Paolo Sandro, The Making of Constitutional Democracy: From Creation to Application of Law, Oxford: Hart Publishing, 2022, 317 pp, pb £59.50. In legal practice, we routinely distinguish the activities of creating law and those of applying it. The distinction not only informs the legal narrative of institutions such as courts, legislatures, and administrative agencies, but more generally it is also part of our everyday understanding in our political communities. Many students learn the distinction, lay people rely on it when conducting their affairs, and so on. The distinction exists in routine life, but is it theoretically sound? According to Paolo Sandro, there is a tendency of some scholars to assume the distinction and benefit from its explanatory fruitfulness but without first defending it.1 In contrast to this group, there is a second group of scholars who explicitly reject the distinction mainly on grounds of rule-scepticism. Here we find legal realists – North Americans, Scandinavians, and continental Europeans – as well as critical legal scholars and legal constructivists. Finally, there is a third group of scholars who, because of other philosophical commitments, undermine the distinction without perhaps intending to do so. Scholars who oppose judicial review of legislation on grounds of the evaluative and open-ended character of constitutional provisions – often leading to judicial law-creation (cf 72–79) – are examples. Against this background, The Making of Constitutional Democracy: From Creation to Application of Law (TMCD) sets out to provide a robust philosophical defence of the conceptual possibility of distinguishing between law-creation and law-application. Although the book title strongly hints at constitutional theory as its main object of study, TMCD is primarily a book on legal theory. Part of Sandro's ambition is precisely to show how complex debates on legal theory bear on the concept and philosophical foundations of constitutional democracy. Sandro's major efforts to establish the conceptual possibility of the distinction are motivated by his belief in one ideal of political theory and by what he takes to be a conceptual necessity derived from the concept of law he works with. The former, which Sandro calls ‘collective autonomy’ (170), points to the value of electing our own representatives for the task of creating law of general application, which can for that reason be treated as legitimate. According to Sandro, collective autonomy requires at least the possibility of distinguishing between law-creation and law-application: ‘For if there is indeed no possibility of applying a legal rule created by a different institution at a previous moment in time, then our current constitutional-democratic frameworks are effectively empty vessels which conceal a power relationship between public authorities and citizens that is very different from the one upon which constitutional democracy is grounded’ (2, emphasis in original). At the root of the distinction, therefore, lies the ethos of modern constitutional democracies. But the need to establish this distinction is also derived, as I said, from the concept of law that Sandro consistently endorses throughout the book: ‘I understand “law”’, says Sandro, ‘as a specific social technique for subjecting human conduct to the guidance of rules (and rulings)’ (24). Yet for the law to fulfil this action-guiding mission, it must be at least conceptually possible to differentiate the act which posits a legal rule from the act which applies it. Put differently, ‘the creation of a normative standard after the relevant conduct has taken place cannot by definition have fulfilled that function [of guiding action], at least vis-à-vis all those forms of behaviour which took place before the (judicial) act of law-creation’ (169). Sandro calls this type of conceptual necessity law's ‘action-guiding requirement’ (17). It is the realisation of collective autonomy, on the one hand, and the fulfilling of law's action-guiding requirement, on the other, that motivate Sandro's project: ‘my ultimate intention’, he says, ‘is to show how the very possibility of distinguishing between the activities of law-creation and law-application represents a necessary (but not sufficient) condition of legitimacy of modern constitutional democracies’ (24, emphases added; see also 212–213). Now you might think that if the threshold is conceptual possibility, then it is a modest one. After all, all that Sandro needs to show is that not every act of law-application is, at the same time, an act of law-creation (cf 169–170). Provided, in other words, that Sandro shows that it is meaningful to speak of acts that ‘merely’ apply the law (without entailing any creative activity from the law-applier (cf 217), he will have succeeded. The reality, however, is that the threshold is high. Three important obstacles stand in Sandro's way; first, rule-scepticism, in its variety of philosophically