1 Introduction

An increasing number of modern constitutions engage in strong constitutional entrenchment—they enshrine material limits to amendment, in the form of so-called ‘eternity-clauses’ that exempt certain constitutional provisions from the scope of the constitution’s procedure of amendment, thus making them immune from any change under the existing constitution. Where there are material limits to constitutional amendment, these are frequently enforced by supreme or constitutional courts that claim the power to strike down procedurally valid constitutional amendments for violation of the material limits of amendment (Gözler 2008; Roznai 2013a, 2014).

Eternity clauses, as well as provisions for their judicial enforcement, one might argue are particularly clear examples of what Ran Hirschl has described as hegemonic preservation (Hirschl 2007). According to Hirschl’s hegemonic preservation thesis, entrenched constitutions protected by judicial review are often introduced by social or political elites who exercise disproportionate and democratically illegitimate influence over the process of constitution-making, and who use that influence to protect their own political preferences, while they know that these preferences are unlikely to survive open democratic contestation in the future. What better way could there be to secure the permanent ascendancy of one’s own constitutional vision than to make certain parts of a constitution altogether unamendable?

The constitutional history of the Republic of Turkey is commonly regarded as a textbook example of hegemonic preservation by way of strong constitutional entrenchment (Özbudun 2011a, pp. 122–129; Hirschl 2012). There have been three constitutions in operation in Turkey since the proclamation of the Republic: the constitutions of 1924, 1961 and 1982. The practice of strong constitutional entrenchment was present from the beginning of the Turkish Republic, and it has become more pronounced with each successive constitution. As new constitutions were made, more eternity clauses were added, and these clauses have come to be judicially enforced by the Turkish Constitutional Court (Anayasa Mahkemesi—AYM). Though Turkey can now look back on a considerable history of democratic multi-party politics, it is faring worse in securing its democratic consolidation than the third wave democracies of South Eastern, Central and Eastern Europe. The practice of strong constitutional entrenchment and of amendment review has often been held to be at least partly responsible for this lack of democratic consolidation (Arslan 2002, 2007; Belge 2006; Gülener 2012; Günter 2012; Koğacıoğlu 2003, 2004; Özbudun 2005, 2011a, 2011b, 2012;Roznai and Yolcu 2012; Tezcür 2009; contra Arato 2016, chapter 5; Arato 2010; Caniklioğlu 2010).

This paper will offer a critical rereading of the history of strong entrenchment and amendment review in Turkey. We believe that Turkey does not fit Hirschl’s model of hegemonic preservation. Attempts to impose that model on the constitutional history of the Turkish Republic offer a one-sided and incomplete diagnosis of what ails Turkish constitutionalism. It is not the excessive strength of constitutionalism in Turkey but rather its normative weakness that is responsible, at least in part, for Turkey’s repeated constitutional crises. In our concluding section, we will suggest an alternative template for understanding Turkish constitutional history—the theory of sovereignty as the power to decide on the exception put forward by Carl Schmitt (Schmitt 1922).

2 Hirschl’s Model of Hegemonic Preservation

In order to assess whether the use of strong constitutional entrenchment and amendment review in Turkey should be described as an example of hegemonic preservation, it is necessary to give a brief outline of Hirschl’s model of hegemonic preservation (Hirschl 2007). Hirschl’s hegemonic preservation thesis has both a descriptive and a normative component. On a descriptive level, the thesis aims to offer an explanation for the global rise, in recent decades, of constitutionalism and judicial review. That explanation, what is more, is meant to cast a critical light on that process, to question whether we should welcome it from a normative perspective. One must guard against the view, however, that all exercises of hegemonic political power through the instrumentality of the courts qualify as instances of hegemonic preservation in Hirschl’s sense. Both the normative and the descriptive components of Hirschl’s model impose constraints on its applicability to particular constitutional traditions. The task of the present section is to outline these constraints, so as to lay the ground for our claim that the hegemonic preservation thesis fails to fit the case of Turkey.

2.1 Hirschl on the Genesis of Strong Constitutionalism

On the descriptive level, Hirschl’s hegemonic preservation thesis claims that the establishment of a strong form of constitutionalism—one characterized by the power of judicial institutions to strike down legislation for lack of conformity to an entrenched bill of rights—takes place as a result of a strategic interplay between political incumbents, judicial elites, and economic elites (Hirschl 2007, pp. 11–12, 38–49). Political incumbents may find themselves in a situation where they have to fear that they will lose power in upcoming elections, as a result of structural changes in the ideological orientation or the social composition of the electorate that make it unlikely that the incumbent legislative coalition will be re-elected at any time in the foreseeable future. In that situation, it may be advantageous for political incumbents, the current hegemons, to introduce constitutionalism and judicial review so as to shield their own preferences from being overturned by future legislative majorities, even though such a shift disempowers the political elite and binds it to the decisions of the judiciary. If the judiciary can be counted upon to use its powers in ways that are broadly in line with the ideological orientation of the current incumbents, the introduction of strong constitutionalism—of what Hirschl also refers to as ‘juristocracy’—is preferable to domination by future incumbents that differ in ideological orientation. Incumbents that introduce constitutionalism, according to Hirschl, can often count on the judiciary’s willingness to decide in line with their own preferences since the judiciary typically hews to a neo-liberal elite-consensus that is shared by economic elites, the higher bureaucracy, as well as the political incumbents that introduce constitutionalism.

2.2 Hirschl’s Normative Criticism of Strong Constitutionalism

Hirschl’s descriptive account of the causal origins of constitutionalism, needless to say, is meant to suggest a normative criticism of the global trend towards juristocracy. If the package of constitutionalism and judicial review is typically introduced in bad faith, for motives other than the ostensible and laudable motive of protecting individual rights and social justice, we should expect its consequences to be normatively problematic. Hirschl discusses two different ways in which constitutionalism turns out to be normatively problematic.

The more prominent strand of normative argument in Hirschl takes off from a Dworkinian starting-point. According to Dworkin, the normative consequences of constitutionalism must be evaluated from an output-oriented perspective: Constitutionalism is legitimate if and only if it is more likely than relevant alternatives to create decisional outcomes that secure distributive justice. If constitutionalism indeed achieves this aim we should not be worried by the fact that decisions taken by constitutional courts are often counter-majoritarian (Dworkin 1997, pp. 1–38). Hirschl concurs that this is the right way to assess the legitimacy of constitutionalism (Hirschl 2007, p. 3). But he holds that a comparative analysis of judicial decision-taking in Israel, Canada, New Zealand, and South Africa shows that constitutionalism typically fails to advance the cause of distributive justice, due to the judiciary’s tendency to favor a Lockean interpretation of constitutional rights as rights of mere non-interference that expresses a neo-liberal bias (Hirschl 2007, pp. 100–168).

A second, decidedly less prominent strand of normative argument intimated in Hirschl’s analysis is closer to the position that is often referred to as ‘political constitutionalism’ and that is associated with authors such as Jeremy Waldron or Richard Bellamy (Hirschl 2007, pp. 186–190; Waldron 2006; Bellamy 2007; Schwartzberg 2007). Here the claim is that constitutionalism is inherently problematic because it necessarily disempowers the people (or their elected representatives) and thus violates the principle of democratic equality in the process of political decision-taking. As Hirschl himself puts the point, strong constitutionalism may ‘undermine the very essence of democratic politics as an enterprise involving a relatively open […] and accountable deliberation by elected representatives’ (Hirschl 2007, p. 187).

Note that these two normative criticisms of constitutionalism stand in tension with one another. The claim that the legitimacy of constitutionalism is to be evaluated on the basis of the substantive quality of decisional output implies that constitutionalism would be legitimate if courts empowered to strike down democratic legislation could be counted upon to take the right decisions—despite the fact that courts are not usually democratically accountable and even if courts decide in counter-majoritarian ways. If, on the other hand, constitutionalism is problematic for the reason that it necessarily violates the conditions of a fair democratic process, as political constitutionalists claim, then it must be problematic irrespective of the substantive quality of the outcomes that it tends to produce. Even the decisions of a court that was deeply concerned to further the cause of egalitarian distributive justice would, presumably, have to be regarded as illegitimate. The very existence of a judicial institution that could potentially overrule the people is, from this perspective, inherently illegitimate. We take it that what makes Hirschl’s attack on constitutionalism distinctive and interesting is that its emphasis is not on this purely conceptual argument but on the empirical claim that constitutionalism in practice tends to serve hegemonic interests. Our analysis will therefore focus on Hirschl’s output-oriented challenge to constitutionalism.

2.3 The output constraint

Hirschl’s preferred way to establish the empirical claim that constitutionalism is typically introduced with the motive of hegemonic preservation is to show that actual judicial decision-taking, after the introduction of constitutionalism, tends to implement hegemonic interests and then to infer the motive of hegemonic preservation from the observed decisional pattern. In line with this argumentative strategy, Hirschl offers a detailed analysis of the substantive normative quality of the actual outcomes of judicial decision-taking in Israel, Canada, New Zealand and South Africa after the introduction of strong constitutionalism with judicial review (Hirschl 2007, chapters 3–5). Clearly, there would not be much of a point in analyzing the actual decisions of courts in different jurisdictions, to show that they are engaged in hegemonic preservation, if any conceivable instance of counter-majoritarian judicial decision-taking was, by definition, to be regarded as inherently illegitimate, as political constitutionalists claim. An output-oriented challenge to strong constitutionalism presupposes that we can establish an empirical connection between the introduction of constitutionalism and the substantive deficiency of decisional outcomes.

It follows that one cannot show that some particular system of constitutionalism is an instance of hegemonic preservation simply by pointing to the fact that the way in which it distributes decisional authority could give rise to counter-majoritarian decisional outcomes. Neither is the claim sustainable simply by pointing to actual instances of counter-majoritarian decisions. To sustain the claim that a system of constitutionalism was introduced for the purpose of hegemonic preservation, and that it is therefore illegitimate, one must not merely show that the outcomes produced by the system in question are frequently counter-majoritarian. One must also show that they tend to be deficient from a substantive normative point of view—i.e. that they tend to inhibit the realization of social justice because they unduly favor a partial interest or perhaps that they often tamper with democratic procedures in a way that unfairly advantages a particular political camp. Without this empirical link between constitutionalism and substantive normative deficiency of decisional output, there is no basis for the inference to the supposed motive of hegemonic preservation and for the normative condemnation implicit in talk of ‘hegemonic preservation’. We call this requirement the ‘output constraint’ on the applicability of the model of hegemonic preservation.

2.4 The Baseline Constraint

Note as well that a normative assessment of the legitimacy of constitutionalism—such as is clearly implied by talk of ‘hegemonic preservation’—must specify a baseline of comparison. Hirschl does not explicitly address this issue, and he appears to imply, at times, that his model is a perfectly general model of the causes and the normative problems of transitions to constitutionalism (Hirschl 2007, p. 42). On closer inspection, though, it is clear enough that Hirschl’s argument appeals to an implied baseline. All the cases analyzed by Hirschl are cases where there was a transition from something like a Westminster-model of parliamentary democracy to a more strongly constitutionalist framework. With the exception of South Africa, all cases were firmly established democracies securely committed to the rule of law before the transition, and even South Africa was a democracy for whites before the end of Apartheid (Hirschl 2007, p. 9).

