1. Introduction
Courts do many good things. Judges carefully consider individual claims and arguments,Footnote 1 and contrast them against the law in light of evidence. Their decisions are argued for, are public, and can be contested in form and content in different hierarchical stages. Additionally, and among other things, these practices are said to contribute to the will-formation of the public sphere and improve the quality of the legislative process.Footnote 2
These are good things, indeed. Yet, there are some virtues that some scholars claim judges have that seem to require further scrutiny before accepting such characterisations. Consider, for example, the notion advanced by some legal scholars that judges are good at morality,Footnote 3 a thesis that has been strongly criticised.Footnote 4 Or the claim that courts engage in deliberation, some even placing the judiciary as the institutional embodiment of central principles of deliberative theory.Footnote 5 Again, and for good reasons, that association has been questioned.Footnote 6
Here, I examine another category associated with the judiciary: political representation. Scholars have generally used three types of arguments to claim that the judiciary is a representative institution. One relies on the fact that judges are appointed through the political process.Footnote 7 A second avenue is to focus on the fact that in judicial procedures, the representation of clients by their attorneys is mediated by judges.Footnote 8 These are important discussions to have, but here I am interested in a third important argument that has not undergone proper scholarly scrutiny.Footnote 9 It is as follows: some commentators take the judges’ duty to argue for and justify their decisions as reasons supporting the claim that they perform representative functions. This claim is usually advanced when the representation of the people at the constitutional level is at stake. These authors thus consider that judicial decisions would be able, due to their argumentative features, to represent the constituents’ better self as expressed in constitutional principles. Conversely (and perhaps ironically) it is implicitly assumed that traditional representative institutions fall short of accomplishing these desiderata.
This article provides reasons to reject the contention that argumentative representation is something characteristic of or to be demanded from judges in a constitutional democracy. Judges look at the law, not at the will of constituents; their practice and the norms ensuing from it are not to be evaluated on the basis of their reflecting or their answering to the will of no-one, but on the basis of their compliance with the requirements or mandates of the legal system. Those norms and practices may effectively, and even oft-times, coincide with what the public prefers or with its best interests. Such coincidences are, however, contingent, and differences between the content of a judicial ruling and the preferences, interests, and so on, of the people they affect with their decision, are not best accounted for as failures in representing or as breaches of some alleged representative function. With this feature in the background, which I will refer to as ‘functional orientation,’ I will show that the notion of an argumentative judicial representation judiciary falls prey to two types of mistakes: descriptive and normative. Descriptive because when claiming that the judiciary is a representative institution due to their argumentative duties, scholars fail to properly account for what courts do and should do. Several other terms of legal and political theory account for what happens in the courtroom, but they do not map squarely onto the concept of representation. And, to be clear, this is not merely about words. Each category conflated with what is allegedly judicial representation (e.g., legitimacy, adjudication, interpretation, etc.) is significantly laden with its own distinctive meaning, and raises theoretical disagreements about its own scope, normative relevance, etc.
But scholars also incur a normative problem. Underpinning the notion of argumentative judicial representation is a concern with imagining mechanisms by which constitutional democracies are kept safe from the sound and fury of short-term, electoral party politics. Courts do this allegedly by embodying something other than transient electoral majorities, namely higher, principled forms of the people inbuilt in the constitution. These may be fine objectives but, I will argue, putting the burden of representation on the courts is a not a suitable path to their achievement. Nothing in this endeavour impedes what we traditionally refer to as representative institutions from representing higher, more principled, forms of the people. I will thus argue that scholars should avoid talking of judicial representation on the pain of incurring two types of fallacies: moralistic and nirvana. Moralistic because since representing is important and it is not properly realised by majoritarian institutions, it is assumed that courts incarnate it. This reasoning, however, unwarrantedly shifts from a desideratum to an affirmation about how the world is. And nirvana because the idea of judicial representation rests on an unfair contrast between the ideal features of the judiciary and the ‘real’ practice of majoritarian institutions.
These questions are related to other pressing issues of legal and constitutional theory, for example, to which domains are “inherently unsuited for adjudicative disposition and should be left to the legislature.”Footnote 10 Yet, my interest here is independent of how these questions are answered. My reflection is rather closer to Weale’s question: “on what grounds can we assign an institutional right to one type of political agent rather than another?”Footnote 11 My claim is that the powers given to judges are not grounded on their putative representativeness.
I proceed as follows. I begin describing three theses claiming (or lending support to the claim) that the judiciary is or can be a representative institution due to its argumentative functions or capacities: Robert Alexy’s ‘argumentative representation,’ Pierre Rosanvallon’s ‘reflexive legitimacy,’ and Philip Pettit’s ‘indicative representation.’ These three are, in my view, the best and most articulate versions of the argument I target here.Footnote 12 I abstract some of their common features under one thesis I call the ‘judicial representation thesis’ (JRT) (section 2).