sophisticated forms. Secondly, legal discretion, affecting law's determinacy. Finally, the alleged impossibility of fixing communicative content before pragmatic enrichment and, consequently, in legal communication, of fixing legal content before legal interpretation. Furthermore, the rigorous way in which Sandro deals with each of these challenges, plus the impressive arsenal of analytical tools he deploys for that task, are further proof of Sandro's ambitious project. The book is also rich in historical, philosophical, and comparative analysis; its bibliographical corpus is, as Felipe Jiménez put it, ‘simply breathtaking’.2 And if, when he speaks of ‘bridging the gap’ between common-law and civilian scholarship (10; cf 162), Sandro means engaging with both familiar giants and rising names in these two legal traditions, then he has undoubtedly succeeded. I think of this book as an attempt to pursue the noble cause of making law possible by distinguishing, and thereby treating differently, the act that creates a norm from the act that applies it. Sandro, we saw, attaches political and jurisprudential value to this distinction. But we also find, recurrently throughout the book, an analytical tool being used for penetrating complex debates in political theory, constitutional theory, administrative law, philosophy of language, and jurisprudence. At some points, it seems as if the distinction is only an instrument for introducing heavy analytical artillery and reshaping certain debates (such as separation of powers) while casting doubt on old ones (such as the Hobbesian state of nature). This argumentative strategy points to another feature of Sandro's book. His account is, as he would put it, ‘unified’ (cf chapter 4) in the sense that all the major claims advanced in the book are tied together around the axis provided by the distinction. Given the complexity of the book and its importance for jurisprudence and constitutional theory, I believe one of the best ways to do justice to it is by offering a careful exposition of its main arguments and some critical and constructive feedback. To this end, we can break down TMCD into three parts. The first part (chapters 1 and 2) puts forward the main argument; the second part (chapters 3, 4 and 5) defends the argument against three possible objections. The final part (chapters 6 and 7) articulates a view of law-application and defends a new approach to separation of powers. The book begins by laying out the basic methodology, then explains the relevant notions of law and political power, and finally it elucidates the conditions for the existence of formal legal systems and political authority. Sandro devotes a significant part to explaining his methodological commitments, which could be characterised as a compromise between conceptual and empirical analyses. Rather than seeing these ‘in opposition to one another’, Sandro argues we should see them as ‘complementary’ (22). This middle ground, says Sandro, is particularly fertile for analysing law as a normative practice. For one thing, conceptual analysis may illuminate, in ways which empirical analysis may not, the normative character of law – by explaining, for example, how the reasons for action the law purports to give may figure in the explanation and possible justification of someone's actions. For another thing, law is a normative practice; as such, it makes little sense to prescind from the results of fact-based disciplines such as history, social anthropology, and natural sciences (cf 21). The main upshot of this methodology3 is that we have good reasons to resist a particular theory if it fails to be sensitive to salient features of empirical reality. And there appears to be one clear example of how this methodology cashes out. After explaining the concept of law and political power that are relevant for the book (23-30),4 Sandro elucidates the conditions for the formation of modern legal systems and political authority. He endorses, as we will see, H.L.A. Hart's thesis of the union of primary and secondary rules (38),5 but having first rejected Hobbes’ account on the ‘passage from the pre-political to the political condition’ (32). Hobbes, we know, thought that the only way to escape the state of nature, in which people strive to survive come what may, is by centralising authority and giving it far-reaching powers to secure peace.6 The problem with Hobbes's analysis, says Sandro, is that there is ‘growing archaeological evidence’ suggesting ‘a quite different picture’ from that offered by Hobbes (32). In particular, this evidence would challenge two central assumptions in Hobbes’ analysis. According to the first assumption, a centralised form of authority is necessary for achieving stability and prosperity. Yet, according to Sandro, the example of certain city-states of Ionia, where individuals would neither rule nor be ruled (isonomia), and where normativity would be mainly customary, ‘invalidates the Hobbesian thesis that such centralised power was necessary to overcome the “state of nature”.’ (34). The second assumption is that the centralisation of authority does not always improve the material conditions of the population involved. As the case of some Mesopotamian states formed around 6000 BCE shows, ‘the centralisation of power, at least at its inception, seems to have actually benefitted only a very small part of the population – the rulers – while leaving the vast majority in the same, if not worse, conditions’ (35). But there are two problems with Sandro's move – apart from the fact that nothing substantive in TMCD turns on Hobbes. First, it is not clear whether the evidence Sandro provides gives us sufficient reason to ‘reject once and for all the Hobbesian tale of the ineluctability of the state’ (11). A threshold problem concerns how we determine the economic, political, and anthropological conditions that allowed these isonomic societies to flourish – if they ever did. Here, barring some brief remarks (at 33), Sandro does not offer a substantive engagement with the secondary literature he relies on. This possible mismatch between these societies and the early modern England surrounding the Leviathan, may cast doubt on the relevance of the former for the latter. Sandro does not consider this possibility. Secondly, how should we understand Hobbes’ claim? Should we take it as a claim on social and legal anthropology or as a normative claim on legal and political theory? The risk of Sandro misfiring here is evident. To undermine Hobbes’ claim Sandro needs to show, first, the type of project Hobbes was pursuing and the extent to which (and why) that project is undermined because of two counterexamples. And the problem with this move is that it can also backfire on Sandro himself. Why couldn't Sandro's theory be undermined by a set of counterexamples? Arguably because while Sandro only makes a case for the conceptual possibility of the distinction, these examples may at most undermine its empirical credentials. Yet recall that as a result of his methodological compromise, Sandro has allowed for the possibility that empirical reality permeates conceptual analysis. One way out for Sandro is to take his methodological compromise as offering some argumentative shield against counterexamples. The account can tolerate a certain margin of mismatch with reality if it is not seriously out of tune. And this is precisely one way, I think, in which we could also read Hobbes. The first part of the book concludes with an analysis of the conditions for the existence of political authority and those which make it legitimate. Existence conditions, says Sandro, ‘are always the same as a matter of conceptual necessity’ while legitimacy conditions ‘are contingent and context-sensitive’ (37). Following the seminal work of Brennan et al,7 Sandro claims that ‘all social norms are constituted by clusters of normative attitudes that exemplify acceptance of a given normative principle … plus mutual knowledge of those normative attitudes between members of the relevant social group’ (37). On this view, the main function of norms is to create spheres of accountability which allow the coordination of small and homogeneous social groups. When these groups become larger and more complex, the initial bulk of primary social rules becomes inefficient (40). As a result, the corrective layer of Hartian secondary norms (of change, adjudication, and recognition) is required (38). The system of norms becomes formal, with one crucial consequence: ‘a distinction between ruler(s) and ruled emerges’ (41). For Sandro, political authority, which ‘cannot be but rule-generated authority’ (41), thus comes into life. The second chapter turns to the legitimacy conditions of political authority. The main claim can be summarised as follows. The passage from non-formal to formal legal system, in which political authority wields the right to rule over others, brings with it the problem of limiting such an authority – Sandro focuses mainly on democratic political authority. The democratic realisation of political power, on the one hand, and its limitation, on the other, become two concerns giving rise to two different ways of legitimating law – the main upshot of political power (cf 270–271). Law is legitimate if it is the product of democratic self-rule or if it provides safeguards against (democratically generated) threats on the democratic system and people's fundamental rights. Sandro traces these two sources of legitimacy to two historically different ways of understanding law: law as lex and law as ius. In the former, law is the realisation of political power; in the latter, its limitation. Constitutionalism emerges as the limitation of law-as-lex by law-as-ius. In this complex structure of juridical checks (‘legal otherness’, says Sandro) lies the root of constitutional democracies. But for this structure to obtain, it must be at least conceptually possible