The claim that constitutionalism is to be regarded as deeply problematic for the reason that it fails to be conducive to a fuller realization of distributive justice will suffice to delegitimize constitutionalism only where there is an alternative that does better on the distributive score, and not significantly worse with regard to the protection of non-economic liberal rights—such as rights to privacy or habeas corpus rights—which Hirschl agrees are important (Hirschl 2007, pp. 108–125, 148). That alternative, for Hirschl, or so it would appear, is a classical Westminster-style parliamentary democracy—one that is not constitutionally committed to an excessive focus on the protection of private property, and thus, supposedly, more likely to be hospitable to the defense of a substantive welfare-state, but that still shows an adequate measure of respect for rights of personal liberty and for the rule of law. Hirschl’s discussion, at any rate, invites the reader to ask what decisional outcomes would have occurred in Israel, Canada, South Africa, or New Zealand without the introduction of judicial review based on a bill of rights. The claim that juristocracy is to be rejected because it tends to prevent the realization of distributive justice thus implicitly assumes that a rights-respecting and pluralist parliamentary democracy is available as a fallback option (Hirschl 2007, p. 38).

To be sure, the dependence on this implicit baseline does not invalidate Hirschl’s criticism of constitutionalism. But it limits the scope of that criticism. It is unclear whether the model of hegemonic preservation will tell us much about the legitimacy of an empowerment of the judiciary in cases where a rights-respecting and pluralist Westminster-style democracy is not available as an alternative to formal constitutionalism. We therefore speak of a ‘baseline constraint’ on the applicability of the model of hegemonic preservation.

2.5 The Agency Constraint

The descriptive component of the model of hegemonic preservation postulates a principal-agent relationship between political incumbents (as well as the economic elite) and the judiciary. But it is crucial to the model that this is a loose relationship, and that the transfer of power from the political elite to the judiciary is a real transfer, not a mere facade that is meant to veil continuing direct control by the current political incumbents. As Hirschl emphasizes, incumbents have no incentive to constitutionalize unless their political power is on the wane (Hirschl 2007, pp. 43–49). A powerful political incumbent that expects to remain powerful will not trade the prospect of continued direct control for juristocracy. Incumbents that constitutionalize, according to Hirschl, fear future impotence in the arena of democratic politics and therefore aim to preserve their influence—or at least the influence of their ideology—by transferring power to courts.

Such a stratagem can work only if courts are perceived, whether rightly or wrongly, as impartial and non-partisan institutions committed to enforcing the law, as Hirschl acknowledges (Hirschl 2007, p. 44). And this perception of courts is incompatible with any form of very direct and visible control of the courts by the incumbents that constitutionalize or by any other political agent. Successful hegemonic preservation requires that there be courts that can draw on strong resources of (perceived) legitimacy and that are backed up by a constitutional tradition that reliably protects their institutional independence. If the agent-principal relationship between political incumbents and courts was based on more direct control, new political incumbents could simply appropriate the role of principal once they take power, and hegemonic preservation through constitutionalism would fail. We will refer to this limit on the applicability of the model of hegemonic preservation as the ‘agency constraint’.

2.6 The Constraints and Turkey

The claim that some constitutional system is an instance of hegemonic preservation has strong critical import. It is meant to imply that the system in question tends to create seriously unjust decisional outcomes and that it came into existence precisely because powerful groups were interested in those outcomes. The applicability of this normative criticism, as well as the genetic claim about the origins of constitutionalism to which it is connected, will be undercut if a case violates the output constraint. What is more, Hirschl’s model is meant to suggest that a society that is said to be subject to hegemonic preservation would have been better off without introducing constitutionalism. This critical implication will fail to hold, even if the output constraint is satisfied, if the case in question violates the baseline constraint. A case might, finally, violate the agency constraint. The decisions taken by courts might express hegemonic interests, but the courts may nevertheless lack true agency, in virtue of a lack of institutional independence. In cases like this, it would make little sense to speak of ‘juristocracy’, to explain hegemonic political control as a consequence of excessive constitutionalism, and to expect that it will disappear as soon as constitutions are more flexible and courts less powerful.

We submit that true cases of hegemonic preservation must satisfy all three constraints. The claim that some constitutional system is an instance of hegemonic preservation would otherwise either be descriptively misleading or its intended normative implications would not be justified. In what follows, we will argue that it is doubtful whether Turkey ever satisfied all three of the constraints on the applicability of the model of hegemonic preservation. To establish this claim, we now turn to a historical analysis of strong constitutional entrenchment and amendment review in the Turkish context.

3 The Constitution of 1924: The Genesis of Strong Constitutional Entrenchment

The practice of strong constitutional entrenchment characterized the constitutionalism of the Turkish Republic from its beginning. The first constitution, of 1924, declared the republican form of state to be irrevocable. This eternity clause, as in France, was not intended to provide a basis for juristocracy. The framers of the constitution, influenced by Rousseau’s understanding of democracy, took parliament to represent the unity of the nation’s general will (Özbudun 2000, p. 35). In line with this implicitly anti-pluralist conception of democracy, the constitution of 1924, like almost all European constitutions of the time, did not create a constitutional court. Rather, it placed the sovereignty of the people in the hands of a legislative assembly.

3.1 The Single Party Era

The intention to create a parliamentary democracy that was to represent the political unity of the people stood in tension with the wider project of nation-building and social modernization that Mustafa Kemal, who had led the revolution against the old Ottoman order, was above all concerned to carry out. The guiding philosophy of this project has come to be referred to as ‘Kemalism’. Kemalism is best described as ‘a set of attitudes and opinions’ with a short and a long-term vision and mission, rather than as a rigid ideological system (Zürcher 2004, p. 181; Alaranta 2014). While the short term goal was the establishment of a secular Republic and a modern, ethnically homogeneous nation state, the long-term goal of Kemalism was ‘to elevate the people to the level of contemporary (Western) civilizations’ (Heper 1985, p. 50). These two goals inspired a series of reforms that were administered from above (Anderson 2009, pp. 414–417; Bali 2012, p. 270; Dunn 1989, pp. 173–198)—at times against staunch popular resistance. These included the abolition of the sultanate in 1922, the declaration of the Republic in 1923, the abolition of the caliphate in 1924, the adoption of republican, secular rule, the introduction of the Latin script, of western attire, and the modernization of the legal system (Ahmad 1993, Karadut 2012; Lewis 2001; Yılmaz 2013; Zürcher 2004, pp. 166–205).

The Turkish modernization project did not grow organically out of society. Hence, the so-called ‘state elites’—the same cadres who initiated and carried through the War of Independence, declared the Republic, and created the new civil bureaucracy—often found themselves at odds with powerful societal actors whose values and interests were threatened by the modernizing mission. The aim to create and protect a strong state that would be able to drive the project of modernization forward, even against resistance, therefore guided much of the later course of the constitutional and political history of the Republic. The principles of the indivisible integrity of the state and of secularism, in particular, were considered as the fundamental pillars of the new nation state that emerged out of the multi-ethnic and multi-religious Ottoman Empire. They were intended to help create a unified political identity which the state-elites regarded as vital for a successful project of nation building.

The Kemalist founders of the Turkish Republic, to enhance the strength of the state, and thus to secure the success of their project of modernization, aimed for a total convergence in political outlook between the legislative and the executive branches of the new government. Mustafa Kemal himself, through a careful vetting of candidates, saw to it that the elections to the second legislative assembly of the Turkish Republic brought forth a parliament that could be counted on to support the Kemalist vision of society (Özbudun and Gençkaya 2009; Tuncay 1981). It was this second legislative assembly, which was almost exclusively composed of members of the CHP, the party founded by Mustafa Kemal himself, that enacted the 1924 constitution (Bali 2012, p. 255).

The convergence between the legislative and the executive, or between the CHP and the state, deepened throughout the single party era (1925–45). The CHP-dominated parliament, in response to local insurrections and the discovery of a plot on the life of Mustafa Kemal in 1925–1926, enacted emergency regulations that prohibited opposition parties. The resulting unopposed control of the CHP facilitated the implementation of the Kemalist project, which sometimes necessitated doing things for the people, despite the people. Yet, it was also responsible for a widening rift between the state and the society, (Mardin 1973, p. 169; Çınar and Sayın 2014, p. 367). Unconstrained by effective parliamentary control, the Kemalist state built a strong military and civil bureaucracy that stood ready to ‘defend the civilizing mission against potential threats from society’ (Shambayati 2008, p. 288; see also Çınar and Sayın 2014, p. 368). The Kemalist elites believed in their own ability and their right to govern the people without their consent, until the latter had reached a level of civilizational maturity that would allow for the introduction of a multi-party democracy compatible with the Kemalist project of modernization.

3.2 The Multi Party Era

A transition to multi-party politics eventually took place in 1945. However, it soon turned out that large parts of society had not yet come around to the Kemalist vision. The opening of political competition was followed by the rapid ascendancy of the Democrat Party (Demokrat Parti—DP), which won a parliamentary majority in 1950. The DP was both more socially conservative and more economically liberal than the CHP. It increasingly channeled popular discontent with the Kemalist project. Unsurprisingly, the DP soon found itself in conflict with the Kemalists in the military, the civil bureaucracy, and the CHP (Çınar and Sayın 2014, p. 368). Nevertheless, it continued to be successful at the ballot box throughout the 1950’s, in part by adopting an increasingly populist discourse. In the face of challenge by the CHP, the DP came to endorse a purely majoritarian conception of democracy that identified democratic legitimacy with success at the ballot box and that rejected constitutional restrictions on the powers of a democratic majority. Ironically, the DP benefited from the 1924 constitution, as the latter, as we have seen, concentrated power in the hands of parliamentary majorities (Çınar and Sayın 2014, p. 369).

Multi-party politics, thus, instead of eradicating the cleavage between state and society, led to its transposition into a division between the Kemalist-controlled bureaucracy and military on the one side and the elected branches of government on the other (Heper 1985; Özbudun 1996; Bali 2012, pp. 263–79; Shambayati in Arjomand 2007, pp. 99–106; Shambayati and Sütçü 2012, p. 109). The growing alienation between the DP and the CHP, in the later 1950s, was a clear indication of a swelling hostility between the political elites leading the DP and the state elites. In the face of a worsening economic crisis, public unrest, and an increasing tendency on the part of the DP to take resort to authoritarian measures against the opposition one of the discontented institutions of the state establishment eventually took matters into its own hands. Democratic politics was suspended in the aftermath of a military coup in 1960.

