On Yaniv Roznai’s Theory of Substantive Unamendability - Y. Roznai, Unconstitutional Constitutional Amendments: The Limits of Amendment Powers, (Oxford University Press 2017), 368 p.

Introduction

There is a growing interest in ‘unconstitutional constitutional amendments’. To couch it in more general terms, this subject concerns the question of whether there are or ought to be any limitations on constitutional amending power, and whether courts can or ought to enforce those limitations. The question can be divided into two broad aspects and analysed accordingly: legality and legitimacy. ¹

Legality of constitutional amendments, in turn, has two aspects. The first is whether a given constitutional amendment is legally valid in terms of some procedural and/or substantive requirements. This is the descriptive or empirical approach. The second, normative, aspect is whether constitutional amendments ought to conform to some superior rules, principles etc. Depending on how one conceives of and formulates those superior rules, principles or values, one’s approach may be regarded as inspired by either natural law or legal positivism.

Legitimacy of constitutional amendments concerns the question of whether the merit of an amendment conforms to any particular political morality. And the legitimacy of substantive judicial review of constitutional amendments may also depend on whether an extraordinary power of the courts is compatible with constitutional democracy. The legitimacy aspect can be approached normatively as well as empirically. While the former approach would first require endorsing a particular view of political morality, legitimacy can be explained in the latter approach with respect to the political morality that a given legal/political order actually maintains or aims to achieve.

Yaniv Roznai’s recent book Unconstitutional Constitutional Amendments - The Limits of Amendment Powers ² offers a comprehensive and sophisticated analysis of the issue. One of the main goals of the book is to supply a general theory of substantive unamendability, ³ which could be applicable to all constitutional amendments in all legal systems. In line with this goal, Roznai’s comparative constitutional law account makes the book an essential source, not specifically limited to this subject, but also for studies of constitutional theory and constitutional design.

Roznai digs deeply enough regarding certain aspects of this issue that he reaches its historical roots. For example, he carefully traces the idea of unamendability, along with a rich literature and case law review, in the history of US constitutionalism. His extensive account also explores the idea and practice of implicit unamendability in many jurisdictions, from India to Colombia and many others. The book offers both empirical and normative explanations. Although Roznai mostly develops and upholds normative arguments and uses normative language, some of his arguments seem to be conceptual as well. Roznai’s methodology deserves a separate analysis, which would go beyond the limits of this short article.

As discussed in the literature, amending power can be limited both procedurally and substantively. Procedural limits refer to formal amendment mechanisms or formal requirements laid down in constitutions. Since the amending power would face no legal limitations as long as it met those requirements, the procedural aspect of the issue causes no serious theoretical conflict with the ideas of constitutionalism and democracy.

Substantive limitations are the focus and main concern of the book. These can be either explicit or implicit. Roznai extensively documents the explicit unamendable constitutional rules incorporated into many written constitutions around the world. He classifies and analyses these explicit unamendable rules in terms of various aspects: their contents, structures, and characteristics. The conclusion he reaches is that there is a growing global trend endorsing the idea of explicit unamendability. He goes on to trace the implicit unamendability in the case law of many jurisdictions – from South-East Asian countries such as India, Pakistan; to African countries such as Kenya and South Africa; and then on to Central and South American countries such as Argentina, Colombia, and Peru. Although some of the jurisdictions Roznai reviews have not acknowledged the idea of implicit limitations, he argues that there has been a migration of the idea of implicit unamendability across the globe.

Roznai next reviews the theories or approaches that attempt to account for the legality and legitimacy of substantive unamendability. He assesses and then classifies these approaches into two main categories. Substantive limitations can derive from either natural law or from international law, which includes regional law such as EU law. In Roznai’s view, these two approaches each display flaws or weakness when they attempt to elucidate the idea of substantive unamendability.

The problem with natural law is, to put it in Roznai’s own words, that

‘natural law theories seem inappropriate to serve as limitations on constitutional amendments. Even if one accepts the presupposition that binding, objective moral principles exist in every society, even those with a “minimal content” of natural law, there is no basis to regard them as the yardstick for determining the legal validity of an amendment. Such a view would necessarily blur the distinction between what the law is and what it ought to be, and would be incompatible with the nature and value of the law as a social institution providing certain measure of predictability.’ ⁴

