Blog of the International Journal of Constitutional Law

Closing Remarks at Symposium on “Comparative Constitutional Change: New Perspectives on Formal and Informal Amendment”

[Editor’s note: In 2014, I organized the inaugural AALS Academic Symposium. The subject of the Symposium was “Comparative Constitutional Change: New Perspectives on Formal and Informal Amendment,” and the program was held in New York City at the AALS Annual Meeting. Half of the papers have been published in the latest issue of ICON; the other half have been published in the American Journal of Comparative Law. I invited Ozan Varol to deliver closing remarks at the Symposium. What follows below is a transcript of his remarks, modified slightly to reflect recent developments. For his latest analysis on the constitutional state of play in Turkey, please see his article, Presidentialism in Turkey:  Is It Already Here?. –Richard Albert]


Ozan Varol, Associate Professor of Law at Lewis & Clark Law School

Closing remarks are supposed to find a common thread that connects all of the divergent ideas that were discussed in the symposium.  That, at first, struck me as a fairly daunting task for a set of symposium papers that run the gamut from formal constitutional amendment and replacement to informal amendment by desuetude.  But as I read the articles, the task did not seem all that daunting.  It occurred to me that a country that I am quite familiar with, Turkey, poignantly illustrates many of the principles that were discussed throughout the day. From the formation of its Constitutional Court, to the unamendable provisions in its Constitution, to their interpretation by the Constitutional Court, and to its recent attempts at constitutional amendment and revision, Turkey provides a salient case study for comparative constitutional change.

Let me begin with the Constitutional Court.

The Turkish Constitutional Court is the creation of a military coup d’etat in 1960.  As I discuss in a previous paper published in the Harvard International Law Journal, the Turkish military in 1960 deposed the authoritarian-leaning Democrat Party government with the promise to hold democratic elections and drafted a new Constitution that is widely considered the most liberal constitution in Turkish history.  The Constitution established numerous individual rights and liberties and enshrined the principles of separation of powers and checks and balances.  It also established a Constitutional Court authorized to conduct judicial review.  The formation of the Court partially supports Tom Ginsburg’s theory of judicial review as political insurance as well as Ran Hirschl’s hegemonic preservation thesis.  The military leaders, foreseeing their inevitable loss of power through democratic elections, created and empowered a sympathetic Constitutional Court to preserve their values and interests.

But the Court’s establishment also supports the alternative explanation offered in Stephen Gardbaum’s paper, which argues that separation of powers concerns are partially responsible for the spread of judicial review.  The military leaders believed that political accountability proved insufficient to check government power and prevent the one-party consolidation of power that sent Turkey down an authoritarian path in the 1950s.  As a result, they created numerous independent institutions in the 1961 Constitution to check the elected branches, including a Constitutional Court and bicameral legislature, ending the supremacy of the one-house Parliament.

The Turkish Constitution also contains a number of unamendable provisions, the subject of several symposium papers.  Article 2, titled “Characteristics of the Republic,” declares that “[t]he Republic of Turkey is a democratic, secular and social state governed by the rule of law.”  Article 4, in turn, states that “the provisions in Article 2 on the characteristics of the Republic . . . shall not be amended, nor shall their amendment be proposed.”   That provision illustrates the expressive function of constitutional amendment rules that Richard Albert discusses in his paper.  The drafters adopted that formal amendment rule in part to reflect the special status that certain principles, primarily secularism, enjoy in Turkish society and to commit future political actors to those principles.

But Article 4 also comes with the design flaw that Professor Albert describes in his article on amending the amending rules.  The drafters of the Constitution did not include Article 4 itself as one of the non-amendable articles of the Constitution.  So it is technically possible to engage in the double amendment tactic that Professor Albert describes, repeal Article 4, and then amend Article 2.  Interestingly, though, that technical design flaw was rectified by a decision of the Turkish Constitutional Court.  In one of its opinions, the Constitutional Court noted that, despite the design flaw, Article 4 of the 1982 Constitution cannot be amended, as doing so would subvert the intentions of the drafters who sought to prevent the amendment of the unamendable provisions.

Let me now illustrate how the unamendable secularism provision has operated in the context of a modern constitutional debate in Turkey.  In a pair of cases in 1989 and 1991, the Turkish Constitutional Court was asked to decide whether laws that permitted the wearing of Islamic headscarves on university campuses violated the unamendable secularism provision in the Turkish Constitution.  In 1989, university regulations prohibited the wearing of religious clothing in universities and these two laws were intended to override them.  The Court struck down the laws on the grounds that the laws violated the secularism principle in Article 2.