4 The Constitution of 1961: The Emergence of the Constitutional Court

The 1961 constitution was prepared by a bicameral constituent assembly convened by the military. A ‘National Unity Committee’ composed of members of the military constituted one of the chambers of the assembly. The military, thus, did not merely convene the constituent assembly, but was also directly involved in the constitution-drafting process. The other chamber was a ‘House of Representatives’ dominated by non-military state elites. It included high-ranking bureaucrats and university professors known for their commitment to the CHP, as well as representatives of the CHP itself. Supporters of the DP were not allowed to participate at all. As a result, the draft constitution, which was prepared by a Constitutional Council composed of twenty selected members of the House of Representatives, closely conformed to the constitutional theses of the CHP. It came into force after a referendum in which it was accepted by 61.7% and rejected by 38.3% of votes cast, a result that signaled that a large proportion of the electorate did not approve of the coup (Özbudun and Gençkaya 2009, pp. 14–17).

Nevertheless, the constitution of 1961 is often described as the most liberal and democratic constitution that Turkey has ever had (Kreiser 2012, pp. 100–101; Özbudun and Gençkaya 2009, p. 16; Zürcher 2004, p. 246). It introduced proportional representation and included a bill of rights. On an institutional level, the framers were concerned to disperse authority and to introduce checks and balances against the perceived danger of a tyranny of the majority. Parliament was given a second chamber (senate) that had the power to block legislation passed by the lower house with mere ordinary majority. The constitution also included stronger protections of judicial independence, of the freedom of the press, and of the autonomy of the universities. Finally, the constitution of 1961 provided for a constitutional court endowed with a power of judicial review of legislation. The AYM was not shy to make use of its authority and soon began to play an influential role in defending the constitutionality of legislation and in protecting citizens’ civil rights (Zürcher 2004, p. 251). Under a charitable interpretation, the constitution of 1961, though introduced by illegitimate means, brought Turkey into line with the general postwar European trend towards constitutional and liberal democracy.

4.1 Origins of Amendment Review in Turkey

The constitution of 1961, like its predecessor, contained a single eternity clause. Article 9 of the constitution made Article 1, which protected the republican form of state, immune to amendment. Though the constitution did endow the AYM with the power to review ordinary legislation, it was silent, at least initially, on the question whether the court had the power to review constitutional amendments. Notwithstanding the lack of a clear authorization by the constitutional text, the AYM, in a series of decisions that began in 1970, held itself to have a power of amendment review (Gözler 2008, pp. 40–49, 64–66, 95–97). What is more, the AYM argued that the substantial characteristics of the Turkish Republic that had been specified in the preamble and in Article 2 of the constitution ought to be regarded as essential elements of the republican form of state, as protected by Articles 1 and 9. On this basis, the AYM claimed the power to review constitutional amendments not only for the conformity of their enactment with the constitution’s procedural rules of amendment, but also for their substantive conformity with the principle of republicanism, as understood in the light of the preamble and Article 2 (Özbudun 2011a, pp. 131–132).

The AYM had first put forward this reasoning in an obiter dictum in 1965. The first annulment of a constitutional amendment took place in 1970 (Decision of June 16, 1970, No. 1970/31, 8 AMKD 313 [1970]). The amendment had relaxed the personal qualifications required for eligibility as a deputy of the National Assembly, by permitting those who had been convicted of certain crimes that would normally have ruled out eligibility (such as defalcation, misappropriation, embezzlement, bribery or fraudulent bankruptcy) to run for parliamentary office in case they had later been pardoned. The decision to annul was justified on procedural grounds, but the AYM repeated its claim to a power of substantive amendment review (Gözler 2008, pp. 40–41). The court argued that it would be absurd to interpret the protection of the republican form of state narrowly as a mere re-affirmation of the principle of popular sovereignty. The AYM consequently held that ‘what is entrenched by Article 9 is not the word Republic, but the Republican regime whose features are identified in the above mentioned constitutional principles’ (8 AMKD 313, at p. 323). It went on to describe the nature of the Republic as a ‘nationalistic, democratic, secular and social state, governed by the rule of law, based on human rights and the fundamental tenets set forth in the preamble’ (8 AMKD 313, at p. 323). In 1971, the AYM again affirmed its power to review amendments on these substantive grounds (Decision of April 3, 1971, No. 1971/37, 9 AMKD 416 [1970]), in a case that concerned an amendment which had postponed the election of the senate by 16 months. In this instance, the AYM came to the conclusion that the amendment did not violate the principle of republicanism.

Critics of the judicial activism that came to the fore in the AYM’s claim to a power of amendment review evaluate the AYM’s stance as an attempt to protect the Kemalist project against any possible democratic change (Oder 2009, p. 269). In other words, they hold it to have been an exercise in hegemonic preservation (Bali 2013, p. 670; Özbudun 1983; Özbudun and Gençkaya 2009, pp. 108–109; Özbudun 2011a, pp. 131–138). This assessment strikes us as problematic, for a number of interrelated reasons.

The AYM was one of the first constitutional courts in the world to claim a power of amendment review. Such review, however, has become an increasingly common and accepted feature of democratic constitutionalism (see for comprehensive overviews: Roznai 2013a, 2013b). It is therefore difficult to dismiss the AYM’s claim to a power to amendment review as an obvious aberration from good democratic practice that could only have been motivated by an interest in hegemonic preservation. By our count (Vinx et al. 2015, pp. 108–209), there are 32 countries in the world with constitutional or supreme courts that are acknowledged to have the authority to review and to strike down constitutional amendments. Many of these polities clearly qualify as consolidated democracies (for instance: Argentina, Brazil, Czech Republic, Germany, India, Italy, Taiwan). There are 12 countries that have seen the actual invalidation, by the judiciary, of constitutional amendments passed by the legislature (Bangladesh, Belize, Benin, Brazil, Colombia, Czech Republic, India, Kenya, Taiwan, Tonga, Turkey, Ukraine).

One especially interesting example of a polity with amendment review is India. The Supreme Court of India successfully established an authority to review constitutional amendments for their conformity with the ‘basic structure’ of the Indian Constitution in the course of the 1970s and early 1980s (Austin 1999). According to the basic structure doctrine, all individual provisions of the Constitution of India are open to modification, but any constitutional amendment must respect the core principles and values embedded in the constitution as a whole (Krishnaswamy 2009; Jacobsohn 2010, pp. 49–58). The SCI arrived at its relevant decisions on the basis of arguments remarkably similar to those put forward by the AYM, and in the face of constitutional amendments that were passed specifically to deprive the court of the power of amendment review. The court was successful though the Constitution of India does not contain an eternity-clause. The persistence of the Supreme Court of India is often credited with having helped to preserve Indian democracy against the dictatorial ambitions of Indira Gandhi (Sathe 2002, p. 85), and the ‘basic structure doctrine’ is now an established and widely copied element of Indian constitutional law (Roznai 2014, chapter 3).

Another noteworthy example of a polity with amendment review is Germany. The German Constitutional Court has never invalidated a constitutional amendment, but it is recognized to possess the power to do so, and it has repeatedly reviewed constitutional amendments (Kommers 1997, p. 48; Möller 2004). The famous decisions on the treaties of Maastricht and Lisbon, which aimed to work out limitations on the constitutionally permissible depth of European integration, were instances of amendment review. In arriving at its decisions in these cases, the German Constitutional Court laid out a sophisticated account of the compatibility of amendment review with democracy (Vinx 2013). According to the doctrine adopted by the court (Murswieck 1978; Böckenförde 1987) there are limits to constitutional amendment that stem from the principle of popular sovereignty. A constitution is to be seen as the product of an exercise of constituent power. Attempts to change the constitution in ways that violate its eternity clause, therefore, must be democratically illegitimate since constituted powers—including the legislature acting under an amendment rule—lack the authority to overturn decisions taken by the people in its capacity as constituent power.

At any rate, the mere fact that a constitutional court exercises strongly counter-majoritarian powers does not suffice, on Hirschl’s account of hegemonic preservation, to establish that we are indeed faced with an illegitimate judicial power-grab. As we have seen, the main strand of Hirschl’s argument is committed to assessing the legitimacy of judicial power in terms of the substantive quality of decisional outcomes. Counter-majoritarian decision-taking, then, is hegemonic preservation, in a sense that supports a negative normative assessment, only if it tends to give rise to decisional outcomes that conflict with the realization of distributive justice or, alternatively, if they corrupt the democratic process so as to unfairly advantage one political camp. It would make little sense to level accusations of illegitimate hegemonic preservation against a court that defends social justice, the principles of the rule of law, the separation of powers, or rights essential to democracy against a legislative majority aiming to exploit what Carl Schmitt described as the ‘extra-legal premium on the legal possession of political power’ (Schmitt 1932, pp. 33–37). To assess whether the AYM’s attempt to claim a power of amendment review, under the constitution of 1961, was indeed an exercise of illegitimate hegemonic preservation, it is therefore necessary to consider the actual decisions that were taken in exercise of the power of amendment review. The attempt to do so suggests that the claim that the AYM was engaged in an exercise of hegemonic preservation, on behalf of a supposedly unified front of Kemalist state-elites in the military, the bureaucracy and the CHP, does not quite fit the facts.

4.2 The Struggle Over Amendment Review in the 1970s

To get at the point, we need to consider the role of the military in the post-coup constitution. Under the constitution of 1961, the military had been given political standing by the creation of a National Security Council (Milli Güvenlik Kurulu – MGK) (Ahmad 2003, p. 123). The function of the MGK, presided over by the president or the prime minister, but staffed mainly by representatives of the military, was to assist the cabinet in matters of national security. Its recommendations soon came to carry a stronger authority than that of mere advice. Through the MGK, the armed forces were, in effect, given immunity from political control, while being explicitly recognized by the civilian authorities as partners and guardians of the new order they had just created. That the military saw itself as the supreme guardian of the constitution became evident in 1971, during the so-called ‘coup by memorandum’ (Zürcher 2004, pp. 257–263). In response to political instability and social unrest, the military delivered an ultimatum to the prime minister that demanded ‘the formation, within the context of democratic principles, of a strong and credible government, which will neutralize the current anarchical situation and which, inspired by Atatürk’s views, will implement the reformist laws envisaged by the constitution’, putting an end to ‘anarchy, fratricidal strife, and social and economic unrest’ (Ahmad 1993, p. 148). The memorandum also announced that the military would assume direct control if such a government was not formed. The prime minister thereupon resigned and was replaced by a successor chosen by the military.

One of the key initiatives of the new government, which mainly relied on the support of the parties of the right, was to pass a large package of 44 constitutional amendments. The general thrust of these amendments was anti-liberal and anti-pluralist. The changes to the constitution allowed for the limitation, by law, of the freedoms protected by the constitution’s bill of rights—a potentially severe restriction of the AYM’s power of judicial review. The amendments also reduced the autonomy of the universities and the press, while they strengthened the role of the MGK (Zürcher 2004, pp. 259–260). That there was an intention to clip the wings of the AYM is clear from the fact that the amendment package explicitly determined that the AYM was to have a power to review constitutional amendments merely on procedural, but not on substantive grounds.

