7.1 Introduction

In recent decades, constitutional review has spread to many corners of the democratic world. However, the de facto success of constitutional review has been accompanied by a growing chorus of profoundly critical assessments in legal scholarship. The view that there is no compelling way to justify constitutional review in a democratic polity has, in some quarters, turned into a near-orthodoxy.

So-called ‘political constitutionalists’, in both the American and the British contexts, argue that no polity that settles disputes over what rights people possess through politically unaccountable judicial institutions whose decisions are often irreversible even by large legislative majorities can be said to be fully democratic. Political constitutionalists do not deny that democracy requires respect for a number of fundamental rights, such as rights to political participation under fair and equal terms, rights to association and assembly, respect for the freedom of opinion and expression, and protection against politically motivated interference with nonpolitical rights. Instead, they argue that a fully democratic community must settle any disputes concerning the content of these rights, as they are bound to arise in any political community, within the democratic political process itself. Leaving the matter to the courts involves subjecting democratic self-government to the potentially dominating control of an aristocracy of the robe.Footnote ¹

Hans Kelsen was one of the inventors of constitutional review. Kelsen drafted the Austrian Constitution of 1919, which was the first European constitution to provide for a constitutional court endowed with the power to strike down unconstitutional legislation and unconstitutional administrative acts.Footnote ² Kelsen also advocated for the introduction of constitutional review in other European states, most notably through his intervention in the constitutional debate on the issue in Weimar Germany.Footnote ³ Although the U.S. Supreme Court successfully claimed the power to exercise incidental constitutional review as early as 1803, in Marbury v Madison, the Austrian Constitution of 1919 introduced a new model of constitutional review, one that has since become a widely copied template for the institutional design of courts that are meant to act as guardians of the constitution.Footnote ⁴ In the Kelsenian model, constitutional review is exercised by a specialised constitutional court that is explicitly empowered to bindingly settle all constitutional questions; these may arise from cases pending in other courts that have been referred to the constitutional court, from executive decisions against which individuals have brought individual constitutional complaints, or from disputes among different organs of the constitution or different levels of government.

Kelsenian constitutional courts wield their powers by constitutional design, not merely by virtue of some judicial arrogation of power that has been accepted, however grudgingly, by other constitutional actors. A written constitution endows these courts with the authority to invalidate unconstitutional norms and not merely to make declarations of incompatibility which are liable, at least in principle, to be overridden or disregarded by legislative authorities. Unsurprisingly, Kelsenian constitutional courts have sometimes become politically powerful institutions.Footnote ⁵ If constitutional review – or judicial review in general – raises problems of democratic legitimacy, such worries, or so it would seem, would apply to Kelsenian constitutional courts with a vengeance.

All this would lead one to expect that Kelsen would have supported what the late Ronald Dworkin called a constitutional (as opposed to a majoritarian) conception of democracy.Footnote ⁶ According to Dworkin, our view of what institutional frameworks or what mechanisms of decision-making ought to be regarded as authentically democratic must be sensitive to the most attractive understanding of the value of democracy. Dworkin further argued that according to the most attractive understanding of the value of democracy, a democratic political system is one that treats all citizens with ‘equal concern and respect’. Therefore, procedures of political decision-taking, such as the majority rule, ought to be assessed in instrumental terms, that is, by reference to the question of whether they are likely to bring forth decisional outcomes that conform to the procedure-independent moral standard of equal concern and respect.

The constitutional conception of democracy provides a straightforward strategy of justification for constitutional review. If constitutional control exercised by a constitutional court makes it more likely that decisional outcomes conform to procedure-independent standards of justification that demand that each citizen be treated with equal concern and respect and if the institutions of democracy are valuable only insofar as they serve that goal, then constitutional control is democratically justified despite its countermajoritarian character. It is not undemocratic to restrict majoritarian decision-taking or to subject it to judicial control if that makes it more likely that all citizens are treated with equal concern and respect. Dworkin argued that judicial control for constitutionality can indeed be expected to have that salubrious effect. A supreme or constitutional court – the ‘forum of principle’Footnote ⁷ – is often more likely to base its decisions on sound justificatory arguments than the government or parliamentary majority of the day.

Needless to say, the standard political-constitutionalist objection to Dworkin’s argument for constitutional review is the charge that it replaces the democratic self-government of citizens with the rule of judges and liberal philosophers who presume to determine the conditions under which democratic decisions might be acceptable in advance of the political process.Footnote ⁸ Although Kelsen was a staunch defender of a robust practice of constitutional review, I suspect that he would have sympathised with the political-constitutionalist attack on Dworkin’s constitutional conception of democracy. In his argument for constitutional review, Kelsen pointedly refrains from appealing to anything like a Dworkinian constitutional conception of democracy. Kelsen would likely have agreed with the view that a community’s fundamental commitments ought to be defined in democratic politics and not by the courts.Footnote ⁹ This stance suggests that Kelsen’s argument for constitutional review must be either particularly interesting or extraordinarily confused. Kelsen must have either managed to identify an argument for constitutional review that does not rely on a constitutional conception of democracy, thus dispelling worries about the democratic legitimacy of constitutional review, or failed to recognise that a majoritarian understanding of democracy entails that there is no compelling justification for constitutional review.

The latter of these two assessments has been defended, with great clarity and vigour, by Michel Troper, one of the few notable legal and constitutional theorists who have paid close attention to Kelsen’s argument for constitutional review.Footnote ¹⁰ This chapter will take the opposite view. It will be argued here that Kelsen’s case for constitutional review is fundamentally sound and has the potential to make an interesting contribution to contemporary debates on the legitimacy of judicial control of constitutionality.