The Court applied an originalist methodology, reasoning that the legislation was inconsistent with the original intentions and vision of the country’s secular founder, Atatürk.  The Court was unwilling to interpret away Atatürk’s original expected application of secularism or view the secularism provision in the Constitution at a high level of generality, which may have provided more interpretive flexibility.  In so doing, the Court implicitly rejected what James Fleming describes as a “philosophic approach” to constitutional interpretation based on political or moral principles, not concrete historical practices.

This relates to a point that Professor Vicki Jackson raises in her paper on the myth of unamendability of the U.S. Constitution:  The harder it is to amend a constitution, the more significant it becomes that those of its parts dealing with general principles be sufficiently general.  Constitutional flexibility, as Tom Ginsburg and James Melton also observe, often generates the conditions for constitutional stability and can prolong the lifetime of a constitution.  But the Turkish Constitutional Court chose to adopt a fairly narrow interpretation of the unamendable secularism provisions in the Constitution, creating an inflexible constitutional rule:  no Islamic headscarves in universities.

In reaching its conclusion, the Court also held that foreign legal norms could not penetrate the Turkish constitutional shield over secularism.  That rejection illustrates the identity construction principle that Ran Hirschl discusses in his paper as to whether and why judges cite foreign law in their decision-making process.  The Turkish Constitutional Court’s attitude towards foreign law is quite similar to the Israeli Supreme Court’s attitude in the Citizenship Law/Family Unification case.  Both courts operate in societies with intense conflicts on secularism and the role of religion.  And in both nations, courts are advancing particular visions of that society’s place in the world and a particular vision of a suitable identity for that society when there are intense societal disputes in the polity on those very questions.  Interestingly, although both the Turkish and Israeli high courts are facing fairly similar constitutional conundrums, you see no citations to each other, no exchanging of notes in these two geographically proximate courts.

After the Turkish Constitutional Court struck down the laws permitting the wearing of Islamic headscarves, there was another option available to the Parliament:  amend the Constitution.  In 2008, the Parliament, led by the politically powerful Justice and Development Party (Adalet ve Kalkınma Partisi), amended the Constitution to state that no one shall be deprived of the right to higher education for whatever reason.  Although the amendment appeared neutral on its face as to religion, it was dubbed the “Headscarf Amendment” in legislative debates.    Its purpose was to override the decisions of the Constitutional Court and allow students to wear religious clothing, including the Islamic headscarf, in higher-education institutions.

The Headscarf Amendment illustrates some of the benefits of constitutional amendment that Professor Jackson discusses in her paper.  Amendment is an important check on the judicial power, especially where, as here, there is a constitutional court employing an originalist methodology to create a rather static constitutional norm. The Amendment also was consistent with the idea of constitutional imperfection.  As society evolves, constitutional norms may become anachronistic and a strict prohibition on the wearing of religious clothing in the public sphere at the foundation of the Turkish Republic may make little sense in the modern context of Turkey.  As Professor Jackson argues, constitutional amendment serves as an important safety valve to refresh the constitution and protect its democratic legitimacy.

The constitutionality of the Headscarf Amendment ended up before the Turkish Constitutional Court.  But the jurisdiction of the Court to review constitutional amendments was far from clear.  In 2008, the Court was operating under the 1982 Constitution, which was drafted after another military coup in 1980.  The numerous counter-majoritarian institutions created in the liberal 1961 Constitution, along with a system of proportional representation, gave rise to two decades of weak coalition governments and frequent power vacuums, which prompted a military coup in 1980.

The drafters of the 1982 Constitution shared Sanford Levinson’s emphasis on the clauses that constitute the Constitution of Settlement—those seemingly intellectually unchallenging clauses establishing basic institutional structures.  They blamed the Constitution of Settlement for the political problems that resulted in the aftermath of the 1960 coup and focused primarily on revising those clauses. The 1982 Constitution restricted or eliminated many of the counter-majoritarian institutions in the 1961 Constitution and converted the bicameral legislature, which was widely viewed as inefficient, to a unicameral legislature, and strengthened the presidency. The 1982 Constitution also restricted the Constitutional Court’s review of constitutional amendments only to procedural grounds.   Under the 1982 Constitution, the Court has jurisdiction to review a constitutional amendment only to ensure that the amendment garnered the requisite supermajority in the legislature and the legislature complied with debate procedures.

When the substantive challenge to the Headscarf Amendment reached the Court in 2008, the Court was confronted with the dilemma that Sam Issacharoff describes in his paper: confront the political branches or accommodate them?  The Turkish Constitutional Court chose confrontation.  Instead of kicking the case out on jurisdictional grounds, the Court established jurisdiction to review and strike down the Headscarf Amendment by holding that the legislature lacked the constitutional authority to propose the amendment in the first place.  The Court concluded that, by enacting the Headscarf Amendment, the legislature had effectively attempted to amend the unamendable secularism provision in the Constitution.  So, according to the Court, its review of the Headscarf Amendment was of the amendment’s form—not its substance.