to distinguish between law-creation and law-application (70, 76–77). This is a remarkable chapter. Possibly its greatest virtue is to offer a general account of the way in which these two types of law (lex and ius) interplay, and how they bear on the emergence of constitutional democracies. Sandro's efforts to give historical context to ius, first in the late Roman Republic (51) and then in England's common law (66-68), are also commendable. If I may press the account a little, I would raise two comments; one on the legal authority of British courts to impose ius over lex today; the other on the grounds for treating ius as juridical. Consider the following passage: ‘Properly conceived of, constitutionalism requires the juridical limitation of the law-making power of the legislature by means of a second type of law that is not disposable by the law-maker via ordinary means, and in some cases not disposable at all’ (71). This juridical limitation, says Sandro, can be concretised via entrenching constitutional provisions (out of the legislature's reach) or by allowing for another type of law, different both in its source and administration to statutory law, to place constraints upon the legislature. Examples of the former we find in the constitutional framework of many civilian jurisdictions; an instance of the latter is Britain's common law. When it comes to British courts, though, two related questions arise: first, on what legal authority can British courts impose ius over lex today? Secondly, do the powers for imposing these limits, as well as the limits themselves, need to be recognised ex ante by law? A British court that creates for itself (ex post) the power to impose limits on primary legislation is a court that stretches its legal authority in ways that might appear controversial vis-á-vis the orthodox theory of parliamentary sovereignty. What courts can do, and often do, is make their way around the statutory language to achieve some desired results. This caveat explains, I think, the following passage: ‘… what British constitutional history indicates is that the courts have found ways to enforce limits upon the legislative authority of Parliament without formally striking down statutes, but rather through creative construction that often goes beyond the letter of the law (under the pretence of giving effect to the intention of the legislator)’ (70, emphases added). But because it is not clear on what legal basis Sandro thinks that British courts can impose ius over lex, doubts arise as to whether the highly contingent relationship between British courts and parliament delivers Sandro's robust notion of ius. At some points,8 it seems as if Sandro is rather interested in advancing a normative account of how courts ought to deal with primary legislation – including, exceptionally, the power to invalidate it.9 If British courts do deliver Sandro's notion of ius, then the question arises as to whether they do so because of the specific content of the limits they impose or rather simply because they impose some limits – whatever these are.10 By contrast, if Sandro wants to defend a normative account of how courts should deal with primary legislation – including, as I said, the exceptional power to invalidate it – then one thing he needs to show is the legal basis, as well as the limits, of those legal powers and whether they need to be previously conferred by law. This lack of definition – concerning the legal basis for imposing ius over lex – also stems from not knowing the grounds on which ius’ limitations are juridical. Are these limitations juridical because they are based on positive sources or because they follow, a la Dworkin, from the best moral justification for the practice of limiting lex via ius? Sandro seems to reject the latter approach (see note 16 below), but this rejection would add more pressure on his account of the legal basis for British courts to impose ius over lex. If, conversely, the limitations are juridical because they are sourced, then the problem of the existing legal source for British courts to impose them upon lex comes to the fore again. (And I doubt, to be sure, that a Dworkinian would find it satisfying to say that ius’ limitations are juridical in virtue of the content-indifferent fact that they have been institutionally posited through a judicial decision). What we miss in TMCD is an account of the grounds for treating ius as juridical and for courts to lawfully impose ius over lex. The first part of the book concludes by showing the extent to which constitutional democracies are premised on the distinction between law-creation and law-application. In doing so, Sandro considers one objection against judicial review of legislation: the open-ended and evaluative character of constitutional provisions that makes it difficult to distinguish, in constitutional adjudication, between law-creation and law-application. According to Sandro, there are two main problems with