Troper’s analysis of Kelsen’s case for constitutional review helpfully distinguishes between what might be called a positive and a negative Kelsenian argument for constitutional review.Footnote ¹¹ The positive argument aims to show that the introduction of constitutional review is necessary for establishing the supremacy of the constitution over other, non-constitutional laws. The negative argument aims to establish that there is no conflict between constitutional review and the principle of popular sovereignty. According to Troper, both arguments fail. Kelsen’s positive argument conflicts with his own theory of legal system, while the negative argument is inconsistent with Kelsen’s own understanding of democracy. My defence of Kelsen’s case for constitutional review will address both of Troper’s attacks in turn.

7.2 Kelsen and the Problem of Constitutional Supremacy

Kelsen called his legal theory the ‘Pure Theory of Law’ (henceforth ‘PTL’). The purity of the PTL, according to Kelsen, is to be achieved on two fronts. On the one hand, the PTL is not an empirical or social scientific theory of law. A legal-scientific description of the law does not involve causal knowledge. The science of the law, though descriptive, is expressed in normative statements. The latter state what is to be done according to the law. They do not offer causal predictions, as one might find them in a legal-realist account of the law. For instance, they do not say that one is likely to be punished if one commits a certain act but that having committed the act makes it the case that one ought to be punished. However, in offering such ought-statements, a legal scientist, according to Kelsen, is not expressing their own moral opinion or offering a practical endorsement of the law’s demands. Although they are formulated in normative language, legal scientific statements are intended to do no more than to describe what is in fact required by the law. They do not express any practical evaluation of the law’s demands. Hence, the discourse of legal science is as sharply distinct from moral as it is from empirical discourse.Footnote ¹²

Troper observes that Kelsen’s advocacy for the introduction of constitutional review is a normative demand based on a conception of what kind of legal order we would have reason to evaluate favourably. It therefore stands in conflict, Troper concludes, with the second demand of purity, which blocks the legal scientist from issuing practical prescriptions.Footnote ¹³ Troper concedes that a Kelsenian might reply that when Kelsen argued for the introduction of constitutional review, he was not speaking as a legal scientist, committed to the demand for purity, but rather as a political theorist, reflecting on the conditions of good legal order. However, this reply will not suffice. As Troper rightly emphasises, Kelsen appears to suggest that his advocacy for constitutional review is based on more than a self-standing normative conviction.

Kelsen presents the institution of constitutional review as the necessary means to realise the legal effectiveness of the constitution. According to Kelsen, a constitution purports to impose legal conditions on the valid exercise of legislative and executive power. Such conditions may take the form of purely procedural rules, but they may also contain substantive requirements that are related to the content of legislative and administrative acts. Kelsen argues that in the absence of constitutional review, the legislature and the executive will be free to disregard these conditions or to interpret them in any way they please. Although constitutional conditions for the exercise of legislative and administrative power purport to be conditions of validity, they are bound to remain legally ineffective in the absence of constitutional review. Thus, Kelsen claims that the only way to give full legal effect to the constitution is to introduce constitutional review.Footnote ¹⁴

This argument for constitutional review, which I will call the ‘argument from constitutional supremacy’, responds to a particular stage in constitutional development that was characteristic of many interwar European polities. The Weimar Republic, to name a prominent example, had a written constitution that contained both procedural rules governing a complex federal system of government and a bill of rights purporting to impose substantive restrictions on legislative action. However, the Weimar Republic did not have a fully developed practice of constitutional review.Footnote ¹⁵ Put concisely, Kelsen’s argument is that the apparent commitment to constitutionalism, as it is expressed in the written constitution itself, makes sense only if it is bolstered by constitutional review. Only through constitutional review will the conditions that the constitution purports to impose on legislative and executive action become conditions of the validity of legislative and executive acts and be legally effective.

In Troper’s reading, Kelsen’s argument from constitutional supremacy is intended to be conformable to the demand that the PTL be pure in the sense of being nonevaluative. Troper understands Kelsen to be claiming that there will be no constitution, in the absence of constitutional review, that is hierarchically superior to ordinary law and provides the basis of the validity of ordinary laws. The very existence of a structured legal system, of a consistent hierarchy of norms, or so Troper’s Kelsen argues, requires constitutional review. Thus understood, Kelsen’s argument from constitutional supremacy, Troper claims, does not conflict with the demand for purity because it is entirely technical. In demanding the introduction of constitutional review, the argument merely identifies a necessary condition for the existence of a legal order with a clearly defined hierarchy of norms, one that allows the constitution to function as the ground of the validity of ordinary laws.Footnote ¹⁶

Unfortunately, Troper goes on to claim, the argument from constitutional supremacy conflicts with certain key assumptions of the PTL’s theory of legal system. According to Troper, the problem can be illustrated by reference to Kelsen’s account of the status of unconstitutional statutory norms.Footnote ¹⁷ Suppose the legislator enacts a statutory norm – through an act that sufficiently conforms to the relevant procedural rules to allow us to recognise that norm as a law, as having membership in the legal system – that is believed to violate some constitutional condition. Kelsen denies that such a norm is to be regarded as invalid. Rather, he argues that it is annullable on the ground of unconstitutionality, assuming that there is recourse to constitutional review. However, Kelsen adds that the norm in question will stand as valid and be enforced until it has been annulled by an organ that is specifically empowered to do so, the constitutional court.Footnote ¹⁸