The AYM, however, did not refrain from engaging in review of constitutional amendments and went on to annul four more constitutional amendments in 1975, 1976 and 1977 (Gözler 2008, pp. 42–46; Özbudun 2007b, p. 261). In doing so, the AYM circumvented the restriction of its power of amendment review to mere procedural review by claiming that constitutional amendments which it judged to be in conflict with its understanding of the republicanism ought to be regarded as procedurally invalid. The constitution’s eternity-clause (Article 9) was held to imply that the constitution simply did not provide for any procedure to introduce amendments subverting the republican form of state. To put the point differently, the AYM argued that the distinction between substantive and procedural review inevitably collapses once an amendment is seen to conflict with a constitutional eternity-clause. An eternity-clause limits the scope of the legislative power that is conferred by a constitution’s procedure for amendment, i.e. it limits what can be done through the procedure. To ascertain whether some amendment is such that it could be enacted by the use of the procedure of amendment, and thus to decide whether the amendment in question is procedurally valid, it is therefore necessary, the AYM held, to determine first whether its substance conflicts with the eternity-clause.

Some legal scholars are highly critical of this argument and treat it as an unscientific political expedient (Özbudun 2011a, pp. 131–137; Gözler 2008, pp. 46–47). However, to reject the AYM’s reasoning one would have to take the view, or so it seems, that the eternity-clause of the constitution of 1961 was not legally enforceable. That position stands in tension with the fact that the constitution, as amended, did provide for a constitutional court endowed with a power of (procedural) amendment review. It also brushes up against the fact that the eternity-clause in question contained an unambiguous prohibition of the use of the amendment procedure for the abolition of the republican form of state. One recent commentator comes to the conclusion that the logic of the AYM’s argument for the inevitable collapse of the substance/procedure distinction, in the face of an eternity-clause, is ‘very strong, even foolproof’ (Arato 2016, p. 242). It is also noteworthy that no less a constitutional-theoretical authority than Hans Kelsen anticipated the AYM’s argument and claimed that ‘any material unconstitutionality is also a formal unconstitutionality’ (Kelsen 2015, p. 29). Kelsen’s stance on the issue, needless to say, could not have been influenced by questions of Turkish constitutional interpretation.

The substance of the four annulments of constitutional amendments that took place after 1971, at any rate, does not seem to substantiate the claim that the power of amendment review was used for the purpose of hegemonic preservation. Two of the amendments that were struck down by the AYM (both of which had been introduced in 1971) had meant to preclude judicial review of the decisions of the Supreme Council of Judges and Prosecutors (Hakimler ve Savcılar Yüksek Kurulu – HSYK), the body responsible for the placement and promotion of judges and prosecutors (Decision of January 28, 1977, No. 1977/4, 15 AMKD 106 [1977] and Decision of September 27, 1977, No. 1977/117, 15 AMKD 444 [1977]). A third decision struck down a constitutional amendment that had permitted the formation of military courts, in time of war, not staffed by professional judges (Decision of April 15, 1975, No. 1975/87, 13 AMKD 403 [1975]). The amendments at issue in these three cases were held to be in violation of the constitutional principle of the rule of law, which, in turn, was regarded as essential to a republican state. The general thrust of these three decisions was to strengthen judicial independence as well as the right to a fair trial. These goals are hardly anti-democratic or specifically Kemalist, and they clearly do not conflict with the realization of distributive justice. The fourth annulment concerned an amendment that had determined that the compensation for the expropriation of real estate must not exceed the value of the property the owner had previously declared to the tax administration (Decision of October 12, 1976, No. 1976/46, 14 AMKD 252 [1976]).

The confrontation between the AYM and a government backed by the military that played itself out in the struggle over amendment review could be regarded as one of the earliest signs of a crack within the Republican alliance, which Belge (2006) traces to the late 1970s (p. 653). While the CHP, under its leader Bülent Ecevit, who had objected to the ‘coup by memorandum’, was moving to the left of the political spectrum, the military was moving to the right. It ultimately came to blame the 1961 constitution for granting too much autonomy to the non-military state elites, and advocated a restriction of judicial independence, so as to create a more docile constitutional court (Shambayati and Kirdis 2009, p. 773). The political turmoil of the late 1970s, which led to increasingly frequent and violent clashes between extremists on the left and on the right, convinced the military of the harmful effects of an excessive dispersal of authority. In the military’s view, the constitution of 1961, by providing unduly extensive rights and liberties to members of a politically immature society, had contributed to an intense politicization of Turkish society. Its checks and balances had also created a divided state that was incapable to provide stability and security. The military therefore intervened once more, in 1980, in order to end what it regarded as a confused and chaotic structure of multiple, independently functioning authorities. What was needed, according to the military, was a reassertion of the supremacy of a sufficiently unified state.

5 The Constitution of 1982: Expansion of Strong Constitutional Entrenchment

If the 1961 constitution reflected the fears of all constituent groups of the Republican alliance, the 1982 constitution reflected, first and foremost, the anxieties of the military. Of all Turkish constitutions, it is the most authoritarian and statist in character; due to the military’s heavy involvement in its composition, which ensured that the military would continue to play an important political role (Arato 2016, pp. 224–225; Gülener 2012, p. 257; Özbudun and Gençkaya 2010). Although the military had been involved in the constitution-making process after the 1960 coup, the process had been led mostly by civilian actors. This was certainly not the case with the 1982 constitution. The military’s distrust towards the civilian bureaucratic agencies culminated in a constitutional text designed to maintain the military as the ultimate guardian and arbiter of the political spectrum (Shambayati 2008, p. 293; Shambayati and Kirdis 2009, p. 773).

The creation of the 1982 constitution was thoroughly dominated by the military. The MGK convened a bicameral constituent assembly, of which the MGK itself made up one chamber. The civilian chamber, the ‘Consultative Assembly’, was even less representative of society at large than the 1961 House of Representatives. Whereas the latter had included representatives of two political parties and of various other institutions of civil society, all members of the former were individually appointed by the MGK. The civilian chamber, what is more, had to defer to the MGK, since the latter claimed an unrestricted power to amend or reject the constitutional draft prepared by the Consultative Assembly, with no machinery envisaged to resolve the differences between the two chambers. Public discussion of the content of the draft constitution was forbidden during the drafting process. Furthermore, the 1982 constitutional referendum was not conducted in a free atmosphere. The MGK made it understood that in case of a rejection of the draft, the MGK-regime would continue indefinitely. The ratification process, thus, was designed in a way that effectively secured approval (the constitution was ratified with almost 92% of the votes). Votes for adoption of the draft-constitution were counted as votes for the election of Kenan Evren, the General who had led the 1980 coup, to the office of President of the Republic. Given that the choice people were given was between a military regime (lacking any democratic legitimacy) and a military regime under civilian disguise (dubiously legitimated and ‘limited’ by a constitutional text adopted through a suspicious referendum), it would be absurd to claim that they were given any real choice at all.

5.1 Eternity Clauses and Constitutional Guardianship under the Constitution of 1982

Though it did continue the constitutionalist tradition that had been started by the Constitution of 1961, the Constitution of 1982 was also designed to assure the continuing political control of those who had made it. The presidency, incompatible with membership in a political party, was given significant authority and powers of appointment, including to the AYM, so as to limit the influence of parliament (Shambayati and Kirdis 2009, p. 774). The role of the MGK was strengthened as well. The number and weight of military commanders in the MGK were increased at the expense of its civilian members. It was also given extensive power with regard to the formulation and implementation of national security policy. Although the council’s role was supposedly advisory in nature, civilian governments were obliged, as a matter of political convention, to comply with its decisions. Through the ‘priority’ consideration given to its policy recommendations, its status was elevated to that of an unequal partner in the ruling bloc: primus inter pares (Cizre 1997, p. 157). Despite the controversies that had surrounded the AYM in the 1960s and the 1970s, the 1982 constitution maintained the principle of judicial review. However, the courts were not to be allowed to review the decisions of the outgoing military regime. The constitution also weakened the guarantees of judicial independence and changed the procedures for the appointment of the justices of high courts, to ensure that the military would continue to influence the composition of the judicial organs (Shambayati and Sütçü 2012, p. 110). Some scholars suggest that the Court now functioned as the ‘administrative attaché’ of the military, which considered the AYM’s role as that of assisting the executive (Shambayati and Sütçü 2012, p. 110; Shambayati and Kirdis 2009, p. 770, 775).

The 1982 constitution contains three clauses (Articles 1–3) describing the character of the state, and these clauses are made eternal by a fourth clause (Article 4) which rules out their amendment. The first clause once again designates the form of the Turkish state as a republic and the second characterizes the republic as a ‘democratic, secular, social state governed by rule of law’. Article 2 also protects ‘public peace, national solidarity and justice’, and the ‘fundamental tenets set forth in the Preamble’. The latter makes reference to ‘Turkish national interests’, ‘Turkish historical and moral values’ as well as the ‘nationalism, principles, reforms, and the modernism of Atatürk’. A further irrevocable provision is to be found in Article 3 of the constitution, which declares the nation and the territory of the state to be one indivisible entity and determines that the state’s official language is Turkish.

What is remarkable about the eternity clause in the constitution of 1982 is that it explicitly protects the expansive understanding of republicanism that, as we have seen, the AYM had earlier attempted to entrench through the mechanism of amendment review. Nevertheless, the constitution also limited the review of constitutional amendments to procedural (as opposed to substantive) review. The relevant constitutional provision (Article 148.2) was carefully formulated in such a way as to forestall the collapse of the substantive and the procedural that the AYM, which was to continue to exercise a power of review over ordinary legislation, had relied on to circumvent the limitation to procedural review first introduced by the amendment package of 1971. The AYM, in other words, was not to have the power to strike down a procedurally valid amendment for lack of substantive conformity to the constitution’s eternity clauses. Given the military’s control over the constitution-making process, there can be little doubt that this limitation of judicial power reflected the preferences of the armed forces.

Though the AYM was denied a power of amendment review, it did rely on the content of the constitution’s unamendable core in exercising the power to dissolve political parties. The relevant provisions of the constitution (Articles 68.4, 69) make reference to the principles protected by Articles 1-4 as grounds for the closure of a political party. In its closure decisions, the court came to rely heavily on the ideological outlook set forth in the preamble and on a rather strict interpretation of eternity clauses (Gülener 2012, pp. 288–289, Yüzbaşıoğlu 1993, p. 130). Not surprisingly, the AYM is often argued to have abused the power to dissolve parties, by endorsing Kemalism as its official ideology, and by acting as the agent of the state elites that had drafted the constitution (Hirschl 2012, p. 324; Koğacıoğlu 2003; Shambayati and Kirdis 2009).

The AYM undoubtedly has a large record of party dissolution case. Six parties were closed under the 1961 constitution and twenty-two under the 1982 constitution. These closures were typically directed against Kurdish and Islamist parties (Algan 2011; Coşkun 2008; Koğacıoğlu 2003; 2004; Öden 2003; Yokuş 2001; Yüksel 1999). Most party-dissolutions were based on the unamendable principles of the indivisibility and territorial integrity of the state as well as on the principle of secularism. Many scholars suggest that the AYM’s decisional record on party closure an abusive misuse of militant democracy (Hakyemez 2000, 2001; Sajo 2004). But there are also voices that argue that the court’s decisions in the party-closure cases exhibit a great perseverance in protecting the foundational values of the Republic (Shambayati and Sütçü 2012, p. 107). They are said to have forced extremist, anti-systemic parties to adopt a milder discourse and to move closer to the centre of the political spectrum.

