• Specific Year
    Any

Weis, Lael K --- "Constitutional Amendment Rules and Interpretive Fidelity to Democracy" [2014] MelbULawRw 14; (2014) 38(1) Melbourne University Law Review 240

CONSTITUTIONAL AMENDMENT RULES AND INTERPRETIVE FIDELITY TO DEMOCRACY

LAEL K WEIS[*]

Fidelity to ‘democracy’ is frequently assumed to be an important evaluative criterion for selecting between competing theories of constitutional interpretation. This article interrogates that assumption by examining a widely deployed argument against progressive judicial interpretation associated with ‘formalist’ theories: the argument that confining constitutional change to the formal amendment process better respects popular self-governance than updating constitutional meaning through judicial interpretation. The article argues that taking this suggestion seriously would require either advocating for substantial reform to existing amendment procedures and practices, or rejecting constitutional entrenchment altogether — positions that are in tension with other commitments that formalist interpretive views commonly share. This analysis ultimately suggests that the preoccupation with democracy is at best a distraction from the real set of issues that guide one’s choice of interpretive approach; at worst, it allows debates about constitutional design to proceed as if they are debates about constitutional interpretation.

CONTENTS

[C]ourts cannot read the Constitution by reference to [principles] that have become accepted since the Constitution was enacted in 1900. ... [T]hey would be amending the Constitution ... without the authority of the people acting under s 128 of the Constitution.

— McHugh J, Al-Kateb v Godwin[1]

The Constitution provides both for formal amendment and judicial reinterpretation. From the earliest days of federation both means of adjustment and change have been followed, to the advantage of the Commonwealth and its people.

— Kirby J, Al-Kateb v Godwin[2]

I INTRODUCTION

Considerations of democracy often appear to play an important role in selecting between competing theories of constitutional interpretation. Nowhere is this better seen than in the familiar debate about constitutional change outside of a constitution’s formal amendment rule, as illustrated by the well-known exchange between McHugh J and Kirby J in Al-Kateb v Godwin.[3] In that case, a majority of the High Court took the position that the failure of the framers to include a right to due process in the text of the Australian Constitution when it was enacted in 1900 means that no such right can now be implied — even though most Australians today think that considerations of substantive fairness should limit government powers of detention. Thus, while acknowledging that social values had changed, that fact was found to be an impermissible source of constitutional meaning.[4]

A common criticism of so-called ‘formalist’ approaches to interpretation, such as that of McHugh J and other members of the majority, is that they are ‘undemocratic’. By ‘formalist’ I mean the family of theories that view the judicial task of constitutional interpretation as analogous to the orthodox task of interpreting ordinary legislation: to give effect to the plain meaning of the text understood in light of the intentions of its drafters.[5] The ‘antidemocratic’ critique of formalism arises from its adherents’ rejection of the ‘living tree’ thesis, or the view that constitutional meaning can and ought to evolve to reflect the values and needs of a changing society. To their critics, who I will refer to simply as ‘antiformalists’,[6] formalists illegitimately privilege the ‘dead hand’ of the past over the will of the present majority, a position that is inconsistent with the view that popular sovereignty is the source of constitutionalism.[7] This is the position taken by Kirby J in his dissenting judgment, where he contends that constitutional meaning must be updated in light of contemporary social values.[8] It is on this basis, his Honour suggests, that the Court both can and ought to find new rights in the text of the Australian Constitution.[9] He criticises the majority for failing to give effect to what is a manifest change in popular will.

Now, one response to this criticism that is available to formalists is simply to deny that the will of the present majority has anything to do with constitutional interpretation — or, indeed, with constitutionalism. For instance, a committed formalist might counter that the rule of law, not democracy, is foundational and ought to guide interpretive practice. Yet, while it is fair to say that formalists do generally privilege rule of law values, most formalists are nevertheless sensitive to the antidemocratic charge and prefer to respond in kind rather than dismiss it out of hand. It is their antiformalist opponents, they insist, who undermine democracy by allowing for constitutional change outside of the textually prescribed amendment rule. Formalists concede that the enactment of a constitution may be best understood as an act of popular will. However, they argue, that very fact implies that the sole means by which a constitution may legitimately be changed is through its formal amendment procedure.

This is precisely the strategy that McHugh J deploys in response to Kirby J. His Honour does not deny that finding a due process right in the Australian Constitution would be desirable; indeed, he goes so far as to suggest that such a result would likely reflect the preferences of ‘a great many’ contemporary Australians.[10] Nevertheless, McHugh J insists that new rights are ‘not to be inserted into our Constitution by judicial decisions’ but ‘by persuading the people to amend the Constitution’.[11] The amendment rule itself embodies the constitutional commitment to democracy; constitutional change outside of that rule, however desirable, is an illegitimate usurpation of the popular will.[12]

The objective of this article is to cast doubt on the utility of ‘democracy’ as an evaluative criterion for selecting between competing interpretive views by drawing attention to this important and under-examined tactic used to criticise defences of progressive judicial interpretation. Far from serving to highlight important differences between competing interpretive approaches, I contend, the preoccupation with democracy has served as a distraction from the real set of questions that guide one’s choice of interpretive approach and therefore warrants critical attention. The line of argument just described is familiar and widely deployed in constitutional law. Yet, the formalist claim about amendment rules has not been subject to theoretical scrutiny. Moreover, claims about interpretive fidelity to democracy on both sides of the debate have been taken at face value.

I will argue that formalists’ strategy is misplaced, largely because they have uncritically bought in to ‘democracy’ as a criterion of interpretive legitimacy. Standard formalist arguments that purport to derive democratic credibility from the founding beg the question, whereas arguments that appeal to ongoing consent are implausible given the obstacles to constitutional change that most amendment procedures present. But more importantly, the relative ease of bringing about constitutional change cannot be conflated with ‘democracy’. Any plausible argument to the effect that amendment rules make formalism ‘democratic’ requires specifying a criterion for democratic adequacy. Taking this suggestion seriously, I argue, would require either advocating for substantial reform to existing amendment procedures and practices, or rejecting constitutional entrenchment altogether — positions that are in tension with other commitments that formalists commonly share.

What this analysis shows is that debating the merits of formalist and antiformalist interpretive approaches based on their supposed ‘democratic’ credentials has obscured the more basic set of concerns that split the two camps, namely: what kind of constitutional change (and how much) is desirable? This study thus has important implications for how theories of constitutional interpretation ought to engage with issues of constitutional design. In particular, it raises serious questions about the propriety of efforts to ‘democratise’ progressive judicial interpretation, as opposed to conceding its inferiority to direct popular engagement and advocating for reforms designed to produce a more robust constitutional politics.

The article’s methodological orientation is driven by its critical objective. In developing the central argument, I will refer to a range of amendment procedures and practices across different jurisdictions. Formalist interpretive theories often purport to be about written constitutions as such.[13] Yet, the scholarship in this area has overwhelmingly taken particulars of the American constitutional system as its point of departure. It is difficult, however, to generalise positions taken on constitutional change in the United States in a manner that allows us to address the central theoretical question raised here. United States Constitution art V is notorious in prescribing a near-impossible hurdle to amendment; as a consequence, changes to the United States Constitution have overwhelmingly occurred outside of art V, most prominently through judicial interpretation. A central concern of American constitutional scholarship has thus been to address whether art V is the exclusive means by which the United States Constitution may legitimately be amended. This makes it tempting to conflate ‘amendability’ with ‘democracy’, when in fact these represent distinct sets of concerns.

By considering a broader range of amendment procedures, we are in a better position to evaluate whether a constitution’s amendment rule itself might be thought to reflect a commitment to democracy, and if so, whether that bears on our preference for formalist approaches to interpretation. As we will see, taking this idea seriously is not necessarily an ideologically divisive point of contention in the way that the issue of ‘extraconstitutional’ amendment appears to neatly divide conservatives and liberals in the art V context.

II FORMALIST THEORIES OF INTERPRETATION

An underlying premise of my argument is that there exists a relatively well-defined interpretive theory of the kind described in the introduction, ‘formalism’, the basic contours of which describe a common mode of constitutional argument. The constitutional theory literature presents two challenges in this regard. On the one hand, formalism is a large and diverse family of interpretive theories. On the other hand, the best-known forms of formalist and antiformalist argument — ‘originalism’ and ‘living constitutionalism’, respectively — are largely associated with American constitutional scholarship. Accordingly, the formalist/antiformalist debate is sometimes (mistakenly, in my view) thought to be confined to American constitutional practice. It is therefore imperative at the outset both to define the kind of view that I take to be exemplary of ‘formalism’ and to demonstrate that formalist arguments (thus understood) are indeed widely deployed.

Turning to the first preliminary point: what makes a view ‘formalist’? At the most general level, what divides formalist interpretive theories from antiformalist interpretive theories is: (1) the primacy that formalists give to the text of a written constitution as the authoritative source of constitutional meaning; and (2) formalists’ rejection of the so-called ‘living tree’ thesis in favour of some version of what I will refer to as the ‘semantic fixation’ thesis. Let us consider these in turn.

Beginning with the primacy of the text, different formalist views take this position for different reasons. Some do so for linguistic or philosophical reasons, holding that a written constitution (like other written laws) is nothing more than its text, including necessary implications from its language and structure.[14] Others do so for institutional or pragmatic reasons, holding that primacy of the text is necessary in order to ensure that the judiciary does not overstep its appropriate institutional role,[15] or holding that primacy of the text is necessary for the constitution to fulfil its proper role in structuring the legitimate exercise of political power.[16]

For present purposes, these underlying differences are unimportant. Of central importance, however, is the second feature that distinguishes formalism from antiformalism: rejection of the ‘living tree’ thesis in favour of the ‘semantic fixation’ thesis.[17] Under the living tree thesis, a constitution’s meaning can and ought to change to reflect evolving social values and needs. Under the semantic fixation thesis, by contrast, the language used in a written law continues to mean what it meant at the time of the law’s enactment.

There is some diversity among formalist views here as well, based on different understandings of what semantic fixation requires. Some formalists (‘original intent originalists’) think that semantic fixation requires discovering the subjective intentions of the law’s drafters whereas others (‘original meaning originalists’ or ‘textualist originalists’) think that it requires broader inquiries into public meaning at the time of the drafting. Still other formalists (‘strict textualists’) reject intentionalism altogether.[18] It is fair to say that subjective intention views have generated the most criticism in the literature, and have largely been discredited. On the other hand, anti-intentionalism is something of an outlier view because, in denying the relevance of a law’s being authored (and by implication, the law’s purpose) to the law’s meaning, it invites a crude ‘literalism’.[19] Therefore, for the sake of clarity — and to give the formalist position on constitutional change a fair hearing — I will focus on what I take to be the most mainstream and best theorised contemporary variant of formalism. This is the variant which holds that a written constitution must be interpreted in light of the plain meaning of its text and the objective intentions of its drafters, and which simultaneously rejects inquiries into subjective intentions regarding the application of particular rules. Henceforth, when I refer to ‘formalism’ I mean to refer to this view.