I then examine political representation. I am interested in identifying a feature found in different views of representation in the circumscribed domain of democratic law-making procedures that few would take as controversial. This feature is the ‘functional orientation’ of representatives towards the will of their constituents. Representation may be more than that,Footnote 13 granted, but the difficulties in specifying what this animal exactly looks like demand this type of methodological stipulation. I conclude that, notwithstanding the disparities of the different conceptions of political representation available in the scholarly literature, we can observe a broad convergence on the etymological definition of representation, to wit making something present again, in this case, someone’s will. This feature shall be at the background of the remainder of the article (section 3).
The question, “is the JRT warranted?” then hinges on whether courts have this functional orientation. My claim is that they do not. The job of courts is not best described by the contention that their duty to argue for and to justify their decisions is a proxy for representation (section 4).
I flesh out the argument in two subsections. My first contention is that the JRT is better described as comprising sundry practices that judges indeed perform and principles that they indeed embody, such as adjudication, argumentation, justification, interpretation, reasoning, legitimacy, legality, inter alia, not representation. Representation may incorporate those features, but the converse—that these features are necessarily linked to representation—does not necessarily hold (4.1). Second, I delve into the reasons why scholars may seek to find representation of the people in the constitution in the judiciary rather than in where it intuitively seems like its more natural place, namely what we ordinarily refer to as representative institutions. I submit that the reason is the longstanding image of majoritarian institutions as incapable of engaging in principled discussions about constitutional politics. Since majoritarian institutions supposedly fail at engaging in the type of discourse and at meeting the justificatory demands required by constitutional politics, the will of the people enshrined in a constitution should be represented elsewhere; courts are the main candidate for filling that role (4.2). As I mentioned above, however, this reasoning incurs a moralistic (4.2.1) and a nirvana fallacy (4.2.2). The final part offers concluding remarks (section 5).
2. Judicial Representation
Some scholars see the judiciary as a representative institution due to their duty to argue for and justify their decisions. They share what I call the ‘Judicial Representation Thesis’ (JRT):
JRT: Courts are argumentative representative institutions. By expressing (a) certain constitutional objective(s) through their arguments, justifications and decisions, they retrieve and/or express the will of the people inbuilt in the constitution.Footnote 14
This section overviews three accounts that are, to different degrees, expressive of the JRT: Robert Alexy’s ‘argumentative representation’ (2.1), Pierre Rosanvallon’s ‘reflexive legitimacy’ (2.2), and Philip Pettit’s ‘indicative representation’ (2.3). The JRT will be my general target, but I will address these three authors directly when needed, for the sake of nuance.
2.1. Argumentative Representation
Alexy’s ‘argumentative representation’ is part of a scheme aimed at justifying judicial review of legislation through his theory of balancing rights and the use of his weight formula.Footnote 15 The overall theory yields a system of constitutional rights and freedoms oriented towards correctness.Footnote 16
Alexy wants to make sense of the fact that courts of law lacking electoral legitimacy are given a prominent place in the determination of the meaning and scope of constitutional provisions in democratic systems. That is, in systems whose legitimacy is predicated upon the will of the people, not the will of judges. As he acknowledges, constitutional review by judges sits at odds with this requirement.Footnote 17 In his view, “[t]he only way to reconcile constitutional review with democracy is to conceive of it … as representation of the people.”Footnote 18
Alexy quickly spots two problems. First, “representation seems necessarily connected with elections, and, second, if representation were possible without elections, why should this kind of representation have priority over representation based on elections?”Footnote 19
The author circumvents these difficulties by conceiving of representation as a relationship between represented and representative whose existence in the case of Parliament is almost primarily determined by elections and majority rule. But, the idea goes,
[a]n adequate concept of democracy must … comprise not only decision but also argument.… For this reason, the connection between the people and the parliament must not only be determined by decisions expressed in elections and votes but also by argument.Footnote 20
Courts generally lack electoral legitimacy, so their relationship with the people is ‘purely argumentative.’ Because Alexy considers that representation “necessarily lays claim to correctness … a fully-fledged concept of representation must include an ideal dimension, which connects decision with discourse.”Footnote 21 This ideal dimension is realised by courts.