Troper notes that if the annullable norm is to be valid for the time being, until it has been annulled by a constitutional court, there must be a hierarchy of norms, and there must be a constitution that validates the impugned norm for the time being. Kelsen claims that norms exist in the mode of validity. What it means for a norm to exist, to be valid, is for it to belong to a legal system. Hence, we can recognise that a norm belongs to a legal system only if it satisfies the conditions of validity laid out by certain higher-level norms of the system. A constitution – a set of top-level norms that lay down the conditions of the validity of subordinate legal norms – must be in place wherever there is a legal system, whether or not that system provides for an institution endowed with the power to annul norms on the ground of unconstitutionality. There is consequently no room for Kelsen to argue, Troper concludes, that a norm that is valid might nevertheless be unconstitutional, or to claim that there is no clearly defined hierarchy of norms, no constitution capable of validating subordinate legal norms, in the absence of a judicial institution empowered to perform constitutional review. The claim that there ought to be such an institution, then, cannot be portrayed as a value-neutral technical demand that does no more than spell out the conditions of the very existence of a structured legal order.Footnote ¹⁹

Troper’s challenge to the argument from constitutional supremacy is certainly ingenuous, but it is based on a misunderstanding of the character and intentions of Kelsen’s argument from constitutional supremacy. As Troper himself notes, a technical argument is an argumentFootnote ²⁰ that aims to validate a hypothetical imperative. Given that a certain end is to be accomplished, what means are necessary or at least pre-eminently suitable for doing so? Troper goes wrong in assuming that the relevant goal, the goal that Kelsen claims can be accomplished only with the introduction of constitutional review, is that of providing the conditions of the very possibility of a validation of legal norms. Instead, the goal is to give effect to the conditions contained in a written constitution that is not yet protected by constitutional review.

To elucidate the point, Kelsen’s analysis of the concept of constitution needs to be more closely examined. Kelsen distinguishes between what he calls a ‘material constitution’ and a ‘formal constitution’.Footnote ²¹ Any legal order necessarily possesses a material constitution. In any legal order, there must be top-level norms that provide the criteria for the validation of subordinate, lower-level norms. These top-level norms must, at the very least, establish the procedures for the valid exercise of legislative power, of the authority to enact general legal rules. In contrast, a formal constitution does not by necessity exist in any legal order. Rather, it is present wherever top-level norms are shielded against repeal or modification in the (ordinary) process of legislation. Such protection is typically (although not invariably) the result of a written constitutional document that confers enhanced legal force on all the norms contained in it. The most common method for establishing a formal constitution is to enact a constitution that protects constitutional norms with an amendment clause that imposes special procedural hurdles on the repeal or modification of all or some of the norms contained in the written constitution. A written constitution might also make certain norms completely immune to change or repeal, as in constitutions that contain so-called ‘eternity clauses’.

Any reasonably careful reader of Kelsen’s constitutional works should see that the argument from constitutional supremacy is concerned with the question of how to give legal effect to the norms of the formal constitution.Footnote ²² As aforementioned, Kelsen’s argument is addressed to a specific constellation in constitutional politics. It is assumed that the addressees of the argument accept and endorse the desirability of a formal constitution that imposes specific conditions on the validity of legislative and executive action or that a formal constitution that purports to impose such conditions is already in place. Given these background assumptions, constitutional review is necessary to give full legal effect to the formal constitution. Suppose, for instance, that the constitution prohibits laws that condition access to public services on religious belief, but that the legislature has enacted a law that blocks access to certain public services on the basis of religious belief. If there is no opportunity to appeal for the annulment of the law, it will remain on the books and be enforced, notwithstanding the fact that it violates a constitutional condition. In other words, the formal constitution will turn out to be a dead letter. The conditions it purports to impose on the validity of statutory norms will be devoid of legal effect. Kelsen argues that if one is committed to the idea that there should be a formal constitution and that it ought to be more than a dead letter, then one must also approve of constitutional review.

Note that the argument from constitutional supremacy, understood in this way, is an argument de lege ferenda. Kelsen does not claim that the mere fact that a formal constitution is in place empowers courts to engage in constitutional review and to strike down statutory norms. For example, Kelsen’s argument does not entail that the Supreme Court of India was legally justified in introducing basic structure review,Footnote ²³ as Troper would appear to suggest.Footnote ²⁴ Troper claims that Kelsen’s argument for judicial review is in effect the same as that put forward by Marshall in Marbury v Madison.Footnote ²⁵ That view is plainly mistaken. Indeed, Marshall held that the mere fact of the existence of a formal constitution entitled the U.S. Supreme Court to exercise constitutional review. Kelsen’s argument from constitutional supremacy claims no more than that it would be fitting for the legislator or the constitutional legislator to endow a constitutional court with the power to exercise constitutional review where a formal constitution is in place or is to be put in place.Footnote ²⁶

It would have been absurd for Kelsen to take the view Troper attributes to him, that is, the view that there can be no constitution in the legal-logical sense, no legal hierarchy, and no validation of lower-level norms by top-level norms without constitutional review. However, Kelsen clearly did not intend to argue for an absurdity of this sort. Once it is recognised that the argument from constitutional supremacy is concerned with giving effect to the formal constitution, Troper’s challenge to Kelsen’s argument from constitutional supremacy turns out to be based on a simple fallacy of equivocation. Consider a statute that appears to violate a constitutional norm but that is valid and, in the absence of constitutional review, cannot be annulled for its unconstitutionality. Here is what Troper has to say about that scenario:

When he speaks of ‘the constitution’, in this passage, Troper must be referring to the constitution in the material sense. As we have seen, Kelsen indeed holds to the view that there will always be a material constitution and concedes that a material constitution may well give validity to a law that conflicts with conditions of validity a formal constitution purports to impose on statutory norms. As we have also seen, Kelsen does not claim that there is no material constitution where there is no constitutional review. He does not argue that constitutional review is necessary because there could otherwise be no constitution that validates lower-level norms. Instead, he claims that a formal constitution may turn out to be legally ineffective if there is no constitutional review or, in other words, that the actual conditions of legal validity may come to differ from those announced in the formal constitution so that the latter becomes a dead letter. Constitutional review is meant to prevent this potential gap between the material and the formal constitution from opening up. When Troper claims that ‘the constitution’ will exist even without constitutional review, he simply avoids the question in which Kelsen was interested. The question in which Kelsen was interested, in putting forward his argument from constitutional supremacy, was not whether there will be some constitution, some set of conditions that in fact determine what laws will stand as valid. The question was which constitution that will turn out to be – the one that was solemnly promulgated in a constitutional document or the one that political powerholders, who may not be very interested in respecting constraints on their legislative and executive authority, manage to impose.

Troper would likely reply to these observations by invoking a second line of argument that figures prominently in his attacks on Kelsen’s argument from constitutional supremacy. Troper argues that a proponent of the PTL could never be in a position to claim that a statutory norm is unconstitutional, so that any alleged contrast between the constitution in the material sense and the formal constitution must be spurious:

It is true that, according to the PTL, there can be no norms that are null. This much follows from the claim that validity is the mode of existence of norms.Footnote ²⁹ However, it should be emphasised that the claim that there can be no norms that are null does not entail that there cannot be purported norms that are null. The instructions issued by the Hauptmann von Köpenick, frequently referenced by Kelsen,Footnote ³⁰ are a case in point. For a norm to be valid, the norm must at the very least have been enacted, as Troper himself puts it, ‘in accordance with the procedure laid down’,Footnote ³¹ it must proceed from a competent authority.

Disputes about constitutionality typically arise in cases where an act does proceed from a proper authority and is recognised, by citizens and officials, not to be altogether null, but where the act is held not to comply with all conditions imposed by a formal constitution. Here, the notion of annullability comes into play.Footnote ³² Acts that pass the threshold of nullity will have legal force until they are annulled by a competent organ. Regarding annullable norms, Troper argues that it is impossible to claim that they are contrary to the constitution before they have been annulled by a competent tribunal. Troper’s challenge is based on a fallacy of equivocation in this case as well. The claim that any valid legal norm must also be constitutional again confuses the constitution in the material sense with the formal constitution. The fact that an unconstitutional law satisfies the procedural conditions that suffice to make it valid, pro tempore and until it is annulled, does not imply that it cannot conflict with one or another element of the full array of requirements of constitutionality provided in the formal constitution.

I suspect that Troper would reply that someone’s view that a procedurally valid law conflicts with some requirement imposed by the formal constitution can be nothing more than an ‘opinion subjective’ until the tribunal empowered to annul has given its verdict.Footnote ³³ However, this reply is ambiguous between two senses of the term ‘subjective’ that need to be kept separate.

The claim that someone’s opinion that a law is unconstitutional is subjective might be understood to mean that that person is not formally empowered, unlike the competent tribunal, to render an authoritative decision as to whether the law in question is unconstitutional or not. My personal opinion that some statute fails to satisfy some constitutional requirement – however true or well justified – has no legal effect. It is evident that the opinions regarding the constitutionality of laws held by private individuals, including legal scholars, are subjective in this first sense.

However, I suspect that Troper’s use of the adjective ‘subjective’ is intended to make a claim that goes well beyond the anodyne observation that private persons lack the authority to annul a law. Troper’s claim that someone’s view that a law is unconstitutional is always subjective before the competent tribunal has rendered its verdict is meant to convey, apparently, that there is nothing to be right or wrong about here and that no opinion on the substantive constitutionality of a law that has been enacted ‘in accordance with the procedure laid down’ could ever be more justified than any other. Troper suggests that any view on the constitutionality of a law that passes the threshold of nullity must involve a value judgement and thus be a function of the will and not of cognition. It follows, of course, that the decisions of the competent tribunal must themselves be legally groundless and that they can never be more than instances of legislation. Accordingly, constitutional review is objectionable simply because it allows the subjective axiological preferences of unelected judges to determine which laws enacted by parliament will stand.Footnote ³⁴

It should be clear that the fact that an opinion on the constitutionality of a law is subjective in the first of the two senses distinguished above does not entail that it is subjective in the second. That it is the opinion of a competent tribunal that decides whether a law is to be annulled for unconstitutionality does not imply that such a tribunal must be legally infallible or that the opinions of private persons or legal scholars can have no merit if they conflict with the tribunal’s opinion. Moreover, there is no need to invoke the second sense of subjectivity to account for the facts that Kelsen tried to capture in his analysis of annullable laws. The interest in legal certainty sufficiently explains why the positive law typically determines that an annullable statutory norm will stand until it has been repealed by a competent tribunal and why it will become permanently valid if a competent tribunal of last instance decides not to annul it. The tribunal, like anyone else, may be right or wrong in its judgement of constitutionality, even though the law makes its decisions final and binding to allow for the settlement of constitutional disputes.