Professor Jeffrey Goldsworthy is one of the most prominent defenders of this brand of formalism. By way of illustration, then, let us consider his presentation of the view. Goldsworthy’s defence of formalism takes its orientation from orthodox methods of statutory interpretation, and in particular, from the longstanding presumption that the object of statutory interpretation is to give effect to the drafters’ intentions. This commitment to legislative intention, he suggests, is best understood in terms of the semantic fixation thesis: intention is relevant, because ‘until they are formally amended, statutory provisions mean what they meant when they were enacted’.[20] To hold that the law means something different than it meant at the time it was enacted is therefore to change the law. The same principles, argues Goldsworthy, apply to the interpretation of a written constitution.[21]

Many defenders of progressive approaches to constitutional interpretation agree with the approach to statutory interpretation thus described. It is the last move, which extends that approach to constitutional interpretation, that they resist. One way to understand this resistance has to do with the fact that constitutions are more difficult to change than statutes. Without progressive judicial interpretation, antiformalists argue, constitutions cannot keep pace with changing social needs and values. They point to examples like Al-Kateb v Godwin,[22] where the application of constitutional rules without the intervention of judicial ‘updating’ leads to results that conflict with the popular will. By contrast, statutes can be amended or repealed through ordinary legislative acts in response to public opinion of the day and therefore do not suffer from this deficit.

Formalists reject this difference between ordinary legislation and constitutions as a basis for accepting the living tree thesis. To the contrary, as Professor Goldsworthy argues, ‘[t]he main function of [an amendment] procedure is ... to prevent the constitution from becoming so out of touch with the community’s needs and values that its legitimacy and authority might be imperiled’.[23] Moreover, they accuse antiformalists of being too fatalistic about the necessity of judicial intervention to overcome outmoded constitutional rules. ‘Even an amendment procedure as difficult as Article V of the United States Constitution’, Goldsworthy conjectures, ‘would surely be galvanized into action to avoid a collapse of constitutional authority’.[24]

In the same vein, formalists argue that the existence of a constitutional amendment procedure is sufficient to save their view from the objection that it is incompatible with popular self-governance. According to formalists, amendment rules should not be thought of as constraints on the popular will; to the contrary,

[t]hose who are restricted by the exclusivity of [the amendment] procedure are government officials ... who might otherwise have had the power to alter the constitution themselves. The strongest normative argument for [formalism] therefore appeals, not to the authority of the dead hand of the past, but to the authority of the living hand of the present: ‘The people’ today, or their representatives, who have exclusive legal authority to change their own constitution.[25]

Therefore, as a corollary, ‘[t]he most powerful objection to [antiformalism] is that it usurps [the] power [of “the people”] to change their constitution’.[26] Just as changing the meaning of a statute outside of amendment or repeal by a democratically elected representative body is an illegitimate usurpation of popular will, formalists argue, so too is changing the meaning of a constitution outside of its formal amendment rule.

So much for what is meant by ‘formalism’. The second preliminary point is that formalism (thus understood) is a familiar form of constitutional argument in jurisdictions with written constitutions.[27] Formalism therefore should be neither conceived nor analysed narrowly in terms peculiar to the American debate about originalism. Indeed, as I have argued elsewhere, formalism better describes interpretive practice in Australia than it does in the United States, and in many respects formalism has a more philosophically sound foundation in the Australian constitutional system as well.[28]

Moreover, even in jurisdictions where formalism does not describe the predominant approach to constitutional interpretation, formalism nevertheless occupies a mainstream place in constitutional argument and is a standard form of critique of judicially immodest constitutional interpretation. For example, although the living tree approach is generally thought to describe the Canadian Supreme Court’s interpretive method, the normative debate about how the Canadian Charter of Rights and Freedoms (‘Charter’)[29] ought to be interpreted has been divided on formalist/antiformalist lines since its inception.[30] Scholarly critics of the Court’s progressive Charter jurisprudence have generally taken a formalist line.[31] Moreover, elements of formalism arguably still persist in the Supreme Court’s interpretive practice.[32] Dissenting judges in particular have sometimes criticised the majority for taking the ‘living tree’ metaphor too far and inserting new provisions into the Charter by judicial fiat.[33]

Similarly, although the Supreme Court of India is today regarded as one of the most ‘activist’ courts in the world, it should not be forgotten that the Court’s early approach to constitutional interpretation was extremely formalistic.[34] Indian judges have overwhelmingly been trained in the English ‘black letter law’ tradition, and the development of the Court’s constitutional jurisprudence since the 1950s has reflected a series of ongoing tensions between its formalist roots, on the one hand, and the progressive political role it has gradually assumed, on the other.[35] Moreover, while these tensions have largely been resolved in favour of antiformalism, they have not entirely vanished — as the Court itself has recently indicated.[36]

In short, while it is understandable that the vast literature on ‘originalism’ and living constitutionalism in American law journals may give the impression to the contrary, the formalist/antiformalist debate is hardly unique to the United States. Rather, the contours of that debate describe a generic dilemma about how to interpret a written constitution, namely: whether the text of a constitution is to be understood in the context of the concrete bargain struck at the time of its enactment, or whether the text of a constitution is to be set free from the conditions of its enactment to suit the needs and values of the times. To the extent that the terms of the scholarly debate have been informed by categories derived from uniquely American preoccupations, this provides greater reason for evaluating the plausibility of the formalist claim about amendment rules in a broader, comparative context.

III THE PLACE OF AMENDMENT RULES IN THE FORMALIST/ANTIFORMALIST DEBATE

A notable feature of the debate between formalists and antiformalists is that both camps have taken constitutionally entrenched amendment rules as givens — fixed points around which their theories must be formulated.[37] Neither camp has seriously engaged with the issue of how amendment procedures are best designed or how they might be reformed, and what values ought to guide those projects. However, whereas the antiformalist agenda has largely been to prescribe acceptable methods for constitutional change that work around amendment rules, the formalist agenda has been to embrace amendment rules in response to the pressure they face to defend their view against the challenge from democracy.

It is fair to say that few antiformalists have taken the formalist response seriously as an answer to that challenge.[38] Yet, given the ambivalence that most antiformalists express about alternative methods of constitutional change — judicial methods in particular — it seems worth examining formalists’ proposition about amendment rules in some detail. In order to do so, however, it is essential that we first carefully situate that proposition. The formalist appeal to amendment procedures is not a freestanding argument about the manner in which a constitution manifests a commitment to democracy, but an argument that occurs in the context of a debate between competing interpretive approaches. There are three points about that debate and its parameters that will be helpful to clarify before proceeding.

The first point is that the debate is not about constitutional change as such. Despite their apparent preference for rigidity, few (if any) formalists would seriously contend that a written constitution should stay fixed for all time. Their disagreement with antiformalists has to do with how a written constitution should be changed, and it concerns only particular kinds of constitutional change: ‘amendments’. Not all changes to a constitution count as amendments. For example, judicial interpretations that simply clarify the meaning of a provision in light of text, structure and necessary implications, and which do not alter the structure, powers or obligations of government, are generally not regarded as amendments. But it is difficult to maintain that they do not amount to change. Moreover, the use of the formal amendment procedure is not sufficient for a change to count as an amendment. For instance, some constitutions are regularly amended to codify uncontroversial interpretations of the kind just described, or to clarify the practical operation of a provision (for example, by making this explicit where it is merely implied or assumed). Although they take place through the formal amendment rule, such technical changes would not count as ‘amendments’.[39]

These are basic points on which most formalists and antiformalists can agree, even if there is disagreement at the margins. For the purpose of my analysis, these points of agreement will suffice: an ‘amendment’ is a change that alters constitutional requirements in a manner that, while still consistent with the substance of the constitution as a whole, cannot simply be inferred using standard tools of legal construction. Typically, amendment will involve changes to the manner in which political power is exercised, modifying the powers of government, creating new institutional arrangements, or redefining rights and obligations.

Formalists and antiformalists may quibble about interpretive methodology across a run of cases, but interpretative problems that involve changes of this kind raise the stakes. Here is where agreement breaks down and we find competing views about whether formal amendment procedures are the exclusive means of legitimately bringing about constitutional change. For shorthand, I will refer to changes that amount to amendments to a constitution and that occur outside of its formal amendment rule as ‘extraconstitutional amendment’.

This brings us to the second point of clarification. The debate has not focused on the problem of extraconstitutional amendment as a general matter, but on the role of the judiciary. Although there are other means by which extraconstitutional amendment can occur,[40] the debate between formalists and antiformalists is about the appropriate scope of judicial review, and it is most often framed in terms of the consistency of judicial review with democracy. On the view of antiformalists, constitutional amendment through judicial interpretation is one way that a constitution is legitimately brought in line with the popular will, and thus can be said to be democratic. This is thought to occur when the judiciary is appropriately responsive to social and political developments or to the force of public opinion.[41] On the view of formalists, by contrast, there are two, and only two, ways that a constitution gives effect to the popular will: first, through enactment and, subsequently, through interpretation of the constitution in light of the objective intentions of the framers; and second, through constitutional amendment using the textually prescribed amendment procedure.

A central task of antiformalism has thus been to demonstrate how constitutional change in the hands of the judiciary is, in effect, change in the hands of ‘the people’. Formalists, by contrast, have had very little to say about how and why permitting constitutional change exclusively through formal amendment procedures can be thought to give effect to popular rule. Formalist arguments instead tend to be formulated in the negative, focused on why allowing the judiciary to amend the constitution offends popular rule, thus shifting the argumentative burden back to antiformalists. They have not offered much by way of positive argument for how and why amendment procedures lend democratic credibility to their own view. Moreover, antiformalists have not pressed this point. This is partly because they have been preoccupied with the formalist claim about enactment: either criticising formalists’ reliance on the founding,[42] or reclaiming the significance of the founding moment.[43] But it is also because they have not needed to. As noted above, most of the scholarly debate has taken place against the particulars of the American constitutional system, which stacks the deck in favour of the antiformalist position.[44]

Here we arrive at the third and final point, which is that the debate has been confined to a problem of constitutional change that is generally thought to be specific to systems with a written constitution. It is only in such systems that the prima facie tension between popular self-governance and constitutionalism appears to emerge. Even if a constitution is grounded in popular sovereignty, the conferral of power through entrenchment constrains its future exercise. By contrast, in constitutional systems that subscribe to legislative supremacy, such as the United Kingdom and New Zealand, there is no special problem of constitutional change: that is to say, the problem of constitutional change is simply a species of the generic problem of changing existing law through the political process. In principle, then, a committed formalist could embrace the idea that constitutional commitment to democracy is most fully realised by an unentrenched constitution, and hold that an entrenched constitution with an amendment procedure is only a second-best option.

There is nevertheless an impression that formalists prefer rigidity in a constitution for rule of law based reasons, which points in favour of entrenchment. For present purposes, we need not take a position on whether formalist commitments imply a preference for entrenchment or whether this is simply a function of the context of the debate, which generally takes a written constitution as its starting point. I will return to this issue in Part V.

IV STANDARD FORMALIST ARGUMENTS BASED ON AMENDMENT RULES

Bearing this context in mind, we can now turn to the task of examining the formalist proposition about amendment rules. That proposition can be divided into two claims. The first claim is that amendments to a constitution that take place through its formal amendment rule are superior to amendments that occur through judicial interpretation because the former practice has democratic credentials that the latter lacks. This claim is less controversial. Even if we accept that a given judicial amendment results in a constitution that in substance better reflects the popular will, I suspect that we would still be inclined to prefer the state of affairs where the same amendment was made through the formal amendment process. Even so, this preference has more to do with the dubious democratic pedigree of the judiciary than with the democratic adequacy of the amendment process. Thus, while as a rejoinder to the antiformalist critique the claim does some work, it functions primarily as an argumentative burden-shifting tactic. As a response to the challenge from democracy it is incomplete at best.