Finally, Alexy recognises that “the existence of good or plausible arguments is enough for deliberation or reflection, but not for representation.”Footnote 22 Yet, he thinks that a middle point between arguments and numbers is possible if a sufficient number of rational people accept those arguments for reasons of correctness.Footnote 23
2.2. Reflexive Legitimacy
Another version of the JRT is Rosanvallon’s idea of ‘reflexive legitimacy,’ an attempt to correct and compensate for what he takes as three flawed considerations of political theory: that voters choice is equated with the general will; that voters are equated with the people; and that politics and legislation flow continuously from the moment of the vote.Footnote 24 ‘Reflexive legitimacy’ questions the view of the people as a monolithic unity with a single will. Rather, the people exist in different times and styles of deliberation, under different labels: the electoral people, the social people, and the people as principle.Footnote 25
Each of these varieties relates to the general will in different ways. The electoral people corresponds to the numerical definition of the general will. The social people is mainly incarnated in the institutions of what Rosanvallon calls counter-democracy, comprising different means by which citizens may monitor elected rulers, mobilize resistance to specific policies, and control corruption.Footnote 26 Finally, the people as principle “refers to an inclusive, egalitarian idea of the general will, grounded in full respect for the existence and dignity of each individual.” Neither of these three images of the people “can by itself claim to be an adequate incarnation of the democratic subject.”Footnote 27 We need the three.
These distinctions bear on the alleged representativeness of constitutional courts. Rosanvallon avows that, “[a]s reflexive third parties, the primary function of the constitutional courts is social and political representation.”Footnote 28 But it is a rather special type of representation by which courts “attest to the existence of the people as principle.”Footnote 29 Their essential role is to make clear “that the sovereign is more than just the party that wins a majority on election day and that no definition of it is sufficient.”Footnote 30
Rosanvallon thus endorses the JRT inasmuch as he thinks that qua reflexive institutions, courts embody a form of representation “of the moral or functional order, structurally distinct from the immediate expression of opinions and interests, which is what elections are all about.”Footnote 31 They strengthen the representative system by playing “a positive role in structuring democracy,” creating more faithful images of the general will,Footnote 32 and preventing “the future from being foreclosed by the party that happens to be in the majority at a particular point in time.”Footnote 33
2.3. Indicative Representation
Representation, says Pettit, may be responsive or indicative in character. In a responsive model, the preferences, attitudes, opinions, etc., of representees “are the causal source of the attitudes displayed by the representer.” Accordingly, representatives should speak or act on the attitudes of those they represent. In an indicative model, by contrast, “the attitudes displayed by the representer are a non-causal sign of the attitudes held by the representee.”Footnote 34 Indicative representers stand for the representees; their actions are indicative of how the representees would act. Just as a painting is indicative of how the subject of the painting looks, “so on this image should representers be indicative of representees; they should be fitted to serve as proxies, not—or not just—as deputies.”Footnote 35
‘Indicative representation’ is presented not only as a plausible abstract ideal, but as one that is manifested in different periods in the history of democracy and in concrete democratic institutions, primarily in proportional representation.Footnote 36
Yet Pettit also suggests that once we understand the distinctiveness and value of ‘indicative representation’ we may then see that some institutions, although not directly elected, can still be accounted for as representative, and so possessing democratic credentials on their own.Footnote 37 Placed at the constitutional level, we can therefore conclude that if these institutions can point at principled ways of interpreting the constitution, we may be able to link such interpretation to the will of the people on certain constitutional matters. Indicative judicial representation would respond to the objection that judges do not track the preferences of citizens, simply by granting the point, as their representativeness does not depend on their retrieving no one’s preferences. Judicial decisions at the constitutional level then are an indication of someone else’s preferences, voices, opinions, etc., expressed in the constitution, not a retrieval of them.
3. Representation and Functional Orientations
One way of determining the cogency of the JRT hinges on how the term representation is used—on its meaning. But determining its meaning is no easy task. As one of the most prominent theorists of representation claims, there is a “striking lack of precision in the literature on political representation.”Footnote 38 The concept is complex, ambiguous, vague, and underspecified. Hence, contrasting the or perhaps even a concept of representation against the JRT leads most likely to a dead end. I would fall prey to the objection that there are other conceptions competing with the one that one prefers.
I will not solve these problems in one subsection. Yet, I think this conceptual mix-up can be attenuated by circumscribing my analysis to what representation does in democratic law-making contexts. That is, in contexts in which the quarrel about the idea of judicial representation finds its place. We care about whether courts, or any institution for that matter, are representative, qua institutions that are part of a democratic system. This does not settle discussions about what political representation is, but it does provide a general vantage point from which different conceptions of political representation may be seen as broadly converging.
Hence, in this section, I submit, without necessarily committing myself to one particular view on what representation is,Footnote 39 that a fixed point in theories of political representation in democratic law-making contexts is that representatives are mandated to trace their decisions back to the will of their constituents; to making their will present again. That is, every conception of political representation concerned with democratic collective decision-making, relates to the relationships between the actions of an individual or a group A, the imputation of those effects to another individual or group B, and the evaluation of the latter about the former on the basis of a criterion X, where X stands for B’s preferences, wills, desires, interests, reasons, etc.
I insist: this is what representation is supposed to do in these particular circumscribed democratic contexts. Representation then gives representative democratic institutions a ‘functional orientation’: agents and institutions engaged in processes of collective decision-making are reasonably accounted for as representatives when the warrantedness of their practices and their results hinge to some higher or lesser degree on their duty to either advance, seek, or pursue the interests of a constituency, or to track their preferences in a manner responsive to it.