Be that as it may, there can be little doubt that the AYM’s decisions on party closure were meant to preserve the Kemalist understanding of Turkey’s constitutional identity. This, however, does not by itself entail that the AYM was engaged in an exercise of hegemonic preservation, in Hirschl’s sense of the term. It is possible for courts to be employed in the service of a political hegemon, but in a way that violates the agency constraint. In the years after the enactment of the constitution of 1982, the AYM was clearly kept on a much shorter leash, by the military, than in the second republic. The military, at least initially, was the hegemon, and while it used the judiciary to get its way where that appeared advantageous, it did not truly cede the power of constitutional guardianship to the AYM, as postulated by Hirschl’s model of hegemonic preservation by judicial empowerment. This is evident in the fact that the court was denied a power of amendment review (and therefore blocked from playing the role of the final interpreter of constitutional identity), in the fact that the military remained outside of civilian judicial control, and in the fact that the military continued to take the initiative in matters of constitutional guardianship when it saw a need to do so, for instance in the so-called ‘postmodern coup’ of 1997 (Zürcher 2004, pp. 299–305) that resembled the earlier ‘coup by memorandum’.

5.2 The Process of Constitutional Reform in the 1990s

The military’s role as a guardian of the constitution, however, came under increasing pressure with the passage of time, and the civilian political elite aimed to regain its influence, not in the least by the use of the power of constitutional amendment. Throughout the 1990s, the constitutional text itself underwent a great amount of change as a result of democratizing reforms, first initiated by Turgut Özal in 1987. One-third of the 1982 constitution has been changed since the 1990s, through eight separate amendment packages, typically in conjunction with European Union accession negotiations that gained momentum at the time (Bilgin in Arjomand 2008, pp. 123–146; Özbudun 2007a; Özbudun and Gençkaya 2009). The general aim of the amendments was to soften the authoritarian nature of the 1982 constitution. Particular targets were references to the sacredness of the Turkish state in the preamble to the constitution and the view that the MGK is a legitimate representative of the nation (Oder 2009, p. 264). The amendments also removed many of the legacies of the military coup, such as prohibitions on trade unions and associations. The constitutional amendments were usually drafted by an All Party Accord Commission in which each party in parliament was represented equally. This compromise-oriented roundtable approach was so successful that it quickly became customary (Arato 2016, pp. 231–235; Arato and Tombuş 2013, p. 430).

The roundtable approach helped to overcome the procedural barriers to amendment put up by the Constitution of 1982 (Article 175). According to the Constitution of 1982, as amended in 1987, an amendment can either be passed by a majority of three-fifths of members of parliament and a subsequent public referendum (the ‘referendum route’) or it can be passed by two-thirds of members of parliament without a subsequent referendum (the ‘consensual route’). The President, however, may decide to submit a bill of amendment that has been passed with two-thirds to referendum. It is also in the power of the president to refer a bill of amendment that has been passed by a mere three-fifths of members of parliament back to parliament for reconsideration, before it is submitted to referendum. In this case, the bill of amendment, if parliament does not change it, must receive a majority of two-thirds of members of parliament before it can be submitted to referendum.

It is not surprising that the parliamentary parties adopted a consensual approach to the passage of amendments. They shared an interest in extending the political authority of civilian politicians and they had to overcome the potential hurdle of a presidential veto to realize their aims. However, constitutional politics in Turkey eventually came to take on a rather different complexion once the Justice and Development Party (Adalet ve Kalkınma Partisi – AKP) had managed to win both a substantial parliamentary majority and the office of the presidency.

6 The AKP Era: Towards a New Constitution?

The AKP government that came to power in 2002 continued the process of constitutional reform. It passed a series of amendment packages designed to satisfy the Copenhagen criteria as well as to further curb the power of bureaucratic institutions over the elected representatives of the people. The AKP’s governmental program of 2006 put forward the project of a general constitutional revision that would create a new and civilian constitution, as had been repeatedly demanded by the European Union. The plan was to enact a completely redrafted constitutional text, by the use of the amendment procedure of the existing constitution. AKP-led Turkey seemed likely to finally consolidate Turkish democracy, i.e. to overcome the alternation of military-bureaucratic and majoritarian forms of authoritarianism, and thus to help the country transcend the ‘winner take all’ political paradigm that had characterized Turkish politics for so long (Çınar and Sayın 2014, p. 366). These hopes, however, were to be disappointed. The AKP’s second term in power came to be characterized by conflicts between the AKP and the Kemalist state-elites still entrenched in the military, the bureaucracy, and the judiciary. These conflicts would derail the project of enacting a new constitution on the basis of a broad political consensus. The AYM played a central role in this struggle and made a renewed attempt to shape the direction of constitutional reform by the exercise of a power substantive amendment review (Özbudun and Gençkaya 2009, pp. 97–112; Arato 2016, pp. 238–247; Bali 2013, p. 674, 678).

6.1 The Constitutional Crisis of 2007–2008

A first major conflict between the AYM and the AKP broke out over the election of a new president in 2007 (Özbudun and Gençkaya 2009, pp. 97–103). At the end of Kemalist President Ahmet Necdet Sezer’s term of office, the AKP seemed to have enough votes in the parliament to elect its own ‘religious’ president and thus to gain much greater control of the process of constitutional amendment.Footnote 1 Yet, it was initially prevented from electing its own candidate. According to the 1982 constitution, as it then stood, a new president was elected by parliament in a maximum of four rounds of voting. The decisional quorum for electing a president was two-thirds of the full membership of the assembly in the first two rounds, but the quorum dropped to simple majority in the subsequent two rounds. The AKP held an absolute majority of parliamentary seats, but was short of a majority of two-thirds. It therefore expected to be able to elect its candidate, Abdullah Gül, in the third round of voting. A retired chief prosecutor of the Republic (Sabih Kanadoğlu), however, argued that a majority of two-thirds was necessary not merely to elect a president in the first two rounds, but also to open the parliamentary session to elect the president. A quorum of two-thirds for opening the session was not obtained, due to the fact that opposition deputies boycotted the election. The AKP’s attempt to elect a new president was followed by a statement of the Turkish Armed Forces that reaffirmed the military’s commitment to defending secularism. The CHP, in its turn, appealed the election to the Constitutional Court, and the AYM, in an extremely controversial ruling, accepted Kanadoğlu’s argument and annulled the election of Abdullah Gül. The Court’s decision appeared to have been influenced by the memorandum published by the military. Critics therefore suggest that the AYM had, once more, acted as a mere ‘attaché’ of the military-led Kemalist bureaucratic elites (Coskun 2010, p. 53).

The deadlock over the presidency led to new elections that were widely perceived as a public rebuke to the Kemalist block. The AKP managed to significantly increase its share of the vote and it subsequently succeeded to elect Gül to the presidency in the new parliament, without provoking any interference on the part of the military. Even before the parliamentary election, the AKP had managed to pass a constitutional amendment, by putting it to a popular referendum despite the resistance of President Sezer, which determined that future presidents would be elected by popular vote and no longer by parliament. President Sezer referred the amendment to the AYM before the referendum, on the ground that its passage in parliament had been procedurally flawed, but the court refused to uphold his complaint.

The AKP reaffirmed its desire for a new, civilian constitution in its election manifesto and the party started the process of drafting a new charter just before the elections. The task was assigned to a commission composed of a group of renowned scholars of constitutional law chaired by Prof. Ergun Özbudun (Özbudun and Gençkaya 2009, pp. 103–105; Özbudun 2011a, pp. 139–150). The AKP planned to rely on its parliamentary majority and its control of the presidency to enact the draft as a constitutional amendment and then to submit it to referendum. If necessary, this was supposed to happen without the assent of the parties of the opposition. At the time, the AKP held a parliamentary majority of 341 seats, eleven seats above the threshold of three-fifths that is required to take a draft amendment to referendum, but below a majority of two-thirds by which an amendment can be ratified in parliament without being submitted to referendum. The draft aimed to expand and protect civil rights and liberties, in line with the Universal Declaration of Human Rights and the ECHR, while preserving the characteristics of the Turkish Republic as a democratic, secular, and social character of the state based on human rights and the rule of law (Özbudun and Gençkaya 2009, pp. 103–104). However, these principles were not to be designated as unamendable. In addition, the preamble was kept very short and concise. The draft, however, was ‘silently shelved’ after the headscarf amendment in 2008 and the ensuing constitutional threat of a closure of the AKP (Özbudun and Gençkaya 2009, p. 105; Köker 2010, p. 328).

The promise to lift the ban on the wearing of headscarves at institutions of higher education had been an important part of the AKP’s election manifesto in 2007, and it became the number one issue of the political agenda in early 2008, after Prime Minister Recep Tayyip Erdoğan had stated that the ban should be lifted (Eligür 2010, pp. 254–275). The Prime Minister’s statement was strongly criticized by the CHP, but it was supported by the second largest opposition party, the ultra-nationalist Nationalist Action Party (Milliyetci Halk Partisi - MHP), whose leader proposed to the AKP to resolve the headscarf issue by passing a single amendment, instead of waiting for the preparation of a new constitution. In contrast to its actions during the conflict over Gül’s presidency, the military now decided to keep its silence, whereas the judiciary took the initiative and issued several warnings. On January 17, 2008, Abdurrahman Yalçınkaya, the chief public prosecutor, warned the AKP that ‘political parties cannot aim at changing the Republic’s principle of secularism.’Footnote 2 Likewise, the Council of State declared that headscarf freedom would be incompatible with the constitution’s principle of secularism.Footnote 3

Despite these warnings, intensive talks between the AKP and the MHP culminated in the passage, with a majority of more than two-thirds of members of parliament, of changes to two constitutional articles: Article 10 concerning equality before the law and Article 42 on the right to education. These changes were formulated without direct reference to the headscarf issue, but they were designed to overturn the ban on the headscarf, by providing that any exception to the right of equal access to education had to have a statutory basis (the ban on the headscarf in Turkish universities had been introduced through administrative and judicial decisions). While it was still unclear whether the amendments would automatically lift the headscarf ban, the CHP and the Democratic Left Party (Demokratik Sol Parti – DSP) challenged the constitutional amendment before the Constitutional Court, arguing that it violated the unamendable principle of secularism and was therefore null and void. Though the amendment had undoubtedly been passed without violation of the constitution’s procedural rules concerning amendment, the AYM concurred and went on to annul it.