This decision illustrates the vagueness of the unconstitutional constitutional amendments doctrine that Ros Dixon and David Landau describe in their paper.  Their article argues that judges have considerable room in using the doctrine to second-guess democratic choices.  They also point out that the doctrine can create a roadblock to the use of amendment procedure as a means of overriding court decisions deemed unreasonable by a majority of citizens.  That was essentially how the Headscarf Amendment decision was interpreted by wide segments of the Turkish population.  If a weaker form of the doctrine described by Professors Dixon and Landau had been in place, it would have been very difficult to establish that the allowance of religious clothing on university campuses is a patently unreasonable deviation from democratic norms.

In his article, Mark Tushnet argues that the doctrine of substantive unconstitutionality might frustrate proponents of fundamental change who, in response, may resort to revolution and violence.  A revolution on a minor scale occurred in Turkey in response to the Court’s decision.  In 2010, two years after the Court rendered its decision on the Headscarf Amendment, the Justice and Development Party proposed a constitutional amendment packing the Court.

In his article, Professor Issacharoff asks, “Can courts risk head on confrontations in the absence of political competition?”  In the Turkish case, the answer was a resounding no.  The Court chose to confront a politically powerful group, in the absence of any credible political opposition to that group, and was slapped with a court-packing plan. As Tom Ginsburg and Nuno Garoupa describe, the judicialization of politics was followed by the politicization of the judiciary.  The court-packing plan was a small part of a broader package of twenty-six constitutional amendments, which included many desirable constitutional reforms.  In a resounding victory for the Justice and Development Party, the Turkish voters approved the amendment package by 58% of the vote.  Although the amendment package included changes to twenty-six different provisions, the referendum required the voters to accept or reject the entire package, with no possibility to vote on individual amendments. In a working empirical study, my co-authors Nuno Garoupa, Lucia dalla Pellegrina, and I find that the 2010 amendments marked a significant, conservative shift in the ideological position of the Turkish Constitutional Court that is increasing in magnitude over time.

Riding on the coattails of its 2010 referendum victory, the Justice and Development Party pushed forward with more extensive constitutional changes.  In October 2011, the Turkish Parliament authorized a Constitutional Conciliation Commission, composed of three members from each of the four parties represented in the Parliament, to draft a new Constitution in order to revamp the 1982 Constitution.  That raised the question that was discussed today in several papers:  What is the source of the authority to write this new Constitution?

The Justice and Development Party essentially made two arguments, though they at times asserted them in inconsistent ways.  They argued that the Parliament has both the constituent power and the amendment power.  The first line of argument is that the Parliament represents the constituent power and therefore has the authority to draft and ratify an entirely new constitution.   This is very similar to the argument that the Fidesz party made in Hungary, as discussed in Kim Lane Scheppele’s paper.  Then-Prime Minister Erdoğan claimed a mandate from his revolution in the ballot box and that mandate, he argued, allowed him to revamp the entire Constitution.  Alternatively, assuming that the constituent power argument failed to persuade, he also argued that the Parliament has the amendment power and would follow the amendment procedures in the current Constitution.

Why does that difference matter?  It matters for the reasons that Professor Scheppele discusses in her article.  A body that lacks the constituent power operates within a much more constrained space.  The distinction may also matter when it comes to the unamendable provisions.  If the Parliament has the constituent power, then the unamendable provisions may not bind the Parliament.  If, on the other hand, the Parliament does not have the constituent authority and is merely relying on the amendment procedures in the existing Constitution, then the unamendable provisions should be binding.

In the constitutional negotiations, a proposed transition to a presidential system assumed center stage.  According to the AKP, a strong, “American-style” president was necessary to bring much-needed stability to Turkey and eliminate the problems that the political system suffered in the past from weak coalition governments.  The opposition argued that this asserted reason was a fig leaf.  At the time, then-Prime Minister Erdoğan was serving his third, and under his party rules, final term as Prime Minister. The opposition believed—correctly, it turned out—that Erdoğan would run for President, providing him a strong incentive to bolster the powers of the office he would soon occupy.

A similar debate occurred in Brazil in 1988.  One of the central debates in the constitutional design process there concerned the adoption of a presidential system of government.  Although the issue was quite contentious, the designers used a temporary constitutional provision to ensure that the question did not derail the design talks.  To reduce the stakes for the opponents of presidentialism, the proponents agreed to the mandatory review of the presidentialism provisions five years after their adoption. The promise of review following a five-year interim experiment with presidentialism significantly facilitated constitutional consensus building. I discuss that process, and other purposes of temporary constitutions or temporary constitutional provisions, in a paper that was published in the California Law Review.  But unlike in Brazil, in Turkey, the proposal for a presidential system, along with other politically contentious proposals, resulted in a deadlock in the constitutional commission and brought the process to a halt.