The second claim is that, by virtue of permitting constitutional amendment exclusively through the formal amendment rule, formalist interpretive theories better respect democracy than antiformalist theories. This claim provides a response to the challenge, but it is undertheorised. It is also far more controversial. A key contention is that the state of affairs where a constitution is grossly out of line with the popular will — and yet constitutional change through the formal amendment rule has failed — is a superior state of affairs, on democracy grounds, to the state of affairs where the judiciary has amended the constitution to bring it in line with the popular will. Suppose, for the sake of argument, that we use significant developments in ordinary legislation as a proxy for gauging the popular will, and that we are evaluating the prospect of a judicial interpretation that would bring a particular constitutional provision in line with the standard manifest in current legislation. Even under these circumstances — which, I take it, are less radical than judicial attempts to divine the popular will by reference to other indicators of social value — formalists insist that their stance on extraconstitutional amendment reflects a commitment to democracy. It is this more controversial claim, which requires considering the democratic adequacy of the amendment process on its own terms, that I now want to examine.

This section considers two standard arguments that formalists offer for their position. The aim in doing so, to be clear, is not to offer a critique of formalist views per se. Nor is it to defend antiformalism. The aim is rather to give the formalist proposition about amendment rules the theoretical consideration it has yet to receive in order, ultimately, to see what it reveals about the contours of the familiar interpretive debate that I have described. The proposition ought to be taken seriously because, as already noted, it is evident that even antiformalists feel some discomfort with leaving significant constitutional change up to the judiciary. Assuming, therefore, that fidelity to democracy is an important criterion for the adequacy of a given theory of constitutional interpretation, if there are persuasive arguments suggesting that strictly confining constitutional amendment to formal amendment procedures adequately respects democracy, then this (at least arguably) calls into question the scholarly preoccupation with reconciling judicial amendment with popular rule. This could have potentially significant consequences for the direction of constitutional theory.

Two caveats are in order before proceeding. Firstly, I will not evaluate formalist arguments against arguments that defend judicial amendment as a first-best option. I am assuming that this is an outlier view and hence falls outside of the core formalist/antiformalist debate. Secondly, I will not evaluate formalist arguments against other extraconstitutional mechanisms for constitutional change besides judicial interpretation. That project would take us well beyond the scope of the question at issue in this article, which is focused more narrowly on the contours of the debate described in the previous section. Putting these issues aside, then, let us consider the arguments.

A Arguments from the ‘Founding’

Some formalists appear to assume that the mere existence of an amendment procedure, any amendment procedure, is enough to confer democratic credibility on their interpretive view. This assumption seems best understood as grounded in considerations that go to the formation of a constitution, or its ‘founding’. A constitution’s founding is often thought to uniquely locate popular sovereignty as the source of constitutionalism. That is because ‘founding moments’ represent episodes of intense public deliberation and engagement that are not found in everyday politics. In this respect, the founding represents a manifestation of ‘the people’ par excellence. Insofar as the resulting constitution can be regarded as constituting the body politic in this manner, the formalist claim is that the written constitution itself — read against the context of its enactment — is the foundational act of popular will for constitutional purposes. All subsequent manifestations of the popular will must operate within its requirements, and this includes changes to the constitution. The democratic credentials of amendment procedures, just like other textual provisions of the constitution, are thus derived from the founding moment. To amend the constitution outside of the formal amendment rule, the argument goes, is to displace popular sovereignty as the ultimate source of the constitution, perversely placing it in the hands of judges.

There are several difficulties with this argument. In the first place, while this kind of bootstrapping manoeuvre may have some plausibility in countries with a central founding moment, many modern democracies cannot trace their written constitution back to such an episode. The argument simply does not get off the ground in countries such as Canada, for instance, where there is no single moment that constituted the body politic. Indeed, as a result of Canada’s gradual separation from the United Kingdom and its ongoing negotiations about the nature of the federal compact, there is no single document that has the status of ‘The Constitution’. Moreover, the important constitutional changes made in 1982 (including patriation and adoption of the Charter) took place through the ordinary institutions of government. Nor does the argument work for ‘codifying’ constitutions that were enacted primarily to formalise existing structures of power,[45] for ‘practical necessity’ constitutions that were motivated predominantly by convenience and efficient coordination of government functions,[46] or for ‘post-conflict settlement’ constitutions that were drafted largely by outsiders.[47]

A second and related difficulty with this argument is that amendment procedures may reflect assumptions about the composition of the body politic that no longer hold, thus rendering amendment procedures politically contestable. In these circumstances, amendment procedures fail to preserve a constitution’s claim to democratic legitimacy, and may give rise to constitutional crises. Indeed, as Sujit Choudhry has argued, one way of thinking about the phenomenon of extraconstitutional amendment is as a response to the failure of amendment procedures to achieve constitutional ‘settlement’.[48] Choudhry’s principal example is the Québec secession crisis in Canada. In his view, the crisis illustrates how amendment rules, far from providing neutral mechanisms for the manifestation of popular will, in fact serve to constitute the popular will. Thus, the Supreme Court’s decision in Reference re Secession of Quebec[49] can be understood as a response to the fact that the procedure prescribed by the Constitution Act 1982[50] could not claim legitimate authority because it reflected politically contestable assumptions about the Canadian people.[51]

Suppose, then, that we limit the argument to countries with a central founding moment and which have a sufficiently settled constitutional politics, so as not to render the amendment procedure politically contestable in this way. Even so confined, the response still cannot succeed. The trouble is that it begs the question: it assumes that the original source of popular sovereignty is retained after enactment of the constitution and that it can be located in the amendment procedure.

The contrast with constitutional systems grounded in legislative supremacy is helpful here. In such systems, the people, through their elected representatives, retain the absolute power to make laws for themselves from generation to generation. Systems that accept constitutional supremacy, by contrast, transfer power from the people to the government. Thus, it has sometimes been thought that amendment procedures do not preserve popular sovereignty but rather relocate sovereignty in the amending body prescribed by the constitution.[52] That body is typically a special (supermajority or complex majority) configuration of the legislature, in federal systems combined with special bodies representing the interests of subgovernments. This does not, of course, rule out an argument demonstrating how an appropriately constituted amending body could stand in for the people (an issue I return to below). The point is just that this cannot be assumed as a general matter, but requires further argument.

Now, one route that some formalists take around this difficulty is to suggest that non-initiation of the amendment procedure is itself an indication of ongoing approval of the constitution. In this way, amendment procedures can be viewed as the locus of popular sovereignty, albeit in a somewhat more attenuated way. The trouble with this tactic, however, is that it invites all of the usual objections to arguments based on tacit consent. Simply put, it is not clear how much we are entitled to conclude from inaction, particularly when a constitution has a very difficult amendment procedure or when the public has no direct role in proposing amendments. Moreover, tacit consent is at best a very weak appeal to democracy, and will certainly fail on more demanding criteria of democratic adequacy. Indeed, it is precisely concerns such as these that popular constitutionalism, a dominant antiformalist approach, exploits when arguing in favour of judicial ‘ratification’ of majority views found in political movements as a means of constitutional amendment.[53] For these reasons, non-initiation alone does not substantiate the claim that permitting amendment exclusively through the formal amendment rule reflects a commitment to democracy.

B The Argument from ‘Practicability’

So far, we have seen that arguments that appeal to the pedigree of a constitution come up short. It is not enough to invoke the status of a written constitution as a founding document, and then point to the existence of an amendment rule. More plausible formalist views recognise this, and suggest that the features of amendment rules matter. In particular, amendment rules must not present insurmountable hurdles to amendment when there is a strong popular majority consensus in favour of change. That is, they must be ‘practicable’.[54]

This point about practicability is made less as a positive argument in favour of confining constitutional change to the formal amendment procedure than as a rejoinder to antiformalists and thus requires some expansion. What antiformalists are really troubled by when they object to formalist arguments on democracy grounds, formalists argue, is amendment procedures that are too difficult.[55] However, as formalists rightly note, this is a different complaint than the complaint that the formalist stance on constitutional meaning and change is incorrect. If an amendment procedure is reasonably practicable, then perhaps there is no complaint on democracy grounds at all.

In order to evaluate this response, the immediate question that must be addressed is how to evaluate practicability. Is practicability a function of principle, or an observation made from practice? Jeffrey Goldsworthy, for instance, argues that a formalist approach to interpreting the Australian Constitution is democratic on the basis that the amendment procedure provided in s 128 meets this requirement.[56] On the face of it, s 128 does not appear to present the hurdle to amendment that United States Constitution art V does.[57] It does not require a legislative supermajority to propose amendments for vote at referendum, and the referendum’s double majority popular approval requirement is a much lower threshold than art V’s requirement of approval by three-fourths of the states.

As is well known, however, in practice s 128 has rarely been used and proposals have rarely succeeded at referendum.[58] Moreover, there is evidence that its disuse is not a function of satisfaction with the constitutional system. Proposals are often poorly understood by the electorate, meaning that success turns on technicalities such as drafting and status quo bias, or on whether the electorate is adequately informed and engaged on issues of constitutional reform, rather than any broad consensus against or in favour of constitutional change.[59]

On the other hand, if we evaluate practicability in terms of frequency of use, then it is not clear that we will single out what formalists are after.[60] According to available studies, the most practicable (in the ‘frequent use’ sense) forms of amendment rules are those where proposals are nominated either by the executive or members of the legislature, and where a complex legislative majority or supermajority is required for adoption.[61] In many jurisdictions, these rules seem to be nearly as effective at constitutional change as amendment through legislative majority (that is, unentrenchment). Although maximally practicable, however, to hold that such amendment rules are therefore democratic seems dubious. Conversely, there is some evidence that amendment procedures that go a step farther than Australia’s and combine referenda with amendment proposal by popular initiative tend to produce even fewer amendments.[62] Yet, at least on the face of it, such procedures seem more democratic than procedures that require proposal by the legislature.

Even if the case could be made for equating maximal practicability with democratic adequacy, however, there are reasons why formalists in particular might be wary of doing so. The first set of reasons goes to rule of law based values, such as predictability and stability, which formalists generally regard as bedrock. Constitutions that change too frequently are thought to be undesirable, and are sometimes criticised as not being constitutions at all.[63] This is a generic point raised against unentrenchment. However, the concern is arguably greater in countries that, notwithstanding constitutional entrenchment, have deep social and cultural divides and a history of conflict and distrust.

This leads to a second set of reasons for resisting the view that practicability is a proxy for democratic adequacy. These are familiar concerns raised about the democratic credentials of pure majoritarian decision rules: permanent majorities may not adequately protect the interests of minorities; and temporary majorities may act for short-term self-interested goals rather than the long-term public benefit, particularly when they know that they are unlikely to maintain majority control. Under conditions where it is not difficult, relatively speaking, to form legislative supermajorities, these concerns may apply with equal force to amendment rules. Yet, it is hard to criticise them on these grounds if practicability is the sole criterion for democratic adequacy, because subsequent supermajorities are equally empowered to overturn those changes.

