (Rijpkema 2018, 140).

This explication of the defensible entity approximates Popper’s methodological definition of democracy insofar as it only posits a ‘methodological’ strategy, self-correction, at the centre of democratic defence. Beyond this restriction, Rijpkema’s theory is agnostic to the concrete form that democracy takes: The ‘essence’ of democracy is only threatened when its capacity for revoking and renewing decisions is undermined.

Rijpkema’s theory thus echoes the procedural institutional conservatism of Popper’s theory insofar as it identifies democracy with an iterative process of social engineering. It crucially differs from Popper’s theory, however, insofar as democracy is conceived as valuable in itself and insofar as self-correction is conceptualized as synonymous with popular sovereignty. Democracy as self-correction, for Rijpkema (2018: 144), ‘means self-government by the people by means of self-correction’. This is an important difference: Because, for Rijpkema, popular sovereignty is inherent to self-correction, it is all the more important to inquire how self-correction actualizes. In institutional terms, Rijpkema (2018, 47), like Popper, opposes the executive ‘enforcement of a utopian ideal’ and, consistent with that anti-utopian reservation, does not offer a fixed blueprint for institutionalizing self-correction. Instead, Rijpkema’s (2018, 153–4) theory proposes ‘three principles of democracy as self-correction’: Evaluation, political competition and freedom of expression. This fairly minimal conception of democracy, according to Rijpkema (2018, 154), offers narrow criteria for judges deciding on party bans while offering much scope for political theorists, legal scholars and the legislative to do further definitional work in particular democratic societies.

Despite this institutional minimalism, however, Rijpkema’s book at least partially immunizes existing conceptions of democratic institutions against ‘self-correction’ and thereby hardens its institutionally conservative outlook. This may not necessarily be a logical consequence of his theory of democracy as self-correction: Self-correction does extend to institutions and may, in principle, be open to various institutional forms. Self-correction, in its most basic sense, is guaranteed as long as ‘there is an institutionalized moment [of evaluation] at which that which does not work can be rejected and new ideas can be embraced’ (Rijpkema 2018, 141, 145). But once the ‘principles’ are applied, received notions of democracy re-emerge:

From the perspective of democracy as self-correction, we might call the right to vote and elections the principle of evaluation. Parties that threaten this principle, by abolishing the right to vote or fundamentally changing the way elections are held, are eligible for a ban (Rijpkema 2018, 140–1)

To be sure, Rijpkema (2018, 141) resists the possibility to formulate ‘hard rules’ to establish when a non-trivial (fundamental) change to the principle of evaluation has occurred – although it is clear that ‘If the articles on the right to vote and elections are removed from a democratic constitution, it would no longer be democratic’. Ultimately, and even though Rijpkema insists that democracy as self-correction is not reconcilable with judicial practices of review that eclipse the legislature, it is for ‘the judge’ to decide (Rijpkema 2018, 141). But this is precisely the point at which even Rijpkema’s minimal conception of democracy restricts the popular will by taking a historical institutional choice as given: Democracy as self-correction occurs through elections, not through non-electoral democratic forms. Even if Rijpkema historicized this claim and insisted only on the contemporary relevance of elections, in the last analysis, the matter is referred to the judge. It is not upon the people to decide how they want to govern themselves. Judges limit this constitutive power.

The version of a procedural institutional conservatism found in Rijpkema’s theory, thus, is indeed more minimal: Neither a vast array of institutions nor an external substantive commitment justifies democratic defence. Only self-correction serves as a normative commitment. Rijpkema’s theory nevertheless remains conservative. It ‘assumes an already existing democracy’ and devises institutions that restrain ‘fundamental’ transformation of the institutions that are identified with its self-correcting essence (Rijpkema 2018, 194).