This surprising decision was heavily criticized by constitutional scholars (Özbudun 2009, pp. 533–538; Özbudun and Gençkaya 2009, p. 109; Özbudun 2011a, pp. 131–138; Roznai and Yolcu 2012). Up until the headscarf decision, the AYM had rejected all petitions that had sought review of the substance of constitutional amendments, thus signalling its acceptance of the limitation of procedural review of amendments (Halmai, 2012, p. 188). Yet in this case, in an apparent reversal of its earlier stance, the AYM ruled that it did have competence to review and to annul an amendment that was in material violation of the constitution’s unamendable articles. The AYM justified that claim by resurrecting the argument that amendments conflicting with the constitution’s unamendable core ought to be regarded as procedurally invalid—a move that led critics to accuse the AYM of having openly violated the constitution’s explicit limitation of its power of amendment review (Coskun 2010, p. 56). A request was subsequently made to the AYM by the Chief Public Prosecutor to close down the AKP, on the ground that it had become a focal point of anti-secular political activities. Six out of eleven members of the Constitutional Court voted for the AKP’s closure. Since seven votes would have been needed to close down the party, the AKP nevertheless narrowly avoided dissolution.

6.2 The Constitutional Amendment Package of 2010

The difficulties and resistance that the AKP encountered in its attempt to draft a new constitution increasingly led the party to resort to administrative measures to achieve its intended goals, and the project to enact a novel constitution was quietly abandoned, for the time being. The eventual lifting of the headscarf ban, for instance, was achieved through a memorandum issued by the AKP-controlled Higher Education Council (Yüksek Öğretim Kurulu – YÖK). The party also reverted to a strategy of gradual change of the constitutional system, and introduced a series of particular constitutional amendments for which it expected to find sufficient public support.

An especially important and controversial constitutional amendment package was enacted in March 2010. The AKP relied on its parliamentary majority to submit the package to referendum, without securing the approval of the opposition, and while refusing any demand, on the part of the opposition, to allow the people to vote separately on the key proposals included in the package. The proposal was adopted with 58% percent of the popular vote in a decidedly contested referendum. The package included twenty-six constitutional amendments, many of which were undoubtedly liberal and progressive, in particular the introduction of an individual right of constitutional appeal (Yazıcı 2010; Özbudun 2011b). But the package also aimed to restructure the AYM and to redefine its powers of review as well as to increase the government’s influence over HSYK (Alessandri 2010; Arato 2010; Halmai 2012).

The CHP challenged the 2010 amendment package in the Constitutional Court on the grounds of a concern for judicial independence, and it based its appeal specifically on the eternity clauses specified in Article 2 and Article 4 of the constitution. It claimed that the changes to the structures of the AYM and the HSYK undermined the separation of powers and were thus in violation of the constitutionally entrenched principle of the rule of law. The AYM affirmed its authority to review amendments for their material conformity with the constitution’s unamendable core and judged some technical parts of the package as unconstitutional, but it nevertheless allowed the bulk of the package to go to referendum.

It is not surprising that the amendment package of 2010 should have been criticized by the opposition. For one thing, the AKP abandoned the consensual form of constitution-making that it had practised in its first term, following the tradition of the 1990s. In 2010, however, the AKP’s approach was majoritarian rather than consensual and the whole campaign polarized Turkish society (Arato 2016, Chapter 5; Alessandri 2010, p. 25). Secondly, the changes made to the composition of and to the procedure of appointment to the Constitutional Court raised the suspicion that the AKP was interested in packing the AYM, so as to reduce the obstacles that the AYM’s claim to a power of substantive amendment review threatened to put in the way of the party’s opportunity to use its parliamentary majority, its control of the presidency, and its ability to win popular referenda for the purpose of further constitutional reform (Alessandri 2010, p. 26; Arato 2016, pp. 247–254; Arato 2010; Bali 2010, 2012, p. 298; Halmai 2012, p. 189).

The amendments increased the size of the AYM from eleven regular and four substitute justices to seventeen permanent justices. Before the passage of the amendment package, all members of the court were appointed by the president of the republic—four directly and the rest upon nomination by Turkey’s five highest courts (other than the AYM) and YÖK. Under the new rules of appointment, the president continues to appoint 14 of the members of the court, of whom ten are nominated by highest courts as well as by YÖK while four continue to be chosen directly by the president. Though the change did not affect the number of direct appointments to be made by the president, the president, in making direct appointments, is now able to choose from a somewhat greater pool of eligible candidates than before. More importantly, three members are now elected by parliament, where a simple majority suffices for an appointment in the third round of balloting.

These changes were certainly not objectionable per se, and they could plausibly be regarded as steps towards democratization of the high judiciary (Yazıcı 2010; Özbudun 2011b). Due to the fact, however, that the AKP holds a parliamentary majority as well as the presidency, and is likely to continue to do so, the changes undoubtedly increased the influence of the governing party on the appointment of judges to the AYM and thus on the court’s composition (Halmai, 2012, p. 189; Arato 2010). The amendment package also provided that any future decision on the part of the AYM to invalidate a constitutional amendment (or to close down a political party) will require a majority of two-thirds (as opposed to three-fifths) of the AYM’s 17 judges. The cumulative result of these changes is that it has become more difficult for the AYM to engage in amendment review against the preferences of a president supported by a parliamentary majority. At least as a matter of politics, the court’s claim to a power of amendment review is therefore unlikely to continue to afford the AYM with a power to control the process of constitutional reform.

6.3 Constituent Power and the Legitimacy of Constitutional Reform

Scholarly opinion on the AYM’s behaviour during the constitutional crises of 2007–2010 has tended to be highly critical. Commentators sympathetic to the project of constitutional reform have described the court’s renewed attempt to exercise amendment review in the headscarf case as an illegitimate usurpation of the constituent power of the people—one that aimed to block any possibility of the replacement, through democratic constitutional change, of an irredeemably authoritarian charter that was foisted upon an unwilling people by a military coup (Özbudun and Gençkaya 2009, pp. 108–109; Özbudun 2011a, p. 134, 143). The AYM’s reliance on the argument about the collapse of the procedure/substance-distinction in amendment review, in particular, is often regarded, by such critics, as little more than a ruthless gambit of hegemonic preservation.

To be sure, this assessment is not without considerable merit. If there is an episode in Turkish constitutional history that does appear to fit Hirschl’s model of hegemonic preservation, the headscarf episode is it. We fully agree that the headscarf decision was substantively mistaken, in that it was based on an interpretation of secularism that was excessively narrow (Roznai and Yolcu 2012, pp. 202–206). The headscarf decision thus satisfies the output-constraint. It also satisfies the agency constraint, since the AYM, in taking that decision, did appear to claim an independent role of constitutional guardianship not directly backed up by a threat of military intervention. Nevertheless, we have misgivings about the constitutional-theoretical assumptions that seem to underpin applications of the hegemonic preservation thesis to the headscarf episode. We are not convinced, to begin with, that the case shows that substantive amendment review as such is inevitably democratically illegitimate. As we have indicated above, we do not think that the AYM’s argument about the collapse of the distinction between procedure and substance in amendment review is obviously wrong, though we agree that it was abused in the headscarf case. Most importantly, the view that the court usurped constituent power in overstepping the constitutional boundaries of its authority of amendment review does not, by itself, answer the question of where constituent power is to be located instead.

Authors who criticise the AYM’s claim to a power of substantive amendment review tend to hold that constituent power should be attributed to the constitutional legislator, i.e. to the parliamentary majority (or the combination of parliamentary majority, presidential approval, and popular referendum) that suffices to pass an amendment in conformity with the current constitution’s (procedural) rules of amendment (Özbudun 2011a, p. 138). Note that the claim that the AYM usurped constituent power is stronger than the claim that it overstepped the boundaries of its competence in striking down the headscarf amendments. It suggests that there is some other constitutional authority that is entitled to exercise the constituent power of the people, and this claim, in the Turkish context, can only be a reference to the constitutional legislator. It should be clear, though, that the claim that the AYM does not have the power of substantive amendment review, even if true, does not entail that constituent power is therefore in the hands of the constitutional legislator. The attribution of constituent power to the constitutional legislator is not a self-evident truth of constitutional theory but a highly contested piece of doctrine. As we pointed out above, the practice in a growing number of influential constitutional traditions explicitly rejects the attribution of constituent power to the constitutional legislator in favour of a distinction between original constituent power and the mere power of constitutional amendment. What is more, there is a growing body of constitutional theory that resists the identification of the constitutional legislator with the constituent power (Ackerman 1991, pp. 266–294; Amar 1995) and that argues that the power of amendment is therefore limited (see Schmitt 1928, pp. 91–122; Murswieck 1978, pp. 168–174; Conrad 1999; Loughlin 2003, pp. 99–113; Grimm 2009, pp. 35–41; Colon-Rios 2012; Roznai 2014, chapters 4–8; Stacey 2016).

Those who defend the attribution of constituent power to the constitutional legislator in the Turkish context face the additional problem that Articles 1–4 of the constitution of 1982 contain unambiguous material limitations of amendment, which seem to preclude an attribution of constituent power to the constitutional legislator—irrespective of whether one agrees with the AYM’s claims to have a power to engage in substantive review of amendments. To be sure, the constitution’s limitation of the power of amendment review to procedural review—provided one rejects the AYM’s argument from the collapse of substance and procedure—will render the constitution’s eternity clauses judicially unenforceable. But is one entitled to infer from this lack of judicial enforceability that the constitutional legislator is, in effect, endowed with materially unlimited authority to change the constitution?

The view that the lack of judicial enforceability of the constitution’s eternity clauses entails that the constitutional legislator itself ought to be the final judge of what the constitution’s eternity clauses permit or forbid is a rather thin reed on which to base an attribution of constituent power (Gözler 2008, pp. 19–22, 54). It is unlikely, to begin with, that the framers of the constitution of 1982 intended to grant constituent power to the constitutional legislator defined by that constitution’s amendment rule. It is unlikely, in other words, that the military wanted the unamendable core of the constitution to be at the disposal of the process of constitutional amendment. It is more plausible to assume that the military decided not to endow the AYM with a power of amendment review for the reason that it saw itself as the final arbiter of constitutional identity.

It might be replied, of course, that the constitution-making process of 1982 had very little democratic legitimacy and should not be regarded as an authentic exercise of constituent power. The intentions of the framers, therefore, should not be seen to carry decisive weight in interpreting the constitution and in determining the scope of the authority of the constitution’s amendment rule. What is more, the fact that the military was apparently no longer willing or able, in the late 2000s, to step up to the role of constitutional guardian by taking direct political control, should be welcomed and the resultant opening for democratic constitutional change be exploited. We agree, but we think the point gives rise to another question. If the constitution of 1982 is not the result of an authentic exercise of constituent power, then why should one accept the view that any constitutional change brought about under its highly specific procedures of amendment is to be regarded as an authentic exercise of constituent power? These procedures, after all, make it comparatively easy for a one-party government to change the constitution by using the referendum route, once it is supported by a solid parliamentary majority and in control of the presidency—a problem that is potentially exacerbated by the fact that a Turkish political party must win at least 10 percent of the vote to be represented in parliament.

The answer to this question cannot be that this is what the constitution says. If the constitution’s legitimacy is held to be fundamentally problematic, then the mere fact that some amendment or a general constitutional revision was produced in conformity with the constitution’s procedural rules cannot, by itself, endow the changes in question with the quality and legitimacy of an authentic exercise of constituent power. The authenticity of an exercise of constituent power, clearly, depends on considerations of legitimacy, and not merely on considerations of constitutional legality—in particular if the latter are based on the provisions of a constitution that is to be replaced because it is held to be democratically illegitimate (Arato 2016, Chapter 5). This is not to say, we hasten to add, that a constitutional revision that relies on the amendment rules of the existing constitution could not amount to an authentic exercise of constituent power. Our claim is, rather, that it would have to satisfy conditions of legitimacy that go beyond mere conformity to the procedural rules laid down in the constitutional text.