Thank you for listening and for allowing me to engage with the timely and topical papers presented in this excellent symposium.

Suggested Citation: Ozan Varol, Closing Remarks at AALS Academic Symposium on Comparative Constitutional Change: New Perspectives on Formal and Informal Amendment, Int’l J. Const. L. Blog, Dec. 9, 2015, at:


6 responses to “Closing Remarks at Symposium on “Comparative Constitutional Change: New Perspectives on Formal and Informal Amendment””

  1. Andrew Arato Avatar
    Andrew Arato

    The Turkish GNA does not have the constituent power, in the same sense as the Hungarian Parliament. It is true that it too has a 2/3 amendment rule that allows bypassing referenda if the president agrees.

    1. There are unamendable provisons. Hungary has none
    2. The constitutional court in 2008 distinguished between the original constituent power and the derived one, in its amedment review. The hungarian court has eschewed amendment review, until recently in in a very partial way, and never made such a distinction.

    With this said the process of dismantling constitutionalism, using court packing, is parallel in the two countries. But in Turkey it ha snot yet succeeded. Erdogan’s de facto presidentialism is political change only, and could have been reversed after june 7, but for the unwillingness of the nationalist mhp to enter into a coalition with the kurdish led hdp. That is a political matter, and not even iformal constitutional change.

    I have argued all this in several articles, and will now be in my forthcoming oup book, next spring.

    1. Ali Acar Avatar
      Ali Acar

      There are some issues, which may lead one to think that informal constitutional change has been indeed inserted to the Turkish constitutional system. One can reach this conclusion especially under some judgments delivered by the new (packed) Constitutional Court, created after the 2010 Referendum. For example, the (new packed) Constitutional Court has radically changed its interpretation of secularism with the so-called 4+4+4 Act, which allowed further (elective) religious curriculum (in addition to mandatory one) in the schools. The Constitutional Court did not find unconstitutional this Act, which could be easily considered so by the old Constitutional Court delivered the Headscarf (Amendment) Decision. The new interpretation of secularism developed by the packed Constitutional Court, together with some practices, like allowing public officials (even to judges) to wearing headscarves, can be in a sense considered as an informal constitutional change.

      In addition to this, it must be noted that there is no possibility that the packed Constitutional Court will conduct a substantive (or concealed behind procedural) review of constitutional amendments in the future. The Court may do so, but this time, it would be for a different aim; i.e. if the Justice and Democratic Party (AKP) loses it power in the parliament in the future, and a different majority undertakes a constitutional amendment with a view to establishing policies radically different from that of the AKP’s, the current Constitutional Court may carry out a substantive judicial review for not allowing those policies.

  2. Andrew Arato Avatar
    Andrew Arato

    Wrong. After June 7 it depended on the MHP whether there would be a coalition to save democracy. That was a political matter. If they did the de facto would have gone up in smoke.

    I did not context other informal constitutional changes, ratified by the packed TCC. After the packing it is a serious question whether Turkey is still any kind of constitutional democracy, where the difference between political and informal constitutional change still matters. One issue however remains open in an electoral authoritarian regime, as it was open in Venezuela too. If elections are still necessary, and if they have to remain relatively free, what happens to informal changes like de facto presidentialism. In my view they would still be threatened.

    You seem to agree with Erdogan, who early this Spring trumpeted the existence of a de facto presidential regime.

    But Erdogan himself seems to agree with me, since he is still actively seeking formal change in the very same d

    1. Andrew Arato Avatar
      Andrew Arato

      direction, the last word was… screwing up the impact.

    2. Ali Acar Avatar
      Ali Acar

      I did not mention presidentialism in my reply. I assume something being there, de facto, does not imply informal change. What I have tried to say by “informal change” has mostly something to do with the Constitutional Court’s position and decisions. In this sense, those (so-called) informal changes concerning the status of secularism in Turkey would hardly be changed once democracy is re-established.

      It is true that there are still some (democratic) forces to threat to those who are seeking for a formal authoritarian regime. And I don’t agree at all with anybody who is seeking to drive Turkey to authoritarianism.

  3. Andrew Arato Avatar
    Andrew Arato

    Dear Ali. I was speaking only about de facto presidentialism, not other alterations of the constitution. So it seems we agree.

    I was trying to be funny with the last two sentences, but obviously failed. No, I do not think you agree with TR Erdogan on anything.

    And no, I doubt he would agree with me on any issue, either.

Leave a Reply

Your email address will not be published. Required fields are marked *