By way of example, the Supreme Court of India’s development of the basic structure doctrine to review the constitutionality of amendments can be viewed as a response to an amendment rule that has proved too practicable on both fronts.[64] In practice, the amendment rules of the Constitution of India have not imposed a significant hurdle to the passage of even quite substantial amendments, because one party (or a coalition) has been able to form a de facto supermajority for India’s first several decades of independence.[65] Thus even critics of the Court’s practice of constitutional review of amendments concede that the ‘basic structure doctrine has served a certain purpose: it has warned a fledgeling democracy of the perils of brute majoritarianism’.[66] Similar concerns have prompted other newer constitutional democracies to make some aspects of their constitutions unamendable.[67]

It might be tempting to discount these concerns as isolated to exceptional cases that do not conform to the general rule that practicability is a good proxy for democratic adequacy. But this is to concede that the practicability of an amendment procedure is not necessarily correlated with how democratic it is, but depends on other background conditions. In fact, it is relatively common for constitutions — even those of stable, established democracies — to distinguish between ‘more fundamental’ and ‘less fundamental’ elements on this basis, and to impose more difficult amendment procedures for the former. For instance, the constitutions of Germany, Italy and Norway all prohibit amendments that undermine core constitutional values: human dignity, a republican form of government, and basic principles of democracy, the rule of law and human rights, respectively.[68] Federal constitutions generally have special procedures meant to protect the political integrity of sub-units.[69] Transformative social justice constitutions, such as South Africa’s, make it more difficult to amend rights provisions and founding provisions than other provisions.[70] Finally, even unentrenched constitutions protect some elements from change; for instance, New Zealand’s constitution ‘semi-entrenches’ certain electoral rules (such as term length, apportionment, and electors).[71] To generalise: we can observe a tendency to make features of a constitution that are regarded as more central to constitutional settlement, and the legitimate exercise of political power, more difficult to amend.

Taken together, these considerations suggest that if we are to view the requirement that constitutional amendment take place only through a constitution’s formal amendment rule as reflecting a commitment to democracy, then something more than practicability must be in play. This is not to suggest that practicability is unimportant or, indeed, that heightened restrictions on amendment of the kind just discussed are ultimately reconcilable with popular rule. It is rather to suggest that focusing solely on practicability is unlikely to single out criteria for democratically adequate amendment procedures. For the formalist argument to go through, what must be shown is how amendment procedures allow the ‘popular will’ (however defined) to emerge both when and how it ought to. Put differently, it must be shown that amendment procedures do not produce too many ‘false negatives’ (producing no amendment when the popular will is in favour of change) or too many ‘false positives’ (producing amendments that are inconsistent with the popular will).[72] That means we need to give a more considered look at how amendment procedures purport to engage — and, indeed, constitute — the popular will.

V ARGUMENTS FROM DEMOCRACY

If formalist theories of interpretation are to be defended on the grounds that the stance they take on extraconstitutional amendment reflects a commitment to democracy, then we require an argument demonstrating the sense in which the amendment process is democratic. That requires, in turn, an inquiry into the features of amendment procedures — at the stage of both proposal and adoption — and how those features purport to ensure that the present generation is self-governing in the relevant sense. In this section, I turn to considerations related to how amendment procedures can be thought to give effect to popular will in the here and now.

It is helpful at this stage to make a few generalisations about the kinds of amendment rules found in contemporary constitutions.[73] Firstly, amendment rules overwhelmingly require that proposals be initiated by legislatures. Proposals generally require support from more than a simple majority of the legislature in order to be submitted for adoption, whether passage by supermajority (the most common decision rule), or a complex extra-majoritarian decision rule (for example, double majority passage). Secondly, amendment rules predominantly have two mechanisms for adoption: legislative vote, or popular referendum (the latter being increasingly more common). These, too, generally impose a more demanding decision rule than approval by a simple majority (whether of the legislature, or of the qualified voters). Thirdly, as noted in the previous section, it is common for constitutions to prescribe a more difficult amendment rule for certain fundamental elements.

My claim is that it is meaningless to analyse the democratic credentials of a given amendment procedure in the absence of some reasonably well-articulated criterion of democratic adequacy. Democracy is a perennially contested concept and there are many different ways of interpreting (and thus giving effect to) the basic principle of popular rule. For that reason alone, one cannot simply invoke democracy as if it represents an uncontroversial and widely agreed upon evaluative criterion. Moreover, the democratic credentials of a given amendment procedure, once specified, may be detached from considerations that guide our evaluation of ordinary (that is, non-constitutional) law-making in terms of its conformity with the requirements of popular self-governance. These considerations, I will ultimately suggest, cast doubt on the utility of invoking democracy as a normative criterion for selecting between formalist and antiformalist interpretive approaches. For now, the important thing to appreciate is that in the absence of a defence of a view that tells us what democracy is and requires, we cannot take procedures that channel the expression of popular will at ‘democratic’ face value. Amendment rules are no exception.

How, then, are we to evaluate amendment procedures in terms of their level of commitment to democracy? It is not possible in an article of this length to canvass all possible theories of democracy on which formalist arguments might be based. Therefore, in what follows I will consider two predominant types of theory: ‘direct democracy’ and ‘representative democracy’. Each type captures some aspects of existing amendment procedures that make them the object of praise or criticism. Moreover, despite lying at opposite ends of the spectrum in terms of the demands that they place on political institutions, each type of view has affinities with aspects of formalism. Direct democracy complements the idea that respect for the founding is the other principal source of the theory’s democratic credentials, whereas representative democracy complements the formalist critique of progressive judicial interpretation, which links popular accountability to election by the people.

It must be emphasised that both theories of democracy are presented as ideal types, and in reality they do not represent mutually exclusive sets of concerns. As a consequence, what follows is necessarily schematic. Proceeding in this way will nevertheless serve to capture the range of argumentative strategies available.

A Direct Democracy

By ‘direct democracy’ I have in mind theories that emphasise the importance of meaningful participation in law and policy-making beyond elections, typically in the form of public deliberation.[74] Theories of ‘deliberative democracy’ fall in this camp.[75] These views tend to be more demanding and are generally reform-driven in their scope; theorists of this stripe find popular disengagement from politics lamentable, and emphasise the importance of active, participatory self-governance. Although typically associated with the view that political engagement is an important good in its own right, direct democracy need not be defended on these grounds. For instance, public participation could be thought to produce more ‘correct’ outcomes as measured against some external standard, perhaps because they do a better job of selecting outcomes that maximally satisfy preferences or because they fulfil an independent criterion of the public good.[76]

If these are the general contours of the evaluative criterion, then it is clear that existing amendment rules and practices are generally not democratic. They provide few vehicles for popular engagement beyond voting, sometimes only indirectly by electing representatives who are empowered to propose and adopt amendments on behalf of the populace. The people may, of course, influence the proposals considered and adopted by their elected representatives through regular political channels. However, this falls short of the requirements for participation and deliberation that direct democracy contemplates. Moreover, even in countries where referenda are used to adopt amendments, typically no more popular engagement is required than a ‘yes’ or ‘no’ vote. Voters may not be well-informed of the issue, and voter turnout may be low — indeed, this may well be by design, in order to obscure controversial issues and make failure or passage more likely. Finally, the premise of making more fundamental aspects of the constitution less susceptible to popular control seems difficult to reconcile with this kind of view.

Formalists thus cannot appeal to existing amendment procedures to argue that their interpretive view better respects democracy by insisting that amendment take place exclusively through the formal amendment process, if by ‘democracy’ formalists mean to invoke a direct democracy based criterion. In order for the argument to go through, formalists would need to commit themselves to a radical transformation of current amendment procedures and practices.

But that might be a good thing. For one thing, commitment to this kind of project would give formalists a more powerful response to antiformalists, who deal with amendment procedures that are difficult and disengaged from public opinion by constructing their theories around them. Many antiformalists hold that the source of constitutionalism lies in popular sovereignty. Yet, they do not consider reforming amendment procedures as an alternative to judicial amendment in circumstances where the constitution has fallen out of step with public opinion. This is somewhat dissatisfying in its complacency.

Here, formalists could follow theorist Joel Colón-Ríos and other advocates of democratising amendment processes and take the position that respecting the people as the ultimate source of political power is not satisfied by a judiciary that is responsive to public opinion, but requires maximally participatory and deliberative amendment procedures.[77] To the extent that formalists are attracted to similar accounts of the significance of a constitution’s founding in order to ground other aspects of their theory, this may seem like a natural extension of the view. Moreover, it is an extension that could prove attractive to liberals and conservatives alike, thus potentially broadening the appeal of formalist approaches.

There are several reasons why formalists might be reluctant to commit themselves to this position, however. First of all, a key criticism of direct democracy is that it too readily assumes that decisions which result from direct popular participation adequately reflect public opinion. But this can be queried under conditions where there is reason to suspect the absence of, inter alia, equal voice and representation of different viewpoints, and deliberation guided by reason-giving. Indeed, some studies suggest that group deliberation, even under optimal conditions, tends to reproduce hierarchies found in society, since what counts as ‘reason-giving’ is influenced by factors related to social standing (such as gender and race).[78] Direct democracy amendment reform thus might commit formalists to other constitutional reforms aimed at promoting substantive equality in political participation. Some formalists may be concerned that such reforms would increase opportunities for judicial activism. However, even those who are not troubled by this possibility may hesitate to commit themselves to such a reform-driven project, as it takes them farther away from their core theoretical and practical objectives.

A second and more serious set of concerns relates to stability and predictability in the framework for governance. Given the emphasis that formalists place on rule of law values, they may be reluctant to expose the constitution to direct democracy. Jurisdictions with amendment procedures that approximate this more demanding requirement provide cautionary lessons here. For instance, Colón-Ríos points to several countries in Latin America that involve (comparatively speaking) substantial public engagement in constitutional amendment through the use of ‘constituent assemblies’.[79] In those countries, however, such procedures appear to be used — at least in part — as necessary expedients for power brokering in the face of threats to political stability from influential stakeholders.

For example, in 1991 Colombia convened a constituent assembly for the purpose of amending the constitution after unsuccessful peace negotiations with guerrilla groups. Although the then-existing constitution granted Congress the exclusive power to propose and adopt amendments,[80] the Colombian Supreme Court of Justice upheld the constituent assembly on the grounds of its claim to popular sovereignty and its potential to achieve peace.[81] The elected Congress was left in place to govern while reforms were being considered. Once convened, however, the constituent assembly went beyond its reform mandate, producing a new constitution.[82] When the government attempted to intervene, the constituent assembly declared itself immune to administrative review; it then ordered Congress to be dissolved and put a special legislative body in its place to take care of government until new elections could be held.[83] Leaving aside whether these extraordinary measures were justified in Colombia’s somewhat exceptional circumstances, the possibility that constituent assemblies could go well beyond their mandate — indeed, even take over government — on the grounds that they are the locus of popular sovereignty suggests that they are unlikely to be attractive to formalists.

Finally, there is a set of related concerns that arise from the risk of exposing constitutional law to interest group politics. In the United States, many state constitutions use direct democracy amendment procedures, combining the popular referendum (a relatively common method for adopting amendments) with proposal through popular ballot initiatives (which is relatively uncommon).[84] In the main, direct democracy procedures emerged during an era when distrust of government ran high, and there was a desire to curb rule by political elites. The experience of states with direct democracy based legislation nevertheless suggests that one must be cautious in assuming that amendments proposed through the popular initiative process better reflect the public interest.[85] Well-funded special interest groups are known to invest a great deal of time and money to put self-serving proposals on the ballot. Moreover, all that is generally required for a popular initiative to appear on the ballot is gathering signatures, a process that involves superficial (at best) public engagement with the amendment being proposed.