The idea of substantive unamendability based on international law, although promising, also has a weakness; the decisive factor concerning the force of international law vis-à-vis domestic law depends ultimately on how the domestic legal system conceives of and regulates the relationship between itself and international law. To put it more succinctly, Roznai shows that, in the final analysis, domestic law will prevail in a conflict with international law. International law could triumph over domestic law concerning the legality of a constitutional amendment; however, if the domestic legal system does not recognise the superiority and direct effect of international law, this has legal significance and bearing only in the field of international law, or in the external espace juridique. For example, if a constitutional amendment is found to be incompatible with the European Convention on Human Rights by the Strasbourg Court, such a decision would have legal (and possibly political) force and consequence only in that particular field of international law and within the Council of Europe. The domestic constitutional legal order might not have a mechanism for recognising the direct effect of that particular decision, or even if it does, the domestic legal order might nevertheless have at its disposal the tools to reverse such a decision. ⁵

Finding both these supra-constitutional approaches inadequate, Roznai proceeds to develop and present his own theory, to wit that the amending power, or what he terms the secondary constituent power, has inherent limitations. He asserts that the limited nature of the amending power stems from the fact that it is a delegated power. The amending power, as a trustee or fiduciary, is bound by a framework which includes, besides the formal procedural requirements, the basic structure or constitutional identity which is drawn and determined by the primary constituent power, or ‘the people’ as the principal.

In Roznai’s account the inherently limited nature of the amending power makes it legally possible and conceivable to impose implicit limitations as well. In this scenario, the written unamendable rules play, in a sense, a declarative role that reveals the limited nature of the amending power. They can only allay a minimum of doubt concerning the idea of substantive unamendability. In short, even if there is no explicitly unamendable rule, there are always implicit limitations, or both types of limitation can coexist. As Roznai puts it,

‘[e]xplicit and implicit unamendability are not mutually exclusive; rather, they are mutually reinforcing. Explicit unamendability should thus be regarded as confirmation or a “valuable indication” that the amendment power is limited, but not as exhaustive list of limitations.’ ⁶

Once the idea of substantive unamendability is built on the inherently limited nature of delegated amending power, Roznai proposes how to interpret and comprehend constitutions. He calls this foundational structuralism, and it furnishes the normative aspect of the theory. Foundational structuralism is presented as the conceptual understanding of constitutions, and at the same time as an interpretive technique for the courts to conduct judicial review of constitutional amendments. It entails that ‘… each constitution has to be regarded as a structure in which all of its provisions are related.’ This is the structuralist component. The foundational component suggests that ‘… this structure is built upon certain pillars and foundations that fill its essence …’ ⁷ Foundational structuralism dictates three consequences: first, the amending power cannot be used to destroy the constitution; second, it cannot destroy the basic principles of the constitution; finally, it must be used in good faith. ⁸ Foundational structuralism is raised as a corollary of the limited nature of delegated amending power on the one hand, and seems to complete that account, on the other.

Roznai thinks that further explanation is needed to refine the idea of substantive unamendability based on the delegated amending power, which completes and is consistent with foundational structuralism. An important clarification submitted in this vein is that the amending power has a spectrum. The idea of a spectrum seems to hinge on the existence of different amendment mechanisms found in various constitutions, or the different amendment mechanisms and procedures envisaged for various provisions of the same constitution. The idea of a spectrum entails the notion that the inclusion of popular elements in the amending power bestows greater legitimacy on an amendment than is enjoyed by an amendment that originates exclusively from an amending power akin to ordinary legislative power. Roznai labels this distinction ‘popular amending power’ versus ‘governmental amending power’. Governmental amending power should be subject to a stricter amending process than popular amending power. For example, invocation of a referendum as an element in the amending power would bestow the amendment with greater legitimacy since it draws the amending power closer to the primary constituent power.

As to enforcement of substantive limits, Roznai states that it is one to thing to argue that there are certain inherent substantive limitations on the amending power; it is quite another to say that those limitations should be enforced by the courts. ⁹ Although limitations may be enforceable by the courts, this does not necessarily have to be the case. Different systems may opt for different mechanisms. One system could be set up granting the courts (or a specific supreme/constitutional court) the power to enforce those limits, while another might entrust supervision to political institutions. This becomes evident from the choices made in different jurisdictions. Even though Roznai seems to favour court supervision, this does not imply any politico-moral or institutional superiority. One point that needs to be underlined is that in the systems that opt for court enforcement of the limitations, judicial review of constitutional amendments can still be possible even when there are no explicit unamendable rules. ¹⁰ Obviously, this is consistent with Roznai’s view that explicit and implicit unamendability can coexist.