Certainly, the question of whether some statutory norm violates a constitutional requirement will often be subject to reasonable disagreement. Even competent legal practitioners who are genuinely interested in identifying the correct legal answer to a hard case may disagree with one another. As Kelsen was well aware, defenders of constitutional review must answer the question of why the task of taking a binding decision in a constitutional conflict about which there can be reasonable disagreement ought to be assigned to a court. Kelsen was happy to admit that to assign this authority to a court is to endow a judicial institution with considerable power, and he was not hesitant to admit that this power is political. He believed that there are good political reasons to embrace constitutional review.Footnote ³⁵

In Section 7.3, we will engage with the political dimension of Kelsen’s case for judicial review. Suffice it to say for now that Troper is not in a position to invoke the problem of reasonable disagreement to assail the legitimacy of constitutional review. Indeed, Troper implicitly denies that there is any such thing as reasonable disagreement about constitutional matters. Whether some statute has been enacted in conformity with the relevant procedures, he intimates, is a question that will rarely be subject to reasonable disagreement among competent lawyers, presumably because it is a question that does not involve axiological judgement. Questions as to whether some statute conforms to some substantive constitutional constraint, on the other hand, are supposedly mere matters of taste, which are not amenable to rational judgement. In this latter case, there simply is nothing to disagree about, although preferences may clash. Obviously, this view presses rather hard on the distinction between procedural and substantive questions. The claim that procedural questions are insulated from reasonable disagreement is not much more plausible than the view that people who find themselves in an axiological dispute must simply be shouting at each other.

To conclude the discussion of Troper’s challenge to the argument from constitutional supremacy, I would like to offer anticipatory replies to two obvious objections to the interpretation that I have presented here. According to the first of these objections, the argument, as I understand it, violates the PTL’s demand for purity or value neutrality. According to the second objection, the argument will be much less interesting than it appears to be at first glance and much less general in its scope if we adopt the interpretation that has been suggested here.

In response to the first objection, I am happy to concede that Kelsen’s argument from constitutional supremacy indeed involves value judgement. However, I do not see why that would show Kelsen to have been inconsistent or to have violated the PTL’s demand for purity. As aforementioned, the argument for constitutional review is an argument de lege ferenda or, more precisely, de institutione creanda; it does not claim to describe the law as it exists, nor does it argue that one can draw a justification of constitutional review from the very concept of law. Therefore, it is unclear why the demand for purity should apply to it. Troper is correct to point out that Kelsen’s argument from constitutional supremacy is intended to be technical, but as we have seen, he misidentifies the end that technique is to serve in this context. The relevant end is not to establish some legal order or other, no matter which, but to give legal effect to the formal constitution. That we ought to do so is a value judgement, of course. In presenting his case for constitutional review, Kelsen assumed that this judgement was uncontroversial, especially in cases where, as in the Weimar Republic, a formal constitution was already in existence. That the argument from constitutional supremacy refers to an end assumed to be valuable does not entail, in any case, that the technical reasoning about how to reach that end is evaluative. Even those who deny the value of some end should be able to agree, as Max Weber argued, that claims as to how that end might be accomplished are perfectly amenable to value-neutral scientific investigation.

According to the second objection, the interpretation of the argument from constitutional supremacy that has been presented here makes the argument much less interesting than it appeared to be. In my reading, Kelsen does not argue that any legal order must practice constitutional review to work as intended or that constitutional review must be legitimate wherever it exists. Moreover, he does not argue that the existence of a formal constitution by itself entitles a jurisdiction’s highest courts to exercise constitutional review and to strike down unconstitutional statutes. Kelsen’s argument from constitutional supremacy presents the much more limited claim that constitutional review is needed to give full legal effect to a formal constitution. It assumes, without defending that assumption, that it is desirable for a polity to have a formal constitution and for that constitution to be given legal effect. However, this modesty is not a deficiency of the argument from constitutional supremacy. It puts the focus of the debate on constitutional review where it belongs, namely, in the context of the question of whether it is appropriate for a democratic polity to embrace formal constitutionalism.

7.3 Kelsen on Constitutional Democracy

This brings us to Kelsen’s negative argument, which is intended to show that constitutional review need not conflict with democracy but can be helpful to its full realisation. This argument is not made entirely explicit in Kelsen’s writings, but it can be pieced together easily enough if one reads Kelsen’s writings on constitutional review in light of his theory of democracy.Footnote ³⁶

Before I embark on this exercise in reconstruction, let me add one important preliminary clarification concerning Kelsen’s argumentative goals. Kelsen’s negative argument does not aim to show that constitutional review cannot conflict with democracy. Certainly, it would be absurd to deny that the compatibility of constitutional review with democracy depends on the content of the formal constitution. Needless to say, a formal constitution could contain substantive provisions that are completely incompatible with democracy. If that is the case, the attempt to give full legal effect to the provisions in question by way of constitutional review would, also needless to say, likewise be incompatible with democracy. However, there is no reason to think that a defender of the democratic legitimacy of constitutional review is committed to defending the view that constitutional review has to be welcomed in all its conceivable instantiations. What they are committed to defending is the view, rather, that a suitably designed constitution and system of constitutional review may be conducive to the flourishing of democracy.