this objection. First, it assumes that vagueness and evaluative language are exclusive of constitutional law, whereas they are pervasive as well in statutory law (73-74). Secondly, the objection makes an unwarranted generalisation: it does not follow from the fact that constitutional adjudication in the US will often be political (by contingent features of that legal system) that the same will be true of all legal systems under an entrenched constitution (74; cf 103–104). For Sandro, the legitimacy of judicial review needs to be assessed on a case-by-case basis and taking into account the specific institutional and normative features of each legal system (75). The first part of TMCD grounds the distinction between law-creation and law-application. The second part defends it against three challenges: rule-scepticism, legal discretion, and one view about the relation between law and language. According to the first challenge, there is no such thing as the distinction between law-creation and law-application; legal rules are largely indeterminate and therefore they are ‘unable to justify judicial decisions’ (91). If the law is thus indeterminate, there is, by hypothesis, no legal rule to be applied. Law-application collapses thus into law-creation. Rule-scepticism, says Sandro, is the ‘lowest common denominator of realisms in law’ (ibid). We may find different philosophical and political commitments underpinning the various forms of legal realism,11 but the gist of it, its conditio sine qua non, is rule-scepticism. Sandro's discussion of rule-scepticism is analytically rich, carefully documented, and comprehensive. For ease of exposition, let us break down his argument into three steps. The first step consists in drawing a distinction between what we could loosely call ‘cognitive’ and ‘non-cognitive’ approaches to adjudication.12 Roughly, the former views adjudication as a process where the judge first establishes the material facts of the case, then identifies the legal rule applicable to those facts, and decides the dispute by applying the former to the latter. Non-cognitivism, by contrast, views adjudication as a largely unconstrained process in which the outcome of the dispute is often determined by the judge's personal reaction to the facts of the case, followed by an ex post and tailormade rationalisation (in legal terms) of the judge's decision. A second step in the argument consists in drawing yet a further distinction between moderate and radical versions of cognitivism and non-cognitivism, and to reject the two radical versions on grounds of descriptive inaccuracy. According to Sandro, radical cognitivism (the rule-formalism that characterised the Enlightenment period) is ‘grossly mistaken’ and should be treated as an ideology (83). On the other hand, radical non-cognitivism,13 according to which the law is globally indeterminate, ‘has been exposed as an epistemically untenable position, a position that simply fails to account for our legal phenomena satisfactorily. The overall workings of law cannot be explained … by a purely empirical observation of what is going on in courtrooms. Too much is left unexplained …’ (102; see also 110–115). So we are left with moderate cognitivism and moderate non-cognitivism. The former, advanced by Hart in response to the American Legal Realists, acknowledges the existence of unclear cases in which the law is indeterminate but locates the source of such indeterminacy outside the law, in the open texture of language. As such, the problem of law's indeterminacy must be understood as part of a wider problem; barring these cases, though, ‘the law discharges its function of communicating effectively standards of conduct’ (84). On the other hand, moderate non-cognitivists, such as Brian Leiter, argue that the law is not globally but only partially indeterminate – particularly in so-called ‘hard cases’, at the appellate level. In ‘easy cases’, by contrast, the law is determinate (102). Sandro's third step is to argue against moderate non-cognitivism, as advocated by Leiter – and also, to some extent, by Riccardo Guastini.14 Drawing from Dan Priel and Frederick Schauer, Sandro argues that it seems difficult to maintain law's global indeterminacy, as the Realists did, while also endorsing legal positivism, as Leiter does. For legal positivism is premised on legal determinacy; were the law globally indeterminate, it could not claim authority in the first place – that claim would be self-defeating (106). Yet if Leiter drops global indeterminacy (so as to be able to maintain both moderate rule-scepticism and legal positivism) much of his naturalised project loses bite. If the number of hard cases is but only a few (moderate indeterminacy), this data does not warrant Leiter's invitation to replace traditional legal analysis (focusing on normative explanations of judicial decisions) for a naturalised jurisprudence (focusing on causal explanations). Leiter cannot, via a naturalised