Of course, not all ballot initiatives succeed at referendum. However, referenda have their own set of difficulties which may well exacerbate some of these concerns. Here, too, sponsoring groups typically invest large sums of money to advocate for their proposals.[86] This influence is exacerbated when voter turnout is low. Moreover, even when special interest influence is not of great concern, at either stage poor drafting can undermine popular legislative intent, making it difficult for members of the public to understand the proposed amendment and, ultimately, making it difficult for courts who must interpret them.[87] This is another pitfall of the popular initiative process, as proposals are typically drafted by groups with little (if any) legislative drafting expertise. It thus cannot be said definitively that these kinds of amendment procedures, although perhaps maximally participatory, are democratic in the relevant sense, because they may deliberately mislead the public or they may otherwise produce results that are in conflict with the interests of a majority of the population.

Now, it is true that these risks are inherent in the ordinary political process as well. The point here is that formalists may well be reluctant to expose constitutional law to the same kinds of risks that ordinary law endures. One possible mitigating factor is empirical evidence suggesting that amendment by direct democracy measures is more difficult and less likely to succeed than amendment through representative bodies.[88] If this is true, then the kind of concerns cited here could prove to be less substantial than one might otherwise suspect. However, that assurance of stability and predictability comes at the cost of claiming that such procedures are democratic.

A more principled solution might be to limit direct democracy procedures to amendments that are ‘fundamental’ in character.[89] Problems with defining this category aside, the trouble is that limiting public participation in this way only addresses the problem if there is reason to suspect that such amendments are more likely to attract considered public attention, so as to mitigate the problems of concentrated minority interests that arise when the public is inattentive. But fundamentality is not an obvious proxy for matters that are likely to enliven public concern. For instance, issues concerning structural features of government often attract less attention than divisive and highly contentious rights issues. Yet, structural reforms may in fact enjoy greater consensus and have more day-to-day impact on the responsiveness and effectiveness of government, thus arguably making them more appropriate topics for direct popular involvement. The relative lack of public attention to issues concerning devolution and local governance in the United Kingdom,[90] as opposed to the high level of public attentiveness to the topic of redefining marriage, is a good example.

In summary, there are two challenges that a direct democracy based criterion for evaluating amendment procedures presents for defending the formalist claim. First, a direct democracy based criterion does not provide a ready-made response: formalists cannot simply appeal to current amendment procedures to lend democratic credibility to their view, but must advocate for reform. Second, although reform might be desirable, formalists must nevertheless address the concern that direct democracy may have undesirable consequences for other constitutional values. These considerations suggest that a direct democracy based criterion, insofar as it calls for endorsing a kind of constitutional politics, is likely to create tensions with other commitments of formalist views.

B Representative Democracy

This brings us to the second type of view that formalists might rely on to show that the amendment process is democratic. I will refer to this type of view as ‘representative democracy’. Such views see representation not as a second-best solution that tends toward the corruption and degradation of the popular will, but rather think of representative institutions as necessary for the instantiation of the popular will. Without representation, on such views, the popular will would not (strictly speaking) exist — for instance, because it would lack coherence, or because it would be too indeterminate.[91] ‘Preference aggregation’ theories of democracy that focus on the structure of voting rules and party systems fall under this heading.[92]

Representative democracy views are, in the main, far less demanding than direct democracy views because they describe the kinds of institutions (for example, voting and political parties) that we are familiar with. Subscribing to this kind of view is thus more likely to allow formalists to avail themselves of existing amendment procedures to claim their commitment to democracy. The drawback is that representative democracy views invite the criticism that they do not so much present normative criteria for democratic adequacy as offer post hoc justifications for the status quo.

Nevertheless, while this may be true of some views, it is not a fair characterisation across the board. Many representative democracy views push in favour of institutional changes designed to ensure that mechanisms of representation function as they ought to. For instance, they may advocate in favour of campaign finance regulation, public information laws, and electoral reforms on the grounds that they are necessary to ensure that policy-making is appropriately responsive to and reflective of majority preferences (however defined). As these kinds of policies indicate, proponents of representative democracy are generally forthcoming about the fact that certain background conditions are required to ensure that the popular will is effectively realised through representative institutions; indeed, this is a basic puzzle built in to the notion of representation that all such views must grapple with. By contrast, as already suggested, there is sometimes a tendency to assume that popular deliberation and participation necessarily produce democratic results. It thus might be said in favour of theories of representative democracy that they have the advantage of putting the fact that procedural rules are a manner of shaping or even constituting the popular will front and centre.

The critical question, then, is how mechanisms of representation ought to operate at the level of constitutional law-making, and whether current amendment rules can reasonably be understood to conform to the demands of a representation-based criterion. The difficulty here appears to have less to do with the ability of citizens to influence the initial proposal and consideration of amendments than it does with the powerful minority veto power over proposed amendments, often at multiple stages in the ratification process.

From the perspective of representative democracy, legislative initiation in particular appears to have several advantages over the popular initiative. Proposals submitted by legislators are typically thought to be better drafted and more carefully researched than those submitted by citizen groups. Legislators have legal drafting expertise that citizen groups typically lack. Moreover, legislators are required to take into account a wide range of preferences and viewpoints, and they are expected to make compromises in the formulation of public policy on that basis. By contrast, citizen groups are (by definition) designed to advance factional policy agendas.

When representative institutions are working properly, then, the amendment proposal process will be a function of normal democratic politics and mechanisms of accountability and responsiveness. Citizens can advocate for amendments in the same way that they advocate for other social policies, and political parties can run on the promise to bring about constitutional reform and obtain a mandate from the electorate. Proposals will be made with community consultation, including written submissions, hearings, and the circulation of discussion papers and reports. Similar considerations would presumably be in play at the stage of adoption through legislative approval as well (confining our attention, for the moment, to decision procedures that rely on legislative deliberation and voting at both stages of the amendment process).

It would thus seem that formalists could offer a respectable argument for the democratic adequacy of this aspect of amendment. The argument is not, of course, immune from objection. For one thing, the most basic mechanism of accountability — regular elections — is not as effective here because unpopular amendments cannot be undone as easily as unpopular legislation. Still, it is not implausible to imagine ways that ordinary political reforms could be used to bring procedures for constitutional amendment by legislative bodies in line with the demands of representative democracy.

The difficulty thus appears to be less with the decision procedure than with the decision rule. Current decision rules require either a supermajority or complex extra-majority vote, and decision rules for adoption in particular tend to be more onerous. Supermajority decision rules are the most common. As a result, legislative minorities hold a veto power over amendment proposals that may otherwise carry broad popular support. The only way to maintain that a minority legislative veto is consistent with popular rule would be to insist that the criterion of democracy requires refining preferences through the considered deliberation and debate of elected representatives so as to produce a near-consensus.

Leaving aside independent objections to this kind of suggestion, strengthening the element of elite deliberation in the criterion puts formalists in a difficult position because they advance similar arguments against the democratic credentials of judicial decision-making. To be sure, part of the objection in that context is the idea that judges are unelected, and therefore unaccountable. However, the objection also makes much of the broader unrepresentativeness and elite status of the judiciary as an institution. It is difficult to see how supermajority voting requirements within legislatures fare much better in this regard: there is no reason to think that a legislative minority is likely to be representative of the population as a whole.

This puts formalists in a difficult position. The proposition that legislators are superiorly placed to initiate constitutional amendments on democracy grounds is already controversial. But to insist that a minority of the legislature is superiorly placed to decide whether a constitutional amendment succeeds or fails cannot be sustained on democracy grounds, even if there were such a consultation process in place. Under circumstances where it is evident that a particular proposal enjoys strong, broad-based support across the electorate and the proposal nevertheless fails to be adopted due to the opposition of a minority faction of the legislature, we would likely conclude that mechanisms of representation have failed — not that the popular will dictated against it.[93]

One way that formalists might address this concern is by confining their argument to hybrid amendment rules, which combine legislative proposal with the popular referendum for adoption. This tactic does not, of course, fully address concerns with the decision rule. Nevertheless, the minority veto problem is arguably less troubling in the case of popular referenda because they appeal directly to citizens rather than indirectly through their representatives. The idea that near-consensus among the citizenry is required to produce the popular will for the purpose of constitutional amendment seems less patently elitist, and therefore less contentious, when applied to the public at large.

There are two problems with this move, however. First of all, insofar as the democratic appeal of popular referenda lies in tapping the unmediated expression of public opinion, they are more plausibly linked to direct democracy. As I have argued, showing that referenda meet the more demanding requirements of a direct democracy based criterion would require reforms to constitutional politics that seem unlikely to be attractive to formalists. But perhaps that point works to formalists’ advantage here. Given that referenda in systems with a legislative proposal rule are heavily mediated by legislative deliberation and debate on the initial amendment bill, might a representative democracy based justification be made?

This brings us to the second problem. A representative democracy criterion relies on the existence of preferences among the electorate, and it requires mechanisms for ensuring the responsiveness and accountability of elected representatives to those preferences. Starting with the status of preferences among the electorate on issues of constitutional reform, as discussed above, there is ample evidence from Australian experience with referenda to suggest that the argument from representative democracy is likely to falter at this step.[94] This is partly an issue with the design and implementation of referenda: the question put to vote may be inadequate to capture relevant popular views, and the question can be easily manipulated to give the artificial appearance of well-formed preferences on an issue where none exists (particularly where it is a yes/no question). But it is primarily a problem of popular engagement. Even in Australia, where compulsory voting means that low voter turnout will not defeat a referendum vote’s claim to reflect public opinion, the public needs to have enough of a sense of the issues at stake in the proposed amendment to form an opinion. In absence of popular engagement, there is simply no reason to think that a referendum vote is representative of public opinion in any meaningful sense. When voter turnout is low, this problem will be even more acute.

Responsiveness and accountability mechanisms present a challenge as well, because they are somewhat attenuated in the amendment context. A ‘no’ vote at referendum is a crude mechanism for holding legislators accountable for failure to adequately consider popular views in the formulation and drafting of a particular amendment proposal. The kinds of political mechanisms used to improve accountability in ordinary law-making, such as improved access to information and campaign finance reform, will not do here because the problem appears to be of a different kind. The problem is not one of greater transparency in decision-making, but of engaging the public on matters of constitutional reform. While granting that these difficulties do not present an insurmountable challenge for formalists, once again it is not an easy case of simply appealing to existing practices in legislative proposal systems that use referenda, but a more difficult case of advocating for reform.

This brings us to the third and final solution available to formalists who want to base their claim on a representative democracy based criterion. That solution is advocating for a majority decision rule. Assuming that the background conditions stipulated by the particular theory of representative democracy in play obtain, amendments passed by legislative majorities would clearly qualify as democratic. Or, perhaps more accurately, constitutional law-making would be at least as democratic as ordinary law-making. The people, through their representatives, would be able to revise constitutional law as they see fit to adapt to their changing needs and values.

If this is correct, then adopting a representative democracy based criterion to evaluate amendment procedures seems to provide formalists with an argument against having an entrenched constitution. To put the point somewhat more contentiously: the claim that formalism is superior to antiformalism on democracy grounds would appear to be most compelling in a system of legislative supremacy. Under circumstances where an unentrenched constitution has fallen out of step with public opinion but amendment through the legislative process has failed, the democratic challenge to the formalist position on extraconstitutional amendment has less purchase. To the contrary, it is the antiformalist position on extraconstitutional amendment that appears to be more difficult to sustain: however dissatisfying lack of political action might be, to claim that judicial amendment is a superior state of affairs on democracy grounds seems dubious under a legislative majority decision rule.