Kirshner’s ‘self-limiting’ theory of militant democracy markedly differs from Rijpkema’s insofar as it neither seeks to offer a baseline definition of democracy nor assumes democracy to already exist. This reluctance to substantively predetermine democracy, however, similar to Rijpkema, stems from an anti-utopianism that resonates with Popper’s The Open Society. Referencing the theory of self-limiting revolution, developed by Solidarność thinker Michnik, Kirshner (2014, 17, 47) emphasizes that theories of militant measures cannot presume ‘a perfect moral community or any other idealized status quo’ and alerts to the essential ‘fallibility’ of democratic revolutionaries. Therefore, his theory cautions against the pursuit of any ‘political utopia’ of democratic uniformity in the name of democratic militancy (Kirshner 2014, 27). Instead, Kirshner’s theory operates under the assumption that democratic defence departs from ‘actually existing and substantially imperfect representative regimes or polyarchies’ (Kirshner 2014, 47). Polyarchy is defined with reference to Cohen and Rogers’ summary of Dahl as characterized by suffrage, political expression, association and office-holding, and access to diverse information sources for ‘virtually all citizens’; officials controlling public policy are elected freely and fairly (Kirshner 2014, 4–5).

Militant democrats, according to Kirshner, should embrace polyarchy as a ‘second best ideal’ rather than using the militant democratic strategy of exclusion for pursuing a democratic utopia (Kirshner 2014, 48). ¹⁵ This ‘intentionally conservative’ conclusion is derived from the moral assumption that ‘democrats’ and ‘anti-democrats’ enjoy an equal right to participation (Kirshner 2014, 24, 59). This right, according to Kirshner, derives from a presumed general interest in political participation under the assumption that ‘a person’s interests deserve the same moral consideration as any other person’s’ (Kirshner 2014, 34). Elevating participatory rights to a delimiting a priori of militant democracy, Kirshner thus postulates that militant democracy relies on an equilibrium between the participatory rights of democrats and anti-democrats. Inherent to this idea of equilibrium, his theory advocates the restrained use of militant measures. Rather than functioning as a means of effectively ousting presumed anti-democrats from the demos altogether, militant interventions are restricted to situations in which participatory rights are in peril: ‘The core justification […] revolves around whether the groups in question have violated the political rights of others’ (Kirshner 2014, 109). Militant democracies can suspend democratic rights only if this has already occurred or can reasonably be expected. And if such a violation has occurred, they still have to balance the ‘serious moral costs’ (Kirshner 2014, 111) of this action against the cost of inaction, with militant interventions figuring only as a means of last resort. This constitutes the principle of self-limitation: Militant interventions are only justified as long as they result in an optimized participatory rights equilibrium compared to the expected result of inaction. If no serious threat to the status quo of participation can be detected, militant interventions are not justified – democrats, thus, cannot use militant repression to further their goal of perfecting democracy.

This is a crucial limitation: The participation that is defended is participation in an existing polyarchy. Kirshner’s theory of militant democracy operates in unidirectionally conservative ways: Participation as it is shall not be undermined. Under status quo conditions, polyarchy is assumed to be the institutional channel through which democrats and anti-democrats alike can pursue their goals (Kirshner 2014, 47–48). Militant democrats should only intervene once either group’s participation in polyarchy is obstructed. Using the politically contested concept of polyarchy to justify the defence of democracy implies that challenges to participation in ‘the institutional arrangements that have come to be regarded as a kind of imperfect approximation’ (Dahl 1971, 9) of democracy move to the centre of democratic defence. ¹⁶ The assumption of an existing polyarchy implies at least a provisional restriction of democratic defence to a specific set of status quo institutions. This does not mean that Kirshner’s theory is necessarily wedded to polyarchy: It could serve the defence of any regime under which ‘participation’ of any sort occurs. It does mean, however, that Kirshner’s theory is wedded to whichever organization of participation characterizes the status quo: A radical transformation of the existing institutional structure would likely imply a categorical shift in the realization of participation. A militant democrat, then, could claim that such transformation strips individuals of the particular actualization of their participatory right in the previous non-ideal regime and is thus illegitimate.

In concluding the reconstruction of militant democracy’s transformed institutional conservatism, both Rijpkema’s and Kirshner’s theories can be placed in a historical development towards proceduralization. Neither of the authors assumes that democratic defence signifies the unequivocal conservation of the status quo in the trivial sense of ‘all that which merely is’. Quite the contrary, their theories appear to encourage or at least emphatically enable transformation without fear of repression. Nevertheless, their theories retain an institutionally conservative thrust insofar as they ultimately contain that transformation within existing institutional structures that serve as the boundary of transformative practices.