Critique of Roznai’s theory

In Roznai’s theory, the legality (and legitimacy) of substantive unamendability relies on the argument that the primary constituent power, while delegating a crucial part of its power to the amending power, sets limitations – either implicit, explicit, or both – on use of the amending power. The main controversy surrounding the delegation theory would seem to be that when it is invoked to account for the implicit limitations the principal is constantly changing – or in the making, as it were; let us assume that this assertion appears in the literature once every generation or so. Therefore, once the constitution is created, it is barely probable that the same intentions or will, i.e. the implicit limitations or constitutional identity, could still be attributed to the principal.

Another underlying assumption contained within the argument is that the people – the true holders of sovereignty or constituent power – can at any time change political and legal course by giving themselves a new constitution. Naturally, once the enterprise of making the new constitution is undertaken, nothing – including explicit and implicit limitations – binds or constrains the primary constituent power. Roznai further expresses the view that when the legality and legitimacy of substantive unamendability are established on the theory of delegation, the criticisms raised against substantive judicial review of constitutional amendments, such as ‘the death hand of the past’ and the ‘undemocratic’ ¹¹ nature of judicial supremacy, would be also eliminated. This argument is, in a way, similar to the idea that the legitimacy of judicial review (of ordinary laws) can be attained by relying on the possibility of amending the constitution in the event of disagreeable court decisions, ¹² or when a need arises to amend the constitution. However, what is less convincing in Roznai’s argument is that while amending a constitution is concrete and feasible, there is no (suggested) mechanism that can call the primary constituent power into play, and, in fact, no tangible mechanism could even be contemplated, since the primary constituent power itself is (taken for granted to be) outside the legal realm.

One possible solution to this difficult question could be offered by the idea of necessarily democratic use of the primary constituent power, which Roznai emphatically emphasises. Roznai believes that the constituent power ‘… belongs solely in the context of democratic theory’, and should be only thought of as ‘… the power of “the people” together to make and remake the polity.’ ¹³ In my view, for this account to hold, three related points need further substantiation: its conception of democracy; ¹⁴ its conception of ‘the people’ making use of their constituent power (as in the ethnos or demos model of the people); ¹⁵ and finally, an account concerning the relationship between the people and their representation (an accounting of collective action by the people).

The argument – that primary constituent power cannot be used other than in a democratic way – is an important conceptual argument. However, accepting the resurrection of the primary constituent power by means of revolution or a coup d’état leading to the end-product of a constitution does not fit in with this seamlessly. Therefore, the underlying argument of the book seems to be built on the assumption that the initial use of the primary constituent power must be democratic. If this is the case, the following questions inevitably arise: does the primary constituent power always emerge and reveal itself in a peaceful and democratic way, or more straightforwardly, can the people themselves use this power democratically? How can we attribute the end-product of a constitution-making process, i.e. the constitution, to the will of people? Another pertinent question: should a referendum by which a new constitution or fundamental constitutional amendment is voted on by the people be subject to a qualified majority requirement? Without substantiating these points, the book’s main arguments – the primary constituent power can always determine its body-politic, abrogate its (current) constitution, and make a new one without limitation – lose part of their theoretical strength.

The conception of the democratic nature of the primary constituent power leads one to surmise two sets of conclusions or scenarios. In the first scenario, the argument amounts to affirmation that the primary constituent power is also inherently limited by the idea of democracy. A vicious circle is then entered into because the requirement that the primary constituent power be used democratically necessarily entails at least some (procedural or substantial) rights or pre-conditions, such as freedom of speech, equal participation rights in the deliberation process of constitution-making, etc. In light of this first scenario, let us assume that a constitution-making process actually culminates in a written document but that the process somehow fell short of providing the required procedural or substantial rights, for instance by not allowing each citizen or the electorate as a whole free and equal participation in the process, or by restricting the freedom of speech (either in accordance with the then existing legal rules or de facto). Would this justify the assertion that the end-product, i.e. the constitution, is not a constitution at all, or bizarrely, that it is unconstitutional? If so, another unavoidable question arises: who, or which organ, would/should resolve this matter? Without disentangling these knotty points, the claim that the primary constituent power can at any time make a new constitution stands on its own as a presupposition or an assumption, just like Kelsen’s concept of Grundnorm. Or, Roznai’s argument could perhaps be understood as serving a legitimising function for the already existing unamendable constitutional rules. However, if those rules do not seek to protect any core constitutional democratic values, as the case could very well be, the argument suffers.

In the second scenario, the argument results in accepting the idea that the necessarily democratic use of primary constituent power makes it possible to regulate certain procedural and/or substantial requirements in advance. This can be useful in making the distinction between total and partial revision. In the case of total revision, satisfying those (procedural or substantial) conditions or rights would be enough to accept or assume the democratic use. In fact, Roznai applauds the benefits of re-emergence of the primary constituent power envisaged by some constitutions. ¹⁶ Yet, the problem with this argument seems to lie in the fact that these are only benefits – merely contingent components – and thus not necessarily required by the idea of the democratic nature of primary constituent power.