At first glance, Kelsen’s theory of democracy admittedly seems committed to a denial of even this modest claim. Kelsen’s analysis of democracy starts from a thesis that seemingly puts it at odds with standard understandings of the value of democracy. It argues that the nature of democracy is not to be explicated with reference to the value of equality but with reference to the value of freedom.Footnote ³⁷

Any social order subjects individuals to a condition that Kelsen describes as the ‘torment of heteronomy’.Footnote ³⁸ Political authorities make laws that purport to be binding, at least until they are repealed, and that are enforced against the recalcitrant by threats of coercion. Individuals who are members of a political society must therefore expect to find themselves compelled to act in ways that conflict with their own preferences and plans. The resulting insult to our autonomy calls for justification. Kelsen argues that democracy differs from other political regime types in that it offers at least a partial reconciliation of our interest in freedom with the unavoidable fact of heteronomy. According to Kelsen, the key characteristic of a democracy is that legislative decisions, or decisions on the enactment of general legal rules, are taken by the use of the majority rule. The majority rule ensures that a collective decision, though it must unavoidably frustrate the preferences of some people, conforms to the preferences of the largest possible number of citizens.Footnote ³⁹

Imagine a community that chooses to adopt a rule of continuing agreement as its rule of collective decision. A collective decision, under a rule of continuing agreement, requires everyone’s approval to come into force, and it remains in force for as long as everyone continues to endorse it. A rule of continuing agreement is tantamount to anarchy. It does not constitute a social order as Kelsen understands the term. According to Kelsen, a social order must be objective in the sense that its demands will continue to bind (and to be enforced) even if some group members come to reject them. A rule of unanimity, in contrast to a rule of continuing agreement, does not run afoul of the requirement of objectivity. Under the rule of unanimity, a collective decision requires everyone to agree to come into force, and it remains in force until it is repealed by a later, equally unanimous decision. A decision that is taken under the rule of unanimity continues to bind individuals who change their mind as time goes on and who come to prefer a different arrangement, at least as long as there are those who would like to stick with the original decision. The rule of unanimity therefore possesses objectivity. It may come to bind members of society in ways that go against their momentary will. The rule of unanimity may even come to bind the vast majority of members of society against their momentary will. A widely shared desire for legislative change might, in the most extreme case, come to be frustrated by the veto of a tiny minority of one. Note that any rule of collective decision that requires a supermajority has a similar effect: It will come into force only if most members of a group agree and will stand unless it is repealed by another, later supermajority, which makes it possible for a minority to hold a majority hostage. As should be evident, any rule of collective decision that puts the power to enact binding rules into the hands of an autocrat or an oligarchic minority will carry a similar potential to frustrate the vast majority of members of a group. Where legislative decisions are taken by one person or by a small group, there can be no assurance that the resulting norms will reflect the preferences of a significant number of addresses of the law.

These reflections show that the majority rule has a unique virtue, as Kelsen argues. The majority rule reduces the tension between the momentary individual preferences of citizens and the legal rules that are concurrently in force as much as possible within a framework of objective social order. Under the majority rule, a decision will come into force if and only if it is approved by a majority and will remain valid and binding until it is repealed by a later majority. The majority rule thus ensures that any collective decision, while it is binding, will continue to be in line with the momentary preferences of a majority of norm-addressees. There will always be more norm-addressees whose preferences are in line with the law than norm-addressees whose preferences are frustrated by the law. This assurance can be provided neither by an autocratic rule of collective decision nor by a rule that moves towards unanimity. Given that social order is necessary, the majority rule is the best we can do. Citizens who suffer a torment of heteronomy, if faced with democratically enacted law, have no reason to complain. The extent of freedom enjoyed by the subjects of the law has been maximised.

Troper rightly points out that this defence of the majority rule appears to conflict with Kelsen’s support for formal constitutionalism.Footnote ⁴⁰ Kelsen defines formal constitutionalism in terms of a rule of constitutional amendment: a formal constitution exists if and only if there are legal norms that are immune from repeal or modification in the ordinary legislative process.Footnote ⁴¹ Where the ordinary legislative process is democratic and majoritarian, the existence of a formal constitution entails that the repeal or change of constitutional norms requires a supermajority of some sort. However, Kelsen’s argument for the majority rule seems to imply that any adulteration of the simple majority rule is liable to reduce the extent of freedom enjoyed by subjects of the law. Kelsen, given the logic of his own argument for democracy, ought to have rejected formal constitutionalism and constitutional review.

The tension between Kelsen’s defence of the majority rule and his commitment to formal constitutionalism is undoubtedly real. However, this observation does not suffice to reject Kelsen’s view of the proper structure of a democratic constitution as incoherent. Kelsen’s reflections on the nature and value of democracy, on closer scrutiny, turn out to contain the resources to mitigate a charge of inconsistency.Footnote ⁴²

The general thesis Kelsen aims to defend in his reflections on democracy is that citizens of a democratic state have reason to regard the norms to which they are made subject as legitimate, even in instances where their own momentary preferences conflict with the law. Now suppose you are a citizen who is outvoted, for the time being, and who must therefore submit to rules that frustrate their preferences. It is not at all clear why the brute fact that the majority have had their way and that the extent of freedom under law has thus been maximised should suffice to reconcile you to your fate. The argument, in this brute form, has the whiff of a utilitarian disregard for the separateness of persons. You will argue that in the democratic state, every citizen’s interest in not seeing their preferences frustrated by the law is equally deserving of respect and consideration.

Kelsen would agree. Although he claims that democracy is to be defended, in the first instance, by appeal to the value of freedom, Kelsen concedes that a full justification of democracy must also consider certain claims of equality.Footnote ⁴³ At the very beginning of his discussion of the value of democracy, in explaining how the torment of heteronomy arises, Kelsen makes the following observation:

To rephrase, if we are all by nature equal in moral status, as a democrat would claim, no one can possess natural authority over anyone else. However, social order must nevertheless be established, and democracy is the form of social order that is most compatible with the assumption of natural equality in moral status. It leaves us all as free as we were before, as Rousseau would have it, or at least with as much freedom as we can reasonably hope to enjoy within the confines of the law, as Kelsen argues, in a somewhat more realistic vein. If Kelsen’s watering down of Rousseau’s promise of equal democratic freedom is not to strip democratic procedure of its legitimating power, then democracy must do more than simply maximise freedom under law. It must affirm our natural equality in political status.