jurisprudence, aim to offer ‘a general account of the legal process as a whole’ (ibid) while also endorsing legal positivism. His moderate rule-scepticism comes at the price of jurisprudential parochialism (as opposed to a general account of law) and eventually abandoning his naturalised project or admitting its superfluousness. Alternatively, Leiter can aim for a general account of law but at the cost of dropping his moderate rule-scepticism and endorsing radical non-cognitivism. The upshot of Sandro's criticism is that moderate non-cognitivism is either untenable or comes at a high price.15 Especially in regard to Leiter, Sandro thinks that his philosophical commitments lead him to radical non-cognitivism (107). Having previously rejected radical cognitivism and radical non-cognitivism, Sandro's critique of moderate non-cognitivism paves the way for claiming that the only model of adjudication compatible with democratic constitutionalism is ‘a (tendentially) cognitive model’ (88). One thing Sandro appears to almost take for granted, though, is that Hart's cognitivism is the default model of adjudication: ‘if you agree with my rebuttal of (any version of) moderate scepticism, then we are left with a simple choice: either (moderate) cognitivism or radical scepticism’ (112). Here, the name of Ronald Dworkin bubbles up again.16 Surely Dworkin's views on adjudication put him alongside Hart in the cognitivism/non-cognitivism divide – if it is possible to fit him there at all.17 As illustrated by his fictitious judge, Hercules,18 Dworkin's cognitivism appears to be more demanding (on judges) than Hart's. Dworkin, moreover, makes plenty of room for the distinction between law-creation and law-application. Recall, his first objection to Hart was levelled precisely against the ‘moderate’ aspect of Hart's cognitivism, ie, Hart's notion of discretion entailed (for Dworkin) that people did not have enforceable legal rights (prior to litigation) and that judges would create law instead of applying it – via legal principles.19 Sandro does not consider the possibility that Dworkin might also have a claim to provide the default model of adjudication. Let me close this section by warning against the idea of taking at face value Sandro's dismissal of a naturalised jurisprudence. It is true that, by rejecting any middle ground between moderate cognitivism and radical non-cognitivism, Sandro translates choice between these two approaches as a binary decision – instead of a continuum where we might have degrees of rule-scepticism and cognitivism. But it does not follow, in my view, that the same holds with respect to the tools we use for studying jurisprudence. The decision to embrace cognitivism does not entail dropping the idea of benefitting from empirical analysis. Likewise, committing to rule-scepticism does not lead one to abandoning traditional legal analysis. But one's choice to embrace cognitivism does commit one to giving up, according to Sandro, adopting a ‘purely’ empirical approach to the study of jurisprudence (cf 102, 112–113). This methodological cost, already anticipated in chapter 1, follows from Sandro's belief in the incapacity of a purely descriptive or empirical theory to account for the normativity of law. For ‘[t]o know what the law is (this is what a theory of law aims at) is equivalent to knowing what a decision maker ought legally to do in the case (and this is what a theory of adjudication does)’ (111). Assuming, of course, that the gap between what the law is and how the case at hand should legally be decided is null (110). In these, so-called ‘easy cases’, law's normativity is exhausted by how a judge legally ought to decide the case. (For the record, I disagree with this approach).20 This view on the normativity of law in easy cases leads Sandro to reject the value of a purely empirical approach to jurisprudence. But what about a hard case? Here, Sandro's elusiveness (cf 111) should be read in light of the proviso above (‘purely’). In other words, we might still benefit from empirical analysis (and greatly so) for studying hard cases and what causes their judicial decisions. The second challenge is the existence of discretion in law. Chapter 4 does not directly deal with the challenge; rather, Sandro puts forward what he calls a ‘unified’ theory of discretion; an account which merges ‘jurisprudential and administrative law insights’ and which ‘bridges the gap between common law and civil law systems’ (14). In doing so, he brings analytical clarity to the topic and paves the way for what I take to be a more targeted regulation of discretion. The discussion can be structured around three parts; the first part provides a rich discussion of the idea of discretion in legal theory, particularly as it figures in the work of Hart (including his ‘lost essay’),21 Dworkin, Kelsen and Klatt (119-142). Sandro's sophisticated and comprehensive analysis makes TMCD one of the best places to start grappling with