For one thing, notice that in this context we can no longer straightforwardly appeal to legislative developments as a proxy for measuring constitutional law’s approximation of the popular will (that is, because any legislative development would simply change the constitution). A more contentious measure of popular will would be required. Moreover, the antiformalist position would potentially broaden the scope for ‘judicial legislation’ across the board, since the relative difficulty in amending a constitution would no longer provide a basis for distinguishing between methods of statutory and constitutional interpretation. These considerations suggest that formalists may have an edge over antiformalists in terms of fidelity to democracy with respect to an unentrenched constitution, at least on a representative democracy based criterion.

In some respects, this conclusion is not surprising. After all, there are strong affinities between formalist methods of constitutional interpretation and orthodox methods of statutory interpretation. The stance that many formalists take on constitutional interpretation is congruent with the approach they endorse for the interpretation of ordinary legislation. Indeed, many formalists think of constitutions as statute-like: directed toward different objects than ordinary legislation, and thus different primarily in content rather than kind. These theorists reject loftier descriptions of constitutions as statements of social aspiration or value, in favour of the more modest view of a constitution as a blueprint for government.[95]

For some formalists, then, the idea that the best defence of their interpretive theory on democracy grounds is found in a system of legislative supremacy might be a welcome conclusion, reaffirming the view they have taken all along about the nature of constitutionalism.[96] For other formalists, however, the prospect of frequent constitutional amendment made possible under a system of legislative supremacy might be thought to threaten the rule of law or related structural values, such as the separation of powers and federalism, that are viewed as at the foundation of constitutionalism.[97] For theorists of this persuasion, I submit, the case for formalism would be better made by appealing directly to those values, and simply denying the antiformalist contention that a constitution ought to reflect the popular will.

VI CONCLUSION

This article has examined the claim that confining constitutional change to the formal amendment process reflects a commitment to democracy. I have shown that taking this suggestion seriously as a reason for preferring formalism over antiformalism would require formalists either to embrace a constitutional politics that is in tension with their emphasis on rule of law values, or to reject constitutional entrenchment in favour of legislative supremacy. I do not take a position here on whether either of these two reform-driven agendas is desirable. My suggestion is only that if formalists do not find these options particularly attractive, then they ought to rethink their response to democracy-based criticisms of their view. Rather than conceding the terms of the debate, they would do better to reject the idea that considerations of democracy ought to guide interpretive practices.

I conclude with some remarks on the broader lesson for constitutional theory that emerges from this analysis. ‘Democracy’ is a term that is frequently invoked in debates in constitutional law, both to define problems and stake out particular positions. However, the manner in which the term is used is often careless, if not purely rhetorical — invoked as a stamp of general approval for practices that purport to respect a vague ideal of popular self-governance.

The discussion in this article illustrates several problems with this practice. For one thing, it overlooks the complexity of democracy as a family of political theories. Consequently, there is an under-appreciation for how evaluative criteria vary between different kinds of democratic theories. A second and related problem is one of translation. It cannot and should not be assumed that the same criteria for evaluating the democratic adequacy of practices at the legislative level apply to evaluating practices at the constitutional level. Thus, as the analysis in this article suggests, although it may be convenient to assume that considerations of democracy in ordinary law-making ought to influence the mechanisms of constitutional change, this simply is not obvious.

Finally, as the formalist/antiformalist debate about extraconstitutional change amply illustrates, appeals to democracy disguise more basic disagreements about constitutional design. One would expect that antiformalists who truly think that democracy is the most important factor in evaluating constitutional change would be concerned primarily with alternatives to judicial amendment. Similarly, one would expect that if formalists were truly concerned about advancing the cause of popular self-governance in constitutional law-making, then they would take one of the two tacks suggested here. Yet, as we know, that is not how the debate has played out. This suggests that the disagreement lies elsewhere: namely, in the relative preference for a rigid, rule-based constitution that guards against the shifting tide of public opinion, as opposed to a more flexible, standard-based constitution that has traction with contemporary social values.

Unless constitutional theorists are prepared to address this set of concerns, invoking democracy as a criterion for evaluating or defending a particular interpretive approach should be avoided. In the meantime, scholarship on constitutional interpretation and judicial review will proceed with greater clarity if debates concerning issues of constitutional design proceed directly on those grounds rather than circuitously, on questions of interpretive fidelity to a vague and ill-defined normative criterion.

Bearing these remarks in mind, I will end by briefly revisiting the debate between McHugh J and Kirby J with which we began. It is noteworthy that Justice McHugh went on to advocate that the Australian Constitution be amended to include a bill of rights in order to prevent future ‘tragic’ cases like Al-Kateb v Godwin.[98] That campaign has thus far not been successful. However, given the difficulties noted with s 128, it is dubious that any conclusions about the will of the Australian people can be drawn from its lack of success. So, should a judge facing an analogous case where the Australian Constitution’s lack of a bill of rights leads to a highly undesirable result await amendment? Or, supposing it can be shown that a majority of Australians support a due process right, should the judge ‘update’ constitutional requirements?

Without attempting to answer this question, this article has demonstrated why invoking considerations of democracy is likely to be counterproductive in resolving the interpretive choice that must inevitably be made. That choice is guided in the first instance by considerations of constitutional design, such as one’s view of the purpose and function of constitutionalism, and how best to conceive of Australia’s Constitution in particular. It is in those terms, and not in terms of how best to give effect to the will of the Australian people, that the interpretive disagreement between McHugh J and Kirby J is properly understood.


[*] BA (Hons), BA (Hons) (Wash); MA, JD, PhD (Stanford); Lecturer, Melbourne Law School, The University of Melbourne. The author would like to thank Cheryl Saunders, JoelColón-Ríos, and Adrienne Stone for their helpful comments on earlier drafts. The author also gives special thanks to Martin Clark and Anna Dziedzic for their research assistance. Any errors are the author’s own.

[1] [2004] HCA 37; (2004) 219 CLR 562, 589 [62], 592 [68].

[2] Ibid 625 [178].

[3] [2004] HCA 37; (2004) 219 CLR 562.

[4] More precisely: while accepting that developments in international law may provide good evidence that social values have changed, the majority nevertheless narrowly confined the interpretive use of international law to discerning the meaning the Australian Constitution had at the time of its enactment. Therefore, post-1900 developments were deemed irrelevant.

[5] I describe what I have in mind by ‘formalist’ theories in more detail in Part II below.

[6] Antiformalism is a broad church. For my purposes, the key feature distinguishing antiformalism from formalism lies in antiformalists’ acceptance of the basic ‘living tree’ thesis. Thus, theorists such as Jack Balkin and Akhil Reed Amar who attempt to meld popular constitutionalism with originalism or textualism are appropriately classified as antiformalists, notwithstanding the fact that they may call themselves ‘originalists’ or ‘textualists’: see, eg, Jack M Balkin, Living Originalism (Harvard University Press, 2011); Jack M Balkin, Constitutional Redemption: Political Faith in an Unjust World (Harvard University Press, 2011); Akhil Reed Amar, ‘The Supreme Court 1999 Term — Foreword: The Document and the Doctrine’ (2000) 114 Harvard Law Review 26; Akhil Reed Amar, The Bill of Rights: Creation and Reconstruction (Yale University Press, 1998).

[7] See, eg, Paul Brest, ‘The Misconceived Quest for the Original Understanding’ (1980) 60 Boston University Law Review 204, 225; John Hart Ely, Democracy and Distrust: A Theory of Judicial Review (Harvard University Press, 1980) 11–12; Samuel Freeman, ‘Original Meaning, Democratic Interpretation, and the Constitution’ (1992) 21 Philosophy & Public Affairs 3, 16.

[8] Al-Kateb v Godwin [2004] HCA 37; (2004) 219 CLR 562, 624 [175], 625 [178]. This is also a position that Justice Kirby has defended in his scholarly writing: see, eg, Justice Michael Kirby, ‘Constitutional Interpretation and Original Intent: A Form of Ancestor Worship?’ [2000] MelbULawRw 1; (2000) 24 Melbourne University Law Review 1.

[9] Al-Kateb v Godwin [2004] HCA 37; (2004) 219 CLR 562, 625–6 [180].

[10] Ibid 586 [48].

[11] Ibid 594–5 [73].

[12] See, eg, Jeffrey Goldsworthy, ‘Interpreting the Constitution in Its Second Century’ [2000] MelbULawRw 27; (2000) 24 Melbourne University Law Review 677, 683–7; Justice Antonin Scalia, A Matter of Interpretation: Federal Courts and the Law (Princeton University Press, 1997) 22.

[13] In this article I follow common practice and use the term ‘written constitution’ when I mean to refer to an entrenched constitution, unless drawing a distinction from an unentrenched constitution. This is somewhat misleading, however, since most unentrenched constitutions include written law. For that reason, I shall refer to the latter as such rather than using the term ‘unwritten’.

[14] See, eg, Jeffrey Goldsworthy, ‘Originalism in Constitutional Interpretation’ [1997] FedLawRw 1; (1997) 25 Federal Law Review 1; Lawrence B Solum, ‘Semantic Originalism’ (Illinois Public Law and Legal Theory Research Papers Series No 07-24, University of Illinois College of Law, 22 November 2008) <http://papers.ssrn.com/abstract_id=1120244> .

[15] See, eg, Raoul Berger, Government by Judiciary: The Transformation of the Fourteenth Amendment (Harvard University Press, 1977); Robert H Bork, ‘Neutral Principles and Some First Amendment Problems’ (1971) 47 Indiana Law Journal 1; Justice Scalia, above n 12.

[16] See, eg, Larry Alexander, ‘Of Living Trees and Dead Hands: The Interpretation of Constitutions and Constitutional Rights’ (2009) 22 Canadian Journal of Law and Jurisprudence 227; Randy E Barnett, Restoring the Lost Constitution: The Presumption of Liberty (Princeton University Press, 2004).

[17] I adopt this terminology from Solum’s ‘fixation thesis’: see Solum, above n 14, 2–4, 59–67.

[18] See Frederick Schauer, ‘Formalism’ (1988) 97 Yale Law Journal 509, 535, where Schauer equates formalism with ‘rulism’ or ‘literalism’. It should be noted that Schauer’s normative defence of formalism is modest and cautious, intended only to demonstrate some of its virtues (against the view’s detractors).

[19] Among the most influential critiques of both views is Brest, above n 7.

[20] Goldsworthy, ‘Originalism in Constitutional Interpretation’, above n 14, 9.

[21] Ibid 12–19.

[22] [2004] HCA 37; (2004) 219 CLR 562.

[23] Jeffrey Goldsworthy, ‘The Case for Originalism’ in Grant Huscroft and Bradley W Miller (eds), The Challenge of Originalism: Theories of Constitutional Interpretation (Cambridge University Press, 2011) 42, 60.

[24] Ibid.

[25] Ibid 57.

[26] Ibid.

[27] One must, of course, be cautious in drawing broad generalisations. For the purpose of the argument I advance in this article, it is sufficient that formalism is an especially familiar form of argument in countries that inherited the British common law tradition, and along with it the method of statutory interpretation described above. The familiarity of formalist arguments in these countries is, perhaps, unsurprising when many of the original constitutions in question were acts of British Parliament, thereby raising the issue of whether methods of statutory interpretation similarly apply to the constitution. In addition to Australia, Canada, and India (discussed below), there is recent scholarship on formalist modes of constitutional interpretation in Malaysia and Singapore: see Yvonne Tew, ‘Originalism at Home and Abroad’ (2014) 52 Columbia Journal of Transnational Law 780.