It might be argued that when the AYM claimed that it had the power to review amendments for their material conformity with the eternity clauses of the constitution of 1982 it was, in effect, confronting the Turkish public with a Hobson’s choice between the acceptance of continued Kemalist dominance under the constitution of 1982, as interpreted by the AYM, or an open and violent revolution to break constitutional continuity (Arato 2016, pp. 243–248). This argument has considerable strength, but it appears to fail to take account of what might have been a third alternative. As Andrew Arato (2016) has pointed out there would have been the option to drive the process of constitutional reform forward through a return to the consensual model of constitutional amendment that had been used in the early 2000s (p. 248). The AYM is barred from reviewing constitutional amendments on its own initiative. Only the president or a group of at least 110 MP’s have the standing to appeal to the court for the review of a constitutional amendment (Article 150). It would thus have been possible to avoid judicial review of amendments backed up by a parliamentary coalition of all major parties acting in accord with the president. Arato argues that the AYM’s renewed claim to a power of material amendment review, if it had been more successful, could have had the effect of channelling the project of constitutional reform back onto the consensual route. Whether this is true or not is hard to say. But we agree with Arato that constitutional reform in a highly consensual mould has a much better claim to be regarded as an authentic exercise of constituent power, from the perspective of legitimacy, than a more majoritarian process orchestrated by one single political party whose constitutional vision is deeply contested.

If it was true that the AYM, and its supposed attempt to engage in hegemonic preservation on behalf of Kemalist state elites, were the major stumbling blocks for the creation of a new and more thoroughly liberal and democratic constitution, then one would expect the project of enacting such a constitution to have succeeded by now. After all, the AYM lost much of its de facto power to resist an AKP-led project of constitutional reform through a deployment of amendment review with the passage of the 2010 amendment package. However, attempts to enact a new constitution have continued to this day, and they have so far failed to produce conclusive results.

6.4 Recent Developments

In the aftermath of the 2011 general elections, the AKP continued to voice a desire to change the constitution. The opposition parties expressed support on the condition that the drafting process would be an inclusive one, so as to encourage and ensure the participation of all political parties, NGO’s, civil society organizations, professional and trade associations, bar associations, and so on. The AKP accepted a return to a consensual strategy of constitution making. It had failed to win the necessary number of seats in parliament to submit amendments to referendum, despite a rising share in the popular vote, as seats in parliament were now split between four parties (Arato 2016, pp. 256–260; Arato and Tombuş 2013, pp. 429–431; Keyman 2012).

An All Party Accord Commission was formed after the 2011 elections, in an attempt to reach a common constitutional text through debate. The committee included three representatives from each of the four political parties represented in parliament (the AKP, the Kemalist CHP, the nationalist MHP and the Kurdish Peace and Democracy Party (Barış ve Demokrasi Partisi – BDP). However, after numerous meetings, debates, and drafts the project was eventually shelved because the parties had reached an impasse. Of the 172 total clauses in the last draft only 60 had been agreed upon. Major disputes related to the preamble and the main principles, the judicial function, and the legislative function. The different constitutional visions turned out to be irreconcilable. While the CHP and the MHP insisted on keeping the eternity clauses of the 1982 constitution, the BDP demanded their removal.

The biggest roadblock to the possibility of consensus, however, is the AKP’s proposal to shift from a parliamentary to a presidential system of government, a demand that has come to dominate Turkish constitutional politics, especially after the election of the former Prime Minister Erdoğan to the presidency in 2014. The AKP government first presented a proposal for a presidential system in November 2012. According to the proposal, the president was to have the power to dissolve the legislative assembly and call for elections, send laws passed by the assembly back to it for further discussion, select half of the members of YÖK, select university rectors, select half of the members of the AYM, the Council of State, HSYK, and elect the president of the Supreme Court. In addition, the president was to be given a right to rule by decrees on matters he or she considers important for general political functioning. Legal scholars suggest that the proposal lacked many of the checks and balances that usually accompany presidential systems of government (Özbudun 2015, p. 126). The major intention, it appears, was to concentrate power as much as possible in the hands of the president.

Due to the resistance of the opposition, and due to its present lack of a majority of three-fifths of the seats in parliament, which would suffice to submit another amendment package to referendum, the AKP has so far been unable to implement its plans for a new constitution that creates a presidential system of government. It remains to be seen whether Turkey will return to a more consensual mode of constitution-making on the basis of the new-found unity among the AKP and some parties of the opposition that resulted from the failed coup-attempt in July 2016. It is quite possible that Turkey will have a new constitution soon, but only time will tell whether that constitution will be less contested than the constitution of 1982.

7 Turkey and the Hegemonic Preservation Thesis

Does the use of strong constitutional entrenchment and amendment review in Turkey conform to Hirschl’s model of hegemonic preservation? We started out from the claim that it does not, and this claim may now seem even more perplexing, given the narrative we have presented. What could be more obvious than the fact that the constitutions of the Turkish Republic have been characterized by the attempt, on the part of Kemalist elites, to impose their vision of political identity on Turkish society? Of course, we do not disagree with the view that the use of constitutional entrenchment and judicial review in the history of Turkish constitutionalism has been shaped by the political hegemony of Kemalist elites. What we argue is that this observation does not suffice to show that the model of hegemonic preservation illuminates the constitutional history of the Turkish Republic.

7.1 Turkey and the Output Constraint

A first problem for applying the model of hegemonic preservation to Turkey arises from Hirschl’s portrayal of elite motives for constitutionalism. It can hardly be said that the Kemalist elites were concerned to impose and to preserve the ascendancy of neo-liberalism. Kemalism was committed to a statist economic approach. The neo-liberalization of Turkey’s economy was pushed forward by conservative politicians, like Turgut Özal, who stood at a certain distance to the Kemalist project (Zürcher 2004, pp. 306–312; Tuğal 2016). It might be replied, of course, that the application of the model of hegemonic preservation could very well focus on some other elite interest(s)—presumably, in the case of Turkey, the preservation of a Kemalist understanding of constitutional and political identity—without robbing the application of the model to Turkey of its explanatory force. Such an argument would still have to assume, however, that the AYM’s conception of republicanism violated the output constraint.

It is unclear whether such an assessment would be defensible with respect to the behavior of the AYM under the constitution of 1961. Recall that Hirschl’s model does not claim that any instance of counter-majoritarian constitutional review that appeals to a strong form of constitutionalism must necessarily be an exercise in a normatively objectionable form of hegemonic preservation. What is required as well—apart from a counter-majoritarian exercise of judicial power—is that the introduction of constitutionalism took place in bad faith, that it was driven by ulterior, non-manifest strategic motives on the part of the constitution-making elites. Such ulterior motives, what is more, should be visible in decisional outcomes that are judged to be anti-democratic and detrimental to the cause of social justice. If a constitutional court defended rights essential to democracy or rights essential to the protection of non-economic individual freedom, by appeal to a charter that is ostensibly designed to protect such rights, it would make little sense to accuse the tribunal of being engaged in hegemonic preservation even if its decisions were strongly counter-majoritarian.

Admittedly, the decision-taking of the AYM under the constitution of 1961 was at times strongly counter-majoritarian. What is more, it was concerned to protect a constitutional identity that the AYM itself understood as Kemalist. But the AYM’s conception of Kemalist republicanism, at the time, does not appear to have been in unavoidable conflict with the goal of distributive social justice or with democratic principles. The AYM’s exercise of the power of amendment review, at any rate, was concerned, for the most part, with the defense of judicial independence and the rule of law. These principles are clearly indispensable to any well-functioning democracy, whether formally constitutional or not. The claim that their judicial protection is to be regarded as hegemonic preservation clearly violates the output constraint.

7.2 Turkey and the Agency Constraint

A further important problem with applying Hirschl’s model of hegemonic preservation to Turkey—one that comes to the fore under the constitution of 1982—is that the history of amendment review in Turkey appears to show that the Turkish case fails to match the model’s picture of elite-empowerment by way of a genuine cession of decision-making authority to the judiciary. As we have seen, the AYM tried to establish an expansive conception of republicanism and to claim the power of amendment review in the 1970s. It aspired to set itself up as the supreme guardian of the constitution. This attempt, however, was thwarted, and the constitution of 1982, created under military control, explicitly stripped the AYM of the power of (substantive) amendment review, even while it came to enshrine the AYM’s expansive understanding of republican constitutional identity. While the AYM was tasked with an indirect form of constitutional guardianship—via the provisions for party-closure—the framers of the constitution apparently did not want it to become the final arbiter of Turkish political identity.

Constitutional scholars have interpreted the AYM’s restriction to procedural review of constitutional amendments as implying that parliament, acting in its capacity as constitutional legislator, can be said to hold constituent power—a view that fit well into the context of ongoing attempts to reform or even to replace the constitution of 1982 by the use of its own amendment procedure. The political history of Turkish constitutionalism, however, suggests a slightly different interpretation. It is unlikely that the military, in withholding the power of amendment review from the AYM under the constitution of 1982, meant to signal that the eternity clauses of the constitution should be understood to be at the disposal of the constitutional legislator. It seems more plausible to assume that the military meant to claim supreme constitutional guardianship for itself. The eternity clauses of the constitution of 1982, in other words, were meant to lay down the limits within which the military, as the true sovereign, in Carl Schmitt’s sense (Schmitt 1922), was willing to allow parliamentary-democratic governance. The AYM, presumably, was to take its cues from that true sovereign, in particular in exercising the power of party closure.

The AYM made a late attempt, in its controversial headscarf decision, to reclaim the power of substantive amendment review. We concede that the headscarf-crisis might plausibly be regarded as an instance of hegemonic preservation in Hirschl’s sense. However, it is revealing that this instance of hegemonic preservation took place after the military had been forced into retreat in the confrontation of the presidential candidacy of Abdullah Gül and was apparently no longer willing to exercise its role as the guardian of the Kemalist understanding of constitutional identity as aggressively as before. Its capacity to do so had, at any rate, been hollowed out by stepwise constitutional reform related to the EU accession process. Given this background, the AYM’s renewed attempt to occupy the role of guardian of the (Kemalist) constitution floundered relatively quickly and ended in political retreat. In retrospect, it is hard to see how the AYM could have succeeded. Its decision was widely perceived to conflict with the explicit constitutional limitation of amendment review to procedural review, and the AYM, at any rate, lacked the legitimacy to successfully engage in radically counter-majoritarian amendment review. As we have seen, the court’s renewed attempt to set itself up as the guardian of the constitution provoked the quick passage of an amendment that significantly reduced its factual opportunity to engage in counter-majoritarian amendment review. If this was an attempt to establish juristocracy, it appears to have failed decisively.