Another important point draws our attention. The way a given constitution is created in the first place is a crucial variable which can contribute to determining the spectrum of the amending power; that is to say, whether the constitution is a the culmination of democratic procedure or it is a kind of imposed constitution, by a coup d’état or by a foreign power. This distinction makes sense, as it could influence one’s perception of the spectrum of the amending power and its associated legitimacy. The substantive unamendability incorporated by a constitution that was created by a coup d’état would obviously enjoy less legitimacy, and amending it could require a less rigid legitimacy threshold.

In my view, it is extremely important to tackle and explore the question of whether a constitution can in fact have a coherent and consistent unamendable core or constitutional identity. When the answer is yes, as it is in this book, the following question would need to be whether the constitutional identity existed from the very inception of the constitution. This inquiry is both relevant and particularly critical for constitutions that are built on compromise between different (political) actors or groups within society, or for constitutions that have somehow been imposed. It is probably for this reason that substantive judicial review of constitutional amendments has, in many jurisdictions, often not been welcomed, either by those in power when a decision goes against their expectations and desires, or by their opponents when a decision supports the government’s political agenda. To illustrate this, let us take two examples: India and Turkey.

The basic structure doctrine in India was initially neither welcomed nor approved. In fact, it was for this reason that the Indira Gandhi government tried to curb the Supreme Court’s power with subsequent constitutional amendments (after Kesavananda Bharati v State of Kerala, the case that upheld the basic structure doctrine) ¹⁷ and/or by putting pressure on, or packing, the Supreme Court. Although current application of the basic structure doctrine in India has been expanded to apply to a wide variety of subjects, it initially arose out of the struggle between the Parliament and the Supreme Court on one particular issue: the scope of the right to property. ¹⁸ Given that the provisions on the right to property (and confiscation of property) were formulated by means of a compromise between the socialist groups and those prone to a more liberal economy in the Constituent Assembly, it was unclear from the outset whether the right to property indeed had always been part of the constitutional identity of India. ¹⁹ In the socio-economic context of India, some sort of socialist rhetoric was at work behind the scenes in the Constituent Assembly, as was clearly demonstrated by the regulation incorporated into Part IV (Directive Principles of State Policy) of the Constitution. ²⁰ Therefore, it remains somewhat unclear which part (Part III or Part IV) truly belonged to the basic structure of the Constitution of India or, possibly, whether both were equally so from the outset. It was only by a decision of the Supreme Court of India that called the basic structure doctrine into being, but it was only over the course of time and under the influence of many complex socio-political events (especially following the aftermath of the Emergency Period), that the doctrine was endorsed in subsequent practice.

Furthermore, it is because of the struggle between Parliament and the Supreme Court that the right to property – which was originally incorporated in the Constitution of India in Article 19 and other related provisions under Part III (the chapter dealing with Fundamental Rights) – was omitted from Article 19, which bestows a specific constitutional protection upon fundamental constitutional rights. The two main provisions of the Constitution concerning the right to property (Article 19(f) and the expropriation of private property (Article 31) were abolished from the Constitution altogether by the Forty-Fourth Amendment Act in 1978. A new provision was added to the Constitution (Article 300A) by the same Amendment Act. This implies that even though there was a right to property in the Constitution of India, it did not belong to the basic structure. Yet, in a hypothetical Supreme Court challenge to the Forty-Fourth Amendment Act, the Court’s response would likely be decisive to resolve the issue of whether the right to property falls within the basic structure.

The Turkish example – and particularly the secular character of its constitution – can be considered in a similar vein. As one of the unamendable principles, secularism has been, without a doubt, among the essential characteristics of constitutional order since the early years of the Republic of Turkey, which was founded in 1923. The founders made an unambiguous choice to form a secular state. However, no firm values were present in society that could provide an underpinning for that choice. This is why constitutional politics and law have been witness to an enduring struggle, at all ‘three track levels’ of democracy, between the various political actors/groups over the secular character of the constitutional order. The Constitutional Court’s controversial 2008 Headscarf Decision, which declared a constitutional amendment unconstitutional – and which Roznai reviews in his book ²¹ – was one of the reasons behind the equally controversial constitutional amendments of 2010, which resulted in changing the composition of the Constitutional Court. In this sense, while the principle of secularism has for quite some time constituted a significant component of Turkey’s constitutional identity, nowadays one could have doubts whether this is still the case, even though that principle is still counted among the unamendable constitutional rules.