Why contemporary militant democracy’s institutional conservatism is problematic

Kirshner and Rijpkema’s recent retheorizations of militant democracy, the above discussion has shown, fall in line with traditional militant democratic thought in at least one respect: They continue a tendency to ‘proceduralize’ the institutionally conservative strategy of militant democracy. Distinct from both the substantive conservatism of Schmitt and the decidedly anti-fascist democratic militancy of Loewenstein, their institutional conservatism echoes Popper’s idea of institutionally encouraging but nevertheless restraining democratic transformation. Their institutional conservatism does not imply a mere perpetuation of the status quo: Their defensible entities are institutional channels for its reform. Moreover, the particular institutional forms that democracy takes are not unjustifiably presumed. It would be logically possible to apply either theory under changed conditions: Self-correction could occur in communal assemblies rather than general elections; and participatory rights could be realized under conditions of sortition. This does not mean that contemporary militant democrats do without certain institutional preferences. Critiquing such latent convictions, however, would amount to a straitjacketing of political theory.

It is, rather, the indeterminate minimalism itself which produces a normative problem for contemporary theories of democratic defence. Because they assume only a minimal institutional baseline, an imperfect approximation of their ideal conceptions of democracy, change towards better approximation of democracy should be allowed for. And to an extent, that seems to be the case: Self-correction ensures that possibilities for change reoccur; a defence of participation in a given participatory system ensures the ‘ability to engage in meaningful political activity’ (Kirshner 2014, 39). But because the de facto demarcation of the defensible entity is cast out from militant democratic theories, the boundary of transformative practice is somewhat paradoxically set by an undetermined status quo. It nevertheless functions as a boundary – with the very real threat of repression lurking underneath. Drawing on the example of democratic sortition, I will first outline how this boundary reappears in the theories of Kirshner and Rijpkema; then I will briefly highlight what kind of limitation on democratic transformation it implies.

Consider the indeterminacy of Kirshner’s core concept of ‘participation’ (see Invernizzi Accetti and Zuckerman 2017, 187–8). Participation can occur in various institutional or, more broadly, social structures. Kirshner would likely agree that in a polyarchical system, participation is unevenly distributed. It is possible to imagine that under such conditions, a radical reshuffling of the existing distribution of participation might be desirable from a democratic perspective. Such a reshuffling, however, might contest basic pillars of polyarchy. This would be the case if participation would be reconceived in categorically different ways which might nevertheless be democratic. A hypothetical political organization aiming for and capable of effecting such radical transformation, however, would clearly pose a threat to the participatory rights instituted under conditions of a status quo polyarchy, as the future form of participation categorically deviates from the previously instituted one.

This abstract scenario can be illustrated with the example of a hypothetical political party that pursues non-electoral forms of democratic organization such as democratic sortition and thereby challenges polyarchy’s virtually universal suffrage under conditions of economic inequality. Proponents of non-electoral democracy might argue that sortition better institutes participatory rights than polyarchy’s political rights do (recently e.g. Landemore 2020, 134–8). Whether that would in fact be the case is subject to theoretical and political disagreement – Kirshner (2019, 68–9) does not believe that sortition would resolve the democratic paradox, but it ‘might improve on the status quo’ of participation. But what is more important is the possibility that a response constructed on the basis of Kirshner’s theory (not a response by Kirshner, who normatively considers sortition a potential improvement) might justify the repression of such a party with reference to a defence of status quo institutions. These institutions guarantee participation, if in a particular, non-ideal and unequal way, and therefore already contain the means for their reform and bettering. A radical break with these institutions, on the other hand, implies uncertainty as to how participation would be organized and how this new form of participation would be distributed. Because existing institutions are emphatically conceived as valuable and because Kirshner’s theory lacks conceptual resources for balancing different forms of participation, a response on its basis would likely err on the side of caution and thereby restrain potentially democratic institutional transformation that points beyond the existing arrangement of participation. ¹⁷