We believe that this outcome should not occasion surprise—if it is indeed true, as many critics claim, that the AYM, under the constitution of 1982, acted as the mere ‘attaché’ of the bureaucratic elite-coalition led by the military. A strategy of hegemonic preservation that relies on installing a juristocracy can only work in conditions where there already is a well-established and securely independent judiciary that can draw on significant resources of legitimacy. This precondition, obviously, is unlikely to be fulfilled where high courts are widely and correctly regarded as mere instruments of their hegemonic political principals. A strategy of hegemonic preservation, therefore, was probably never really feasible in post-1982 Turkey. If the principal-agent relationship between an elite and a court is based on a one-sided subordination of the court, and if it relies too heavily on being backed up by the raw political power of the principal, as seems to have been the case in Turkey after 1982, it will likely be too late to switch to the strategy of hegemonic preservation by juristocracy once the principal’s raw power begins to crumble. The reserves of judicial legitimacy that are necessary to make the strategy succeed cannot be built up quickly, and as long as its direct control is not threatened, a hegemonic political principal is unlikely to prepare for a loss of power, by allowing judicial institutions the independence that they would need to establish their legitimacy as trusted actors in their own right. After all, hegemonic political principals will prefer to exercise power themselves, instead of ceding it to the judiciary, as long as they can get away with it. We conclude that while the AYM’s decisions under the constitution of 1982 often appear to have satisfied the output constraint, the AYM’s position, with the exception of a brief transitional episode, failed to satisfy the agency constraint.

7.3 Turkey and the Baseline Constraint

It is difficult, naturally, to come to definitive conclusions with regard to whether Turkey satisfies the baseline constraint. The process of constitutional reform continues and only time will tell whether the apparent failure of juristocracy in Turkey will usher in a stable and pluralistic democracy. But there is some reason to be doubtful about the claim that Turkey satisfies the baseline constraint. If the AYM had really been engaged in a successful long-term project of hegemonic preservation, then the reduction of its control over the process of constitutional amendment after the constitutional crisis of 2007–2008 ought, presumably, to have led to a successful conclusion of the process of constitutional reform. At least so far, this hope has not been fulfilled. It stands to reason that an allegedly excessive juridification of constitutional politics in Turkey was not the sole cause of the repeated failure to enact a new constitution. The divisions among the major political parties about the content of a new constitution are more than reflections of a conflict between narrow elite interests and the will of the people. They are grounded in deep divisions among the electorate about Turkey’s constitutional future.

7.4 Turkey and Carl Schmitt

President Erdoğan, while advocating the introduction of a presidential system of government, recently stated that Turkey is already operating under a de facto presidential system. The task of constitutional reform, in the President’s view, is now to adapt the constitution to the changed political circumstances.Footnote 4 This remark is strikingly reminiscent of Georg Jellinek’s famous claim that constitutions are formed by a ‘normative power of the factual’ (Jellinek 1914, pp. 337-344). In Jellineks’ view, the legal force of the written constitution, or of the constitution as a system of positive legal norms, depends on whether it accurately reflects the true balance of political power. In Carl Schmitt’s version of the theme, the prevailing balance of power is taken to be an expression of the nation’s true political identity (Schmitt 1928, pp. 75–99). The source of constitutional legitimacy, from this point of view, is not a compromise among all major political camps, but a successful claim, on the part of one camp, to establish itself as the true embodiment of the nation and to impose its vision of constitutional identity.

In the context of Turkish constitutional history, President Erdoğan is rather more justified in adopting this perspective than his political opponents typically care to admit. After all, the military did not consider its own role of constitutional guardianship, as expressed most clearly in the original version of the constitution of 1982, to be a result of positive constitutional authorization. Rather, it held that role to rest on its unique capability to safeguard what it saw as the true political identity of the Turkish Republic. In assuming such a position, the military saw itself as defending the authoritarian project of modernization pursued by the Kemalist elites of the early republican period. Such invocations of meta-constitutional legitimacy are perhaps best understood as claims to what Schmitt understood by sovereignty (Schmitt 1922; Vinx 2015). By way of conclusion, we would like to suggest that Schmitt’s theory of sovereignty might well provide a more fruitful template for understanding Turkish constitutional history than the hegemonic preservation thesis.

Schmitt famously argued that sovereignty is the power to decide on the exception (Schmitt 1922, p. 13). This thesis must not be interpreted too narrowly. Schmitt did not mean to claim that sovereignty is a formal competence created and limited by the provisions on emergency powers that may be contained in a written constitution. What Schmitt wanted to argue, rather, is that the sovereign is the person or institution, whoever it may be, that possesses the de facto power to set legality aside in a perceived situation of existential political crisis (Schmitt 1922, pp. 13–14). Schmitt also held that there can be no stable constitution in a state that lacks a sovereign with the de facto power to suspend legality, for the reason that the dependable and predictable application of any legal norm requires a situation of social and political normality (Schmitt 1969; Scheuerman 1996; Croce and Salvatore 2013, pp. 30–45).

The situation of normality, for Schmitt, is characterized by the absence of fundamental disagreement over questions of political identity. Such disagreement, Schmitt fears, will turn appeals to (constitutional) legality into nothing more than another instrument of partisan infighting (Schmitt 1931, pp. 12–48). Stable (constitutional) legality therefore requires the manufacture of a unitary ‘concrete’ political identity whose establishment must precede the onset of constitutionally organized democratic politics (Schmitt 1928, pp. 20–36, 226–234). For Schmitt, it is the task of sovereignty, in taking the decision on the exception, to determine the contours of a polity’s identity. Schmitt recognizes that a sovereign’s decision on the exception is not declarative but constitutive: Any decision on whether there is a crisis, and one that warrants a suspension of legality, inevitably rests on a decision about what is to count as a normal or abnormal situation. And that decision can only be taken in light of a concrete understanding of political identity. The power to decide on the exception thus fuses with the power to define a community’s political identity. Once a state of exception has been declared, the sovereign, if necessary through purely discretionary action, will attempt to establish or to restore what he defines to be social normality, before the emergency ends and the rule of law resumes. It is the factual success in this political task of social normalization that validates a claim to sovereignty, not the authorization by some positive norm of constitutional law (Schmitt 1993, pp. 23–34). Sovereignty, thus, is a purely political power, though one that is necessary, according to Schmitt, to sustain stable and legitimate constitutional legality.

Schmitt’s theory of sovereignty was developed as a critique of the liberal-democratic constitutionalism of the Weimar Republic (Dyzenhaus 1997; Caldwell 1997). For scholars like Hans Kelsen, the core of the liberal-democratic constitutional project consists in the attempt to defuse political conflict through the avoidance sovereign decision and of any irrevocable commitment to a particular understanding of political identity. Kelsen believed that conflicts over political identity should be settled within the constitutionally organized democratic political process, and not through a sovereign decision that precedes constitutionally organized politics. Any democratic decision on a question of political identity, therefore, is to be valid only for the time being, since it ought to be open to be modified by future democratic constitutional change (Kelsen 1929; Vinx 2007). In Schmitt’s view, this form of liberal democratic constitutionalism is dangerous to the very existence of political community because it permits the erosion of social homogeneity and normalizes the contestation of political identity. A democratic constitutionalism that dispenses with sovereignty runs the risk of overburdening the constitutionally organized political process and of undercutting both its stability and its legitimacy (Schmitt 1931, 1932).

It is not our aim in this paper to assess whether Schmitt’s analysis of the conditions of stable democratic constitutionalism is correct. Our claim is, rather, that Turkish constitutional politics, throughout the history of the Republic, has been driven by a recognizably Schmittian political logic. More often than not, the relevant actors behaved as though they were convinced of the soundness of Schmitt’s analysis. Those who held the reins of the state have seldom been willing to submit themselves unreservedly to the rule of democratic constitutional law. Rather, they typically raised the claim to be the sole authentic representatives of what they took to be the true identity of the Turkish nation. According to the logic of Schmitt’s theory of sovereignty, such claims to representation are not dependent on formal constitutional authorization and they are not falsifiable by democratic elections. They consider the creation and the protection of a form of political belonging that must precede constitutionally organized democratic politics (Müller 2016, pp. 27–32). They are valid as long as those who raise them have the factual political power to make them succeed. A purely political claim to representation of this sort—as Schmitt’s theory of sovereignty illustrates—gives rise to an authority to defend the nation’s identity by resort to extra-constitutional means. In the light of this Schmittian logic of sovereignty, Turkish constitutional history might be understood as a continuous conflict over constitutional identity, one that has tended to take the form of a struggle over who is sovereign, over who determines when and where strict constitutional legality must give way to make room for a purely political form of guardianship of the integrity of the nation.

The Schmittian interpretation of Turkish constitutional history, at any rate, explains why it is difficult to apply the hegemonic preservation thesis to the case of Turkey. Where constitutional authority is held to rest, ultimately, on a purely political power that results from a successful claim to represent the antecedent unity of the nation, written constitutions can do little more than to reflect the outcomes of political struggles over the contours of political identity. Such constitutions are normatively weak, in the sense that they cannot long withstand dramatic changes in the balance of political power. As a result, they do not provide the surplus of legitimacy that would make it feasible for elites to engage in an opportunistic strategy of hegemonic preservation by judicial empowerment. Appeals to constitutional legality will fail to impress where there is no underlying democratic constitutional consensus among competing political camps and where there is no stable distinction between ordinary and constitutional politics. This is not to say that political actors who claim Schmittian sovereignty have no incentive to engage in constitution-making. Normatively weak constitutions that reflect political power, rather than constrain it, fulfill an important symbolic function. They ratify the meta-constitutional claim, on the part of those who have succeeded to gain political control of the process of constitution-making, to be the true representatives of the political identity of the people. One of the most salient ways for a political agent to validate the claim to be a Schmittian sovereign is to successfully exercise constituent authority. It is therefore unsurprising that constitutional politics is often hotly contested even where its outcomes can be expected to be normatively weak.

It should be obvious that Hirschl’s model of hegemonic preservation has little relevance to democratic constitutions that are normatively weak. The model takes the baseline of a stable and pluralist democratic constitutional tradition for granted and points to the danger that a political constitution built on such a tradition may come to be corrupted by economic elites that leverage the perceived legitimacy of formal constitutionalism for their own partial ends. In the context of a constitutional politics in which the competitors contend for the role of Schmittian sovereign, such a strategy is very unlikely to succeed, and we have argued that there is rather less evidence than has often been suggested that it has been systematically employed in Turkey. Attempts to explain recent Turkish constitutional politics in terms of the hegemonic preservation thesis were, in a sense, too optimistic. They assumed that Turkish constitutional politics had already left the logic of Schmittian sovereignty behind. The constitutional devices analyzed here—strong constitutional entrenchment and amendment review—have indeed been abused for deeply illegitimate hegemonic ends in the course of recent Turkish constitutional history. But these abuses are more likely to have been expressions of a (highly contested) claim to Schmittian sovereignty than instances of a strategy of hegemonic preservation through judicial empowerment. If this is true, a less entrenched constitution and a curtailment of the power of constitutional review are unlikely to suffice to consolidate Turkish democracy.