The democratic state respects us as equals only where it gives equal consideration to everyone’s interest in freedom. To meet that requirement, the democratic political process, centred as it is on the majority rule as the dominant rule of collective decision, must therefore be supported by a number of background conditions. Suppose again you find yourself in the position of an outvoted citizen. Here are a number of further observations, observations going beyond the brute claim that freedom under law has been maximised, which, given you find yourself in a healthy democracy, might help you to find reconciliation with social order: The laws to which you are compelled to submit have been enacted as the result of a free and fair process of political competition. All citizens enjoyed an equal right to advocate for their views, individually and in organised groups, without having to fear any disadvantage in expressing their views. All social groups had adequate access to the public sphere to make themselves heard. The decision that has now been made, though it is binding pro tempore, is open to revision. It may come to be repealed at a later stage if a new majority so pleases. All citizens have the right to work, as individuals or groups, towards convincing their fellow citizens to repeal the offending law.

Kelsen is fully aware that the mere presence of legal or constitutional guarantees of freedoms of opinion, assembly and equal democratic participation, however necessary, does not suffice to establish the background conditions for the proper functioning of the majority rule.Footnote ⁴⁵ Outvoted citizens must know that there is a real prospect (and not merely the legal possibility) of an alternation in legislative and governmental authority as a result of the democratic process. There must be a party system that aggregates individuals into two opposing groups that are committed to the indefinite continuation of fair and peaceful political competition that honours the ‘rules of the democratic game’. Both sides must recognise each other as legitimate contenders for legislative authority and be willing to accept outcomes of elections that favour the other side. Where this is the case, one will likely see two large parties or coalitions of parties that aim to occupy the middle ground in an attempt to build an electoral majority. Truly democratic parties are willing to compromise with their opponents and to offer and act on a plausible interpretation of the common good, one that is not designed to buy the support of the narrowest constituency that might have a chance to win out in skewed electoral competition but to appeal to a broad array of voters on both sides of the central political cleavage.Footnote ⁴⁶

If both the legal and the factual conditions of properly functioning democracy are fulfilled, an individual who is outvoted now can reasonably hope to be a member of a future majority and to then enjoy the benefit of being subject to law that is in line with their personal preference. A flourishing democracy honours our equality in political status by making us all potential members of coalitions that exercise legislative authority until they are replaced, in free and fair elections, by a successful opposition. We are all included in a temporally extended popular will. Instead of speaking of the majority rule, Kelsen concludes, it would be more appropriate to speak of a ‘majority-minority rule’.Footnote ⁴⁷ A truly democratic majority can exist only where it is challenged by legitimate opposition, a minority that is a legitimate government in waiting.

To conclude, the simple freedom-based argument for the majority rule does not offer a full account of Kelsen’s conception of democratic legitimacy. Once the necessary supplements that ensure majoritarian democracy’s ability to offer reconciliation with social order have been brought into the picture, one can appreciate the fact that there is no necessary conflict between democracy and formal constitutionalism. Whether formal constitutionalism and constitutional review conflict with democracy depends on the content of the formal constitution. An ideal Kelsenian formal constitution would limit itself to protecting the procedures as well as the individual rights and liberties that are essential to properly functioning democracy. Constitutional review that enforces a formal constitution with that content serves to protect minorities against the tendency of a majority to entrench itself in power by rigging the rules of the game to deprive the minority of a realistic prospect of a democratic alternation in power. Thus, constitutional review defends the equality in political status that we must accord to each other as democratic citizens if the majority rule is to be able to legitimate its outcomes.Footnote ⁴⁸

To reemphasise a point I made at the beginning of this section, Kelsen’s negative argument is a modest one. It does not claim that there can be no flourishing democracy without formal constitutionalism and constitutional review. It could be that conventions and informal norms are strong enough, in this or that polity, to maintain the integrity of the democratic process. Moreover, Kelsen does not argue that formal constitutionalism and constitutional review are compatible with democracy in all their conceivable instantiations. A formal constitution may contain provisions that are designed to frustrate the collective self-determination of equal democratic citizens. What the negative argument claims is that there is no inherent conflict between democracy and formal constitutionalism bolstered by judicial review. A democratic constitution can be designed in a way that makes formal constitutionalism supportive of democracy.

7.4 The Contemporary Relevance of Kelsen’s Argument

It might be argued that the conclusion of the negative argument – that there is no inherent conflict between democracy and formal constitutionalism bolstered by constitutional review – is altogether too modest to be interesting. I will elaborate further to establish that formal constitutionalism and constitutional review, when designed in the right way, can be expected to support the flourishing of democracy.

It might seem that the truth of the claim is obvious. If a formal constitution establishes and protects the rights essential to the functioning of a democratic political system, then giving effect to that constitution through constitutional review, to turn it into a legal reality as opposed to a dead letter, must be conducive to the flourishing of democracy. Unfortunately, the simple combination of Kelsen’s positive and negative arguments does not suffice to establish that democracy must benefit from suitably democratic formal constitutionalism. There are at least two critical replies to Kelsen’s argument that are available to political constitutionalists.

The first arises from what I call the ‘argument from futility’. A proponent of the argument from futility agrees that the legitimating power of democratic procedure depends on more than just the majority rule. The democratic rules of the game must be observed by all parties, political rights must be respected, and there must be a real prospect of alternation in power. The argument from futility claims that the question of whether these background conditions of legitimate majoritarian rule are fulfilled depends entirely (or almost entirely) on political-cultural factors and not on whether there is a formal constitution bolstered by judicial review. Either a society does not need formal constitutionalism to establish a flourishing democracy, because a fully developed democratic political culture exists, or else formal constitutionalism and constitutional review will be powerless to prevent ‘democratic backsliding’, because neither key political actors nor the citizenry at large are deeply committed to democratic conventions.