[28] Lael K Weis, ‘What Comparativism Tells Us about Originalism’ (2013) 11 International Journal of Constitutional Law 842. It bears noting that Professor Goldsworthy pioneered what is now hailed as the ‘new originalism’ well before the view gained ascendency in American constitutional scholarship, as acknowledged in Solum, above n 14, 19.

[29] Canada Act 1982 (UK) c 11, sch B pt I.

[30] In the early days of the Charter, Justice Bertha Wilson described the two interpretive choices as follows: the ‘textual approach’ which ‘relies on accepted techniques of statutory interpretation’, and the ‘contextual approach’ which ‘views the text of the Charter as identifying certain basic values’: Justice Bertha Wilson, ‘Decision-Making in the Supreme Court’ (1986) 36 University of Toronto Law Journal 227, 245, 247; see generally at 245–8. As the Court’s Charter jurisprudence developed along progressive lines, formalist criticism emerged. For example, a relatively mainstream criticism of the Supreme Court’s interpretation of the guarantee of ‘fundamental justice’ under s 7 of the Charter in Reference re BC Motor Vehicle Act [1985] 2 SCR 486 is that those words were deliberately chosen by the drafters to avoid American-style substantive due process jurisprudence; yet, this was the interpretation it was given notwithstanding the Court’s acknowledgement of the drafters’ intent: see, eg, Peter W Hogg, Constitutional Law of Canada (Thomson Carswell, 5th ed, 2007) vol 2, 805–6.

[31] For example, in a well-known article by Professors Robert Hawkins and Robert Martin, the authors criticise the ‘contextual’ approach advocated by Justice Wilson on the grounds that it is ‘anti-democratic’, allowing judges to ‘amend’ the constitution in defiance of the constitutional bargain struck in the 1980s: Robert E Hawkins and Robert Martin, ‘Democracy, Judging and Bertha Wilson’ (1995) 41 McGill Law Journal 1, 12–19. They further argue in favour of the ‘textual’ approach she describes: at 10–11. For a general discussion of scholars who criticise the Court along formalist lines, see Kent Roach, The Supreme Court on Trial: Judicial Activism or Democratic Dialogue (Irwin Law, 2001) 74–80.

[32] In a recent article, Bradley Miller has sought to complicate the commonplace understanding of the Court’s constitutional practice by illustrating that original meaning of constitutional provisions has some interpretive purchase: Bradley W Miller, ‘Beguiled by Metaphors: The “Living Tree” and Originalist Constitutional Interpretation in Canada’ (2009) 22 Canadian Journal of Law and Jurisprudence 331. Miller suggests that invoking original meaning occurs both in majority judgments that depart from precedent, and in dissenting judgments:at 346–7.

[33] See, eg, Reference re Remuneration of Judges of the Provincial Court of Prince Edward Island [1997] 3 SCR 3, 180–1 [314]–[316] (La Forest J); R v Prosper [1994] 3 SCR 236, 287 (L’Heureux-Dubé J).

[34] For instance, in its first constitutional rights judgment, Gopalan v Madras [1950] INSC 14; [1950] SCR 88, 129, Kania CJ stated that ‘the Courts are not at liberty to declare an Act void because in their opinion it is opposed to a spirit supposed to pervade the Constitution but not expressed in words’. In that case, the Court gave a narrow interpretation to art 21 of the Constitution of India based on its text and original meaning. As was the case with s 7 of the Canadian Charter, art 21 was self-consciously drafted to avoid invoking American notions of substantive due process, using the language ‘procedure established by law’ rather than ‘due process of law’: at 107–8. The Court later departed from this view and adopted the substantive due process reading, holding that art 21 ought to be interpreted according to principles of natural justice: Maneka Gandhi v India [1978] 2 SCR 621. For an excellent discussion of the progression between these cases and the conflict between these competing interpretive approaches, see Manoj Mate, ‘The Origins of Due Process in India: The Role of Borrowing in Personal Liberty and Preventive Detention Cases’ (2010) 28 Berkeley Journal of International Law 216, 232–53.

[35] See generally S P Sathe, Judicial Activism in India (Oxford University Press, 2nd ed, 2002) ch 2.

[36] In Koushal v NAZ Foundation (Unreported, Supreme Court of India, Singhvi and Mukhopadhaya JJ, 11 December 2013) [52], the Court indicated suspicion about the use of foreign precedent to expand the categories of persons protected from discrimination under art 15. It is noteworthy that some of the litigants in that case advanced formalist arguments, ‘refer[ring] to the constitutional assembly debates on Article 15 to show that the inclusion of sexual orientation in the term “sex” was not contemplated by the founding fathers’ and ‘stress[ing] that Courts, by their very nature, should not undertake the task of legislating’: at [16.14]. The Court did not go quite that far in its judgment, but the result nevertheless suggests that formalism still has currency in constitutional argument.

[37] This is an important observation made by Joel I Colón Ríos, ‘The Counter-Majoritarian Difficulty and the Road Not Taken: Democratizing Amendment Rules’ (2012) 25 Canadian Journal of Law and Jurisprudence 53. As will become clear, my own view is that the role amendment procedures have played in the debate between formalists and antiformalists, and the United States Constitution-centrism of that debate, help explain why theorists have given so little attention to these normative issues.

[38] See, eg, Mitchell N Berman, ‘Originalism is Bunk’ (2009) 84 New York University Law Review 1; Cass R Sunstein, ‘Justice Scalia’s Democratic Formalism’ (1997) 107 Yale Law Journal 529.

[39] This is most likely to occur in jurisdictions with more detailed and code-like constitutions; Germany and India are two examples: see Donald P Kommers, ‘Germany: Balancing Rights and Duties’ in Jeffrey Goldsworthy (ed), Interpreting Constitutions: A Comparative Study (Oxford University Press, 2006) 161, 171–2; Mahendra Pal Singh, ‘India’ in Dawn Oliver and Carlo Fusaro (eds), How Constitutions Change: A Comparative Study (Hart Publishing, 2011) 169, 172–85.

[40] For instance, through enactment of comprehensive regulatory statutes that revise the background or baseline against which constitutional provisions operate. See generally William N Eskridge Jr and John Ferejohn, A Republic of Statutes: The New American Constitution (Yale University Press, 2010).

[41] There are a number of different ways this has been theorised. Representative views include: Bruce Ackerman, We the People: Foundations (Belknap Press, 1991); Barry Friedman, The Will of the People: How Public Opinion Has Influenced the Supreme Court and Shaped the Meaning of the Constitution (Farrar, Straus and Giroux, 2009).

[42] See, eg, Brest, above n 7.

[43] This is the thrust of recent work by Jack Balkin and Akhil Reed Amar: see above n 6.

[44] This is due both to the extraordinarily difficult amendment procedure set out in United States Constitution art V and the United States Supreme Court’s significant role in reordering social and political relationships.

[45] The Swedish constitution, which consists of four distinct written and entrenched fundamental laws, is one example. The Regeringsformen [Instrument of Government 1974] [Sveriges Riksdag trans], one of the four constitutional documents, codified conventions and understandings based on previous constitutional documents that had emerged over the past 50 years regarding the parliamentary system and the role of the monarch.

[46] Australia is an excellent example: Cheryl Saunders, The Constitution of Australia: A Contextual Analysis (Hart Publishing, 2011) 9–13, 18–21. The move toward Federation was driven largely by practical necessities owing to geographical considerations and not by a distinct sense of nationhood. The Commonwealth of Australia Constitution Act 1900 (Imp) 63 & 64 Vict, c 12 was not patriated until 1986: see Australia Act 1986 (Cth); Australia Act 1986 (UK) c 2.

[47] Japan and Germany, respectively, are two well-known and successful examples of constitutional democracies whose fundamental laws were drafted by outsiders.

[48] Sujit Choudhry, ‘Ackerman’s Higher Lawmaking in Comparative Constitutional Perspective: Constitutional Moments as Constitutional Failures?’ (2008) 6 International Journal of Constitutional Law 193, 197.

[49] [1998] 2 SCR 217. The question before the Court was whether Québec could unilaterally secede from Canada, and whether secession could occur without a prior constitutional amendment. The Court held that if a ‘clear majority’ of Québec’s population supported secession at referendum, Canada would be obligated to enter negotiations: at 221.

[50] Canada Act 1982 (UK) c 11, sch B.

[51] Choudhry, above n 48, 221. Briefly, this is due to the fact that constitutional patriation arrangements with the United Kingdom — which included the amending formulas — were adopted without Québec’s consent. For a discussion in the context of constitutional change in Canada, see Tsvi Kahana, ‘Canada’ in Dawn Oliver and Carlo Fusaro (eds), How Constitutions Change: A Comparative Study (Hart Publishing, 2011) 9, 36–9.

[52] For example, this concern was raised in American legal scholarship in the first half of the 20th century, of a system that clearly has both a central founding moment and a relatively settled constitutional politics: Lester Bernhardt Orfield, The Amending of the Federal Constitution (University of Michigan Press, 1942) ch 5; Max Radin, ‘The Intermittent Sovereign’ (1930) 39 Yale Law Journal 514, 523–5. On this view, the reassertion of popular sovereignty occurs only when the constitution is replaced.

[53] See especially Ackerman, We the People, above n 41.

[54] Goldsworthy, ‘The Case for Originalism’, above n 23, 57.

[55] Ibid 57–60.

[56] Ibid 51–2.

[57] Briefly, s 128 requires that proposed amendments be passed by a majority of Parliament (either a majority in both houses or twice in one house), and then submitted to a referendum. At referendum, proposed amendments must be approved by a majority of the electorate and by a majority of the states. In addition, proposed amendments that change a state’s boundaries or its representation in Parliament must pass by majority in the affected state.

[58] To date, Australia has had 44 referenda with only 8 successful votes. The last successful amendments were made in 1977. In all but one of the eight referenda attempted since 1977, the vote failed in every state. The exception was in 1984 on a proposed reform to Senate terms, which was only successful in Victoria and New South Wales: see John M Williams, ‘The Constitutional Amendment Process: Poetry for the Ages’ in H P Lee and Peter Gerangelos (eds), Constitutional Advancement in a Frozen Continent: Essays in Honour of George Winterton (Federation Press, 2009) 1, 1–3.

[59] See Chief Justice Murray Gleeson, ‘The Shape of Representative Democracy’ [2001] MonashULawRw 1; (2001) 27 Monash University Law Review 1, 2; Richard McGarvie, ‘The Wisdom of Hindsight — The 1999 Republican Referendum — Lessons for the Future’ [2001] UNDAULawRw 3; (2001) 3 University of Notre Dame Australia Law Review 11; Graeme Orr, ‘The Conduct of Referenda and Plebiscites in Australia: A Legal Perspective’ (2000) 11 Public Law Review 117; Cheryl Saunders, ‘The Parliament as Partner: A Century of Constitutional Review’ (Research Paper No 3, Parliamentary Library, Parliament of Australia, 2000).

[60] Cf Zachary Elkins, Tom Ginsburg and James Melton, The Endurance of National Constitutions (Cambridge University Press, 2009). The explanatory theory that Elkins, Ginsburg and Melton defend holds that the relative ‘flexibility’ of a given constitution is a predictor of its relatively greater endurance: at 81–3. In defining the relevant criterion of flexibility for testing the model’s hypothesised correlation, the authors acknowledge that neither the apparent (facial) ease of a particular amendment rule nor the actual rate of amendment are sufficient: at 100–1.