On the other hand, if the constitutional identity or the basic structure resides in the constitution from the very beginning, this prompts the question of which interpretive technique is more appropriate to unveil that identity. If the basic structure or constitutional identity reflects the will and expression of the primary constituent power, originalism might be one of the more suitable techniques for uncovering it. Especially if a theory grants superiority to the will of the primary constituent power, originalism is expected to be among the appropriate alternatives, especially for certain vague or overly broad unamendable rules. Therefore, it could be expected that the book would address whether originalism could play a role in determining and describing the basic structure.

Roznai’s comparison, summarised above, between popular amending power versus governmental amending power – and the accompanying arguments – is vulnerable to being abused as a catalyst. Especially in the context of countries where the amendment procedure applicable to all constitutional articles includes popular elements like a referendum, the power of qualified majorities (or populist governments) could enable the government to exploit or abuse the referendum. ²² Although Roznai addresses and discusses this issue, a caveat should apply to the unconditional attribution of a higher degree of legitimacy to the referendum. Again, Turkey provides a good example for the discussion.

The Turkish Constitution has only one amendment procedure (Article 175), and it applies to all articles. Knowing this, the recent constitutional amendments should immediately attract our attention. Those amendments were approved by means of the referendum held on 16 April 2017 by a majority, albeit a slim one, of the electorate. The amendment bill had a very wide scope: among many other matters, it fundamentally changed the system from parliamentarism to presidentialism. The question that needs to be considered is whether these amendments have also brought about fundamental change in, or change of the basic structure of, the Constitution. Some scholarly opinion ²³ as well as comments made by the Venice Commission ²⁴ suggest as much. It has been argued that, through said amendments, Turkey has turned away from democracy. By granting so much power to the President, the amendments have nearly obliterated certain values at the core of constitutionalism, like the separation of powers – not to mention the core of the Turkish Constitution. By doing so, the amendments ‘make changes to the entrenched essential features of the Constitution’ and have been labelled unconstitutional. ²⁵

The discussion surrounding the cases of Turkey and India brings me to my final point. Roznai hints at the very beginning of the book that his main arguments are by and large appropriate for, and applicable to, modern constitutional systems that subscribe to the rule of law, respect for fundamental rights, the constitutional ideal of limited government, the people as the source of all governmental power, etc. ²⁶ However, this emphasis becomes blurred in the ensuing pages – the underlying notion becomes particularly ambiguous when he says that foundational structuralism is applicable to any type of constitution and unamendability. Closely connected with this, Roznai points out that his theory represents a substantivist approach to constitutionalism. Namely, that it somehow stands opposite the proceduralist approach. He seems to equate this distinction with the one between positivism and natural law, and he appears inclined to associate his foundational structuralist approach with natural law. ²⁷ If this is the case, it leads to an ambiguity that prompts a number of questions. First of all, as I have illustrated earlier, Roznai himself casts doubt on the ability of natural law to account for substantive unamendability. Second, if ‘… the delegation theory and the distinction between primary and secondary constituent powers can apply in different types of regimes…’ ²⁸ it can hardly stand as a normative theory. Understood this way, the theory would run into difficulty with regard to explicitly unamendable rules that do not seek the protection of core constitutional democratic values, as Roznai himself has shown could be the case with some constitutions. ²⁹ As a result, the entire theory might be considered by taking into account the extent to which it constitutes a framework for assessing whether a (proposed) constitutional amendment is consistent with the foundational structuralism and whether judicial interpretation of a constitutional amendment by a court is performed in accordance with the foundational structuralist understanding of the given constitution. Viewed this way, the theory would be more like a positivist theory, since it would be devoid of any particular (normative) substance.

Conclusion

Roznai provides us with a comprehensive treatment of the perplexing subject of unconstitutional constitutional amendments. His metaphoric invocation of the Sword of Damocles, which also appears on the book cover, aptly symbolises its perplexity. The book itself will come to be valued as an important resource on that issue as well as on the more general subjects of constitutional theory and constitutional design. Roznai’s profound accounting of comparative constitutional law makes the book essential reading, especially for those who study the issue of constitutional amendments. I believe Roznai’s sophisticated theory offers a promising explanation of substantive unamendability. The overall thesis of the book can be read, though, as part of a larger project. Roznai himself in fact hints at this towards the end of the book. I am confident that the parts yet to come, or perhaps already being developed – especially on the theory of primary constituent power – will prove a worthy contribution to this subtle account.