The ‘argument from disagreement’Footnote ⁴⁹ presents a second challenge to Kelsen’s case for formal constitutionalism and constitutional review. It is best understood as a refined version of the claim that democracy is inherently incompatible with formal constitutionalism and constitutional review. A proponent of the argument from disagreement concedes that formal constitutionalism and constitutional review are not always futile. There may be cases in which judicial guardianship of a democratic formal constitution will turn out to protect democracy. Nevertheless, formal constitutionalism and constitutional review, while they are not inherently incompatible with democracy, are inherently incompatible, or so it is argued, with the full or perfect realisation of the democratic ideal.

The proponent of the argument from disagreement, like the proponent of the argument from futility, concedes that a flourishing democracy can exist only where there is general respect for the rules of the democratic game and the fundamental rights that sustain the democratic process. However, the argument goes on to claim that the interpretation of these rules and rights is subject to reasonable disagreement, even among citizens who are honestly committed to achieving collective self-government. The only democratic way to resolve such disagreement, the only way fully compatible with the fundamental assumption of equality in political status, is to leave such matters to the democratic process itself and not to rely on judicial institutions interpreting and applying a formal constitution.

Proponents of the argument from disagreement concede that it may be worthwhile to adopt formal constitutionalism and constitutional review in some instances. There may be societies whose political culture would not be able to support democracy without these crutches. However, the resulting instantiation of the ideal of democracy will be second rate in quality. It will fail to amount to a ‘core case’ of democracy. Therefore, formal constitutionalism should not be introduced where it is not needed to safeguard democracy, and we should remain aware that its presence, where it might be needed, indicates a deficiency in democratic development.

I will close by offering a defence, if only in very rough outline, of Kelsen’s case for formal constitutionalism and constitutional review against the arguments from futility and disagreement.

The soundness of the argument from futility depends on the truth of its empirical claims. To assess this argument, one would have to investigate empirically whether there are instances in which formal constitutionalism and constitutional review have prevented or at least impeded democratic backsliding. One would also have to determine whether such beneficial consequences of formal constitutionalism are more likely than effects detrimental to the proper functioning of democracy. An investigation of this kind cannot be undertaken here, of course. At any rate, the proponent of the case for futility bears the burden of empirical proof. What they must show to refute Kelsen’s defence of constitutional review is not merely that constitutional review can turn out to be detrimental to democracy, which is undoubtedly true, but that it should be expected to have that effect even where the content of the formal constitution that is to be given legal effect is focused on the protection of the integrity of the democratic process. That one can easily find examples of the abuse of constitutional review for purposes of ‘hegemonic preservation’ is neither here nor there – unless we think that the defender of constitutional review must show that all conceivable instances of the practice must be democratically legitimate.

What, then, about the argument from disagreement? I have already conceded that democracy can flourish without formal constitutionalism. However, we should be wary of the claim that the ideal of democracy is more fully realised in such instances than in formally constitutionalist polities. It may be that the seeming stability of successful democratic institutions that are not supported by formal constitutionalism is owed more to social, ideological, and ethnic homogeneity and to the absence of deep political conflict than to a principled commitment to the rules of the democratic game. Furthermore, the fact that the latter have never been made constitutionally explicit may, in a time of crisis, encourage and facilitate the adoption, in political practice, of a narrowly majoritarian understanding of democracy, one that identifies the will of the people with the will of a current majority, or even only with the will of a plurality whose support, due to vagaries of a distorting electoral system, suffices for a party to attain a majority of seats in a legislative assembly. If democracy is understood as simple majority rule, a minority’s opposition to the majority’s policies and its attempt to engage in political organisation and advocacy with a view to bringing about the democratic repeal of these policies may come to be seen as illegitimate and as a form of undemocratic resistance to the true will of the people. In that case, the minority will no longer be regarded as a majority in waiting but as a group of uncertain loyalty that consists of people who do not truly belong to the polity and who should not be offered a level playing field in electoral competition.

Proponents of the argument from disagreement would concur, I assume, that such a process would amount to a corruption of democracy and that it would turn democratic procedure into an empty shell that must fail to have legitimating force. However, if that concession is made, a blanket rejection of formal constitutionalism seems unnecessary. Formal constitutionalism can be a powerful symbolic repudiation of the majoritarian corruption of democracy. In light of recent events, it would be foolhardy and arrogant to blithely assume that the polities that supposedly qualify as ‘core cases’ of democracy will never find themselves in circumstances in which the opportunity for contestation afforded by formal constitutionalism might turn out to make a difference to the prospects of democracy.

In other words, the argument from disagreement does not speak to our current political predicament. It trades on the optimistic assumption that democracy can be taken for granted, at least in the more ‘advanced’ and ‘civilized’ nations of this world. The main virtue of Kelsen’s case for formal constitutionalism and of his democratic theory more generally is that it turns out, by contrast, to be surprisingly realistic. It does not trade in idealisations but asks how democracy can be stabilised in circumstances of conflict and crisis. This question is now more relevant than Dworkin and his political-constitutionalist opponents could have foreseen when they initiated the modern debate on the democratic legitimacy of formal constitutionalism and constitutional review in the long Indian summer of postwar liberal democracy. Kelsen’s unique blend of legal and political insight deserves much more attention.