[61] See Donald S Lutz, ‘Toward a Theory of Constitutional Amendment’ in Sanford Levinson (ed), Responding to Imperfection: The Theory and Practice of Constitutional Amendment (Princeton University Press, 1995) 237; Astrid Lorenz, ‘How to Measure Constitutional Rigidity: Four Concepts and Two Alternatives’ (2005) 17 Journal of Theoretical Politics 339; Bjørn Erik Rasch and Roger D Congleton, ‘Amendment Procedures and Constitutional Stability’ in Roger D Congleton and Birgitta Swedenborg (eds), Democratic Constitutional Design and Public Policy: Analysis and Evidence (MIT Press, 2006) 319. But see Elkins, Ginsburg and Melton, above n 60. Although the authors are sceptical of actual amendment rate as an adequate measure of constitutional flexibility — a conclusion that one might be tempted to draw from the work of Lutz and others — they appear to accept the basic empirical point, namely: that there is at best a weak correlation between perceived versus actual difficulty of a given amendment rule: at 76, 100–1. Accordingly, Elkins, Ginsburg and Melton factor information about amendment rate into their model’s definition of flexibility.

[62] Lutz, above n 61. One difficulty with generalisations here is that few constitutions have such proposal rules. Moreover, as noted by James Melton, most empirical work in this area has insufficiently isolated the effect of popular involvement on amendment rate at the proposal versus approval stage: James Melton, ‘Constitutional Amendment Procedures: A Summary and Critique of Existing Measures’ (Paper presented at LUISS Guido Carli University, 19 June 2012).

[63] See, eg, Elkins, Ginsburg and Melton, above n 60, 82, 209; Kathleen M Sullivan, ‘What’s Wrong with Constitutional Amendments?’ in Alan Brinkley, Nelson W Polsby and Kathleen M Sullivan, New Federalist Papers: Essays in Defense of the Constitution (Twentieth Century Fund, 1997) 61.

[64] Some provisions may be amended by simple majority. For example, new states may be formed by simple majority, and Legislative Councils in states and certain territories may be created or abolished by simple majority: Constitution of India arts 4, 169, 239A. Most provisions require, however, that amendments pass by an absolute majority of both houses, and by two-thirds of the members present and voting in each house. In addition, there are special procedures for amendments affecting the states: at arts 368(1)–(2).

[65] For a survey of some of these significant amendments, see Raju Ramachandran, ‘The Supreme Court and the Basic Structure Doctrine’ in B N Kirpal et al (eds), Supreme but Not Infallible: Essays in Honour of the Supreme Court of India (Oxford University Press, 2000) 107, 109–17; Singh, above n 39, 174–78, 179–80. The Constitution of India has also undergone many minor administrative amendments which have not been particularly controversial or divisive, meaning they would have been likely to pass with relative ease irrespective of the composition of government: see Singh, above n 39, 169, 175–6.

[66] Ramachandran, above n 65, 130.

[67] See, eg, Ustav Bosne I Hercegovine [Constitution of Bosnia and Herzegovina] art X (rights and freedoms); Constituciones de Honduras [Constitution of Honduras 1982] art 374 (form of government; national territory; and presidential term limits and certain eligibility requirements); Türkiye Cumhuriyeti Anayasası [Constitution of Turkey 1982] art 4 (republican form of state; democratic, secular and social state governed by the rule of law; indivisibility of state; national language; flag; anthem; and capital). See generally Richard Albert, ‘Constitutional Handcuffs’ (2010) 42 Arizona State Law Journal 663; Aharon Barak, ‘Unconstitutional Constitutional Amendments’ (2011) 44 Israel Law Review 321. All translations of constitutions referred to in this article are by the Comparative Constitutions Project, unless otherwise specified.

[68] See Grundgesetz für die Bundesrepublik Deutschland [Basic Law of the Federal Republic of Germany] art 79(3); Costituzione della Repubblica Italiana [Constitution of Italy 1948] art 139; Kongeriket Norges Grunnlov [Constitution of Norway 1814] s 112. Norway’s Constitution prohibits amendments that alter the ‘principles’ and ‘spirit’ of the Constitution; while referencing the specific principles listed in s 2 (‘State Values’), s 112 is potentially much broader than that.

[69] See, eg, Australian Constitution s 128; Constitution of India art 368(2).

[70] Constitution of the Republic of South Africa Act 1996 (South Africa) s 74. Under s 167(4)(d) the Constitutional Court is authorised to rule on the constitutionality of any amendment: see generally Bernard Bekink, ‘Balancing Constitutional Stability and Flexibility: An Evaluation of the Constitutional Amendment Procedures’ [2004] Journal of South African Law 657,661–4.

[71] Electoral Act 1993 (NZ) s 268. The supermajority requirement for amendment, of course, can itself be amended by a simple majority.

[72] Bruce Ackerman, ‘Higher Lawmaking’ in Sanford Levinson (ed), Responding to Imperfection: The Theory and Practice of Constitutional Amendment (Princeton University Press, 1995) 63, 85. As discussed below, formalists generally appear to be more concerned about false positives, whereas antiformalists (such as Ackerman) are more concerned about false negatives.

[73] The best empirical surveys available are those listed in above n 61. Although there are limitations to advancing an argument that generalises across the run of amendment rules, the dominant patterns identified nevertheless serve to highlight the principal challenges for the claim under consideration.

[74] For a survey of some of these views, see Archon Fung and Erik Olin Wright (eds), Deepening Democracy: Institutional Innovations in Empowered Participatory Governance — The Real Utopias Project IV (Verso, 2003); James Bohman and William Rehg (eds), Deliberative Democracy: Essays on Reason and Politics (MIT Press, 1997).

[75] There is not a perfect mapping between the two, however. Some theorists of deliberative democracy reject direct democracy in favour of representative democracy on the grounds that the latter is more likely to produce deliberation of the relevant kind: see, eg, Nadia Urbinati, ‘Representation as Advocacy: A Study of Democratic Deliberation’ (2000) 28 Political Theory 758. Nevertheless, as an ideal type, deliberative democracy is generally thought to require some form of public deliberation. For the sake of precision and to avoid confusion with views that blend elements of representative and deliberative democracy I have chosen to use the term ‘direct democracy’ here rather than ‘deliberative democracy’.

[76] Such views are commonly referred to as ‘epistemic’ theories of democracy: see, eg, David M Estlund, Democratic Authority: A Philosophical Framework (Princeton University Press, 2008); David Estlund, ‘Beyond Fairness and Deliberation: The Epistemic Dimension of Democratic Authority’ in James Bohman and William Rehg (eds), Deliberative Democracy: Essays on Reason and Politics (MIT Press, 1997) 173.

[77] Joel I Colón-Ríos, Weak Constitutionalism: Democratic Legitimacy and the Question of Constituent Power (Routledge, 2012). Colón-Ríos’s argument, in brief, is that a constitution is ‘democratic’ only insofar as it preserves popular constituent power, and that preserving popular constituent power requires meeting a direct democracy criterion. His view assumes, although does not defend, a direct democracy criterion. This is perhaps not so obvious, when constituent assemblies are themselves representative bodies of a kind.

[78] See, eg, Lynn M Sanders, ‘Against Deliberation’ (1997) 25 Political Theory 347.

[79] Colón-Ríos, Weak Constitutionalism, above n 77, 92–4, 160–5.

[80] Constitución Política de la República de Colombia 1886 [Constitution of the Republic of Colombia 1886] [Organization of American States trans] art 218.

[81] See Colón-Ríos, Weak Constitutionalism, above n 77, 92.

[82] Constitución Política de la República de Colombia 1991 [Constitution of the Republic of Colombia 1991] [Organization of American States trans]. It should be noted that this had the sanction of the Supreme Court, which struck down the limits imposed by Congress on the mandate of the assembly when challenged. For a discussion of this case and the mandate of the constituent assembly, see Colón-Ríos, Weak Constitutionalism, above n 77, 93–4.

[83] Donald T Fox, Gustavo Gallón-Giraldo and Anne Stetson, ‘Lessons of the Colombian Constitutional Reform of 1991: Toward the Securing of Peace and Reconciliation?’ in Laurel E Miller and Louis Aucoin (eds), Framing the State in Times of Transition: Case Studies in Constitution Making (United States Institute of Peace Press, 2010) 467, 473–4.

[84] Currently, 18 states allow amendments proposed through popular initiative; the other states require proposal through legislatures (using either a supermajority or complex majority decision rule). For an overview, see William B Fisch, ‘Constitutional Referendum in the United States of America’ (2006) 54 American Journal of Comparative Law 485, 494–7.

[85] Fisch cites some of the relevant critical literature: ibid 497. See also Elizabeth Garrett, ‘Who Directs Direct Democracy?’ (1997) 4 University of Chicago Law School Roundtable 17.

[86] Garrett, above n 85, 22.

[87] On the special interpretive difficulties that direct democracy legislation presents for courts, see Jane S Schacter, ‘The Pursuit of “Popular Intent”: Interpretive Dilemmas in Direct Democracy’ (1995) 105 Yale Law Journal 107.

[88] Lutz, above n 61, 254.

[89] This is the solution Colón-Ríos adopts, although he is less concerned about these kinds of risks: Colón-Ríos, Weak Constitutionalism, above n 77, ch 7.

[90] Craig Prescott, ‘The Union, Constitutional Change and Constitutional Conventions (and English Regionalism?)’ on UK Constitutional Law Association (3 April 2013) <http://ukconstitutionallaw.org/2013/04/03/craig-prescott-the-union-constitutional-change-and-constitutional-conventions-and-english-regionalism/> .

[91] A classic statement of this view is found in John Stuart Mill, Considerations on Representative Government (Parker, Son, and Bourn, 1861).

[92] Here, too, it should be noted that there is not a perfect mapping between these categories. As noted, some theorists defend representative democracy on deliberative grounds: see above n 75. Nevertheless, as an ideal type, the view is more closely associated with the aggregation of preferences through voting and elite competition, as in the classic Schumpeterian model: see Joseph A Schumpeter, Capitalism, Socialism and Democracy (Routledge, first published 1942, 2010 ed).

[93] What we think about the merits of a proposed amendment that fails for this reason is, of course, a separate consideration. That is, this is not to deny that there might be other reasons for thinking that a minority veto is a good thing in some circumstances; these just are not reasons that go to considerations of democracy in constitutional law-making.

[94] See above n 59 and accompanying text.

[95] This is one version of the thesis that I have termed constitutional formalism, which is the view that a constitution is a text like other legal texts: Weis, above n 28, 855. It is a view that has long described the mainstream approach of the High Court of Australia: see Jeffrey Goldsworthy, ‘Australia: A Devotion to Legalism’ in Jeffrey Goldsworthy (ed), Interpreting Constitutions: A Comparative Study (Oxford University Press, 2006) 106.

[96] I suspect that theorists of a formalist persuasion in Australia and Canada, who tend to be attracted to legislative supremacy more so than their American counterparts, are likely to fall into this camp.

[97] I suspect that many American formalists would take this position; indeed, many American constitutional scholars view entrenchment as a necessary feature of a constitution. Justice Scalia would almost certainly fall into this camp; despite his apparent preference for majoritarian democracy, he unequivocally rejects majoritarian constitutionalism: see Justice Scalia, above n 12, 46–7.

[98] [2004] HCA 37; (2004) 219 CLR 562. See, eg, Justice M H McHugh, ‘The Need for Agitators — The Risk of Stagnation’ (Speech delivered at the Sydney University Law Society Public Forum, 12 October 2005).