This problem can be illustrated even more clearly with Rijpkema’s theory, which offers a more explicit threshold for intervention. Recall its assertion that parties seeking to abolish or fundamentally change voting procedures may justifiably be banned. Where would that leave proponents of non-electoral democracy? Of course, the question of what constitutes a ‘fundamental change’ is subject to debate. Installing a wholly different institutionalization of democracy, however, would likely qualify. Even a generous interpretation of Rijpkema’s theory, which would abstract from its explicit claim that elections in particular constitute the boundary of transformation, points to this problem. Political organizations that seek to ‘fundamentally’ alter the institutional ordering of democracy are eligible for a ban. Returning to the example of a hypothetical political party that pursues non-electoral forms of democratic organization such as democratic sortition, it might be assumed that abolishing elections and replacing the existing form of parliamentary representation would constitute a fundamental change. The response to such a party, as discussed above, is then left to the judge who operates exclusively with the criterion of ‘fundamental change’ and therefore lacks the conceptual tools to differentiate between democratic and undemocratic radical transformation. The constitutive people, thus, loses its power to restructure democratic politics in thoroughgoing, and potentially democratic, ways.

Why, then, is this democratically problematic? Militant democratic theories, of course, explicitly seek to set boundaries to democratic practice. But importantly, contemporary militant democratic theories seek to justify the restriction of democratic practice in ways that are consistently democratic. Their theories are neither justified by a substantive conservatism, as is the case with Schmitt’s Legality and Legitimacy, nor do they claim a distinctly anti-fascist motivation which carries Loewenstein’s militant democracy. Instead, their anti-totalitarian theories merely claim that in order to defend democracy, certain limits on transformation are required. Yet, these limits are difficult to sustain without removing a crucial resource of power from the democratic sovereign. Even though contemporary militant democratic theories only defend democratic procedures, they constrain popular sovereignty with legal sovereignty: The existing legal order itself limits its transformation; this ultimately implies that (institutional) innovations which are external to the existing legal order are inhibited (on the distinction between popular and legal sovereignty, see e.g. Maus 2015, 40–1). Rather than exclusively obstructing the abolishment of democracy, militant democracy’s procedural variant of institutional conservatism, despite its limited openness to transformative participation, might thus restrain potential improvements via democratic innovation. Doing so, it is not consistently procedural. Instead, it ties procedures to the conservative intention of restraining change to the existing order.

To be sure, this would not be a relevant immanent criticism if the authors of contemporary militant democratic theories assumed the existence, or at least a fixed knowledge, of the institutions that would be characteristic of some kind of ideal democratic society. But the insight that this is not the case, in fact, constitutes one of their most crucial improvements over traditional militant democratic theories which do assume such substantial knowledge. Given the insight that democratic institutions are pervasively non-ideal and asymmetrical, that they reproduce structural inequalities, restraining their transformation in institutionally conservative ways seems difficult to justify.

Conclusion

The above reconstruction of militant democratic theories’ institutional conservatism demonstrates that militant democracy’s persistent theoretical preference for institutions that restrain the transformation of the existing order is significantly transformed throughout the idea’s trajectory. I have reconstructed this trajectory as a process of proceduralization: Contemporary militant democratic theories, unlike the substantive institutional conservatism of Schmitt and the ambiguous institutional conservatism of Loewenstein, reproduce the anti-utopian and anti-totalitarian position developed by Popper to justify restrictions on democratic practice without substantive commitments to a fixed status quo.

But, as I have argued, this procedural institutional conservatism produces unique normative problems of its own. If substantial commitments are cast out from militant democratic theories, their restrictions on democratic practice target transformative demands that radically point beyond whichever institutional ordering of democracy happens to characterize the status quo. Amendments to that institutional ordering can only happen in a gradual, contained manner without changing the existing order in fundamental ways. This, I have argued, is particularly problematic under conditions of non-ideal institutions and potentially democratic challenges that point beyond them.

Within the scope of this article, it is not possible to offer a full-fledged alternative to the procedural institutional conservatism of contemporary militant democracy. In concluding, however, I would like to emphasize that an alternative might take on the task of defending democracy while emphasizing, rather than restraining, the democratic capacity for even radical transformation. Such an alternative might be useful particularly under circumstances in which an alienation from existing democratic institutions might drive challenges to them. Their rigorous critique, one might hope, could lead to democratic innovation and institutional reappropriation. For that to happen, however, the route to radical change cannot be categorically foreclosed.