It could be objected that this description is limited to what Rehfeld terms the “Interest and Responsiveness Account” of representing. If, arguendo, we take Rehfeld’s analysis as correct, my account could be deemed incomplete, as it would not consider the full panoply of phenomena falling under the label ‘representation.’ Representing as such, the objection would go, “does not require the pursuit of another person or group’s interests, nor does it require that an action be done in a manner responsive to them.”Footnote 40 Instead, Rehfeld tells us, the activity of representing has three senses, unrelated to interest and responsiveness. I will consider these senses for they articulate the objection and allow me to dispel it in an orderly fashion.
The first sense refers to the exercise of the social power that individuals acquire when they take on the social status of representative. In contexts of democratic decision-making, “this often occurs after an individual is elected; in non-democratic or non-institutional contexts this can occur through informal audience recognition, often including the process of ‘claims-making.’”Footnote 41 In this first case, representing happens if the actions undertaken are done by someone using a social power: “[w]hen those actions are not an exercise of a representative’s social power, they are not cases of ‘representing,’ they are merely cases of voting, deliberating, negotiating, and so on.”Footnote 42
The second sense derives from the first and refers to those instances where individuals holding the social power to act as representatives are not performing the activities typically attributed to them qua representatives. In Rehfeld’s example, this sense “explains the case in which we might say, ‘Karen represents the third district in Congress,’ even when Karen is sleeping.”Footnote 43
The third sense refers to the act of “conveying one thing for another,” much as symbols convey a message, or language conveys a meaning. This sense is not related to the exercise of social powers but “merely refers to the successful conveyance through language or other means of one thing for another.”Footnote 44
Nevertheless, none of those senses prove the JRT. The first two senses identified by Rehfeld are question-begging for the purposes of my argument. Both versions assume that the office is already one occupied by someone with social status of representative. For example, Alexy’s explicit acknowledgement that the representativeness of the court is not based on election or numbers implies that additional argumentative steps are to be taken to grant the judiciary such status. If anything, this recognition shows that the judges’ authority is one that does not naturally derive from their status as representatives. In contrast, my discussion presupposes that the nature of judicial office is not of such kind. My discussion is about whether the actions of officials whose selection is—barring exceptional cases where they are elected—supposed to place them outside the political sphere, can be nevertheless taken as representative because the content of those actions could be indicative of something akin to the people’s will. The same applies to the second sense as it is “derivative on the first.”Footnote 45
The third sense poses, prima facie, a bigger challenge to my account. It implies, just as Rosanvallon’s and Pettit’s accounts do, that there are ways in which we can talk about representation that are detached from the particular function the official is supposed to carry forward. Forms of representation emerge, the argument goes, from the actions of the court whether its members have been selected for those purposes or not. One can still analyse the performance of a court through the lenses of political representation. We can still ask the question: ‘did X judicial decision incarnate the will of the people?’
I grant this possibility. Yet, I argue that such possibility can only be contingent if allegiance to the law is not to be lost as the primary function of the court. Put differently, the content of a decision may coincide with what the public prefers or with its interests. But it may also not, because the legal system applied by the court may not provide the tools for it to issue decisions that will be aligned with the people’s will or with the people’s interests. And this, which I here refer to as the functional orientation of a court towards the law, implies that an evaluation of the court as negative or wrong because ‘it failed to fulfil a representative role it is expected to fulfil’ will miss the target.
Moreover, should one use the language of representation to account for what courts do because they convey one thing for another, one should simply assume that representation is anything that the court decides. Inasmuch as courts translate the meaning of a text or of a practice into decisions, then whatever a constitutional court decides will be imputed to the people.
But this is wrong for two reasons. One will be discussed below in a brief reference to originalism or legislative intent. Here I will dispel the second one. It is this: describing what courts do and should do as representation in the sense described abuses the term. It says too much and thus too little. Such line of reasoning would easily lead us to conclude that any public institution, inasmuch as its powers can be traced back to an original constitutional moment and inasmuch as every act of application of the law involves the conveying one thing for another, is a representative one. But certainly not any institution exists for the sake of expressing the people’s will, nor their actions are to be evaluated on the bases of whether they represented someone. Applying the language of representation to describe what any institution does by virtue of its partaking in a representative system waters down the term conceptually and normatively; it overshadows the boundaries separating representation from other concomitant terms. And, as I show below using the example of Central Banks, it makes us lose sight of the evaluative tools we have to assess the performance of institutions designed to pursue objectives that could run counter to representative ideals, courts among them.
So courts are functionally oriented towards the law, and representative institutions are functionally oriented towards the will of their constituents. Of course, different theories will interpret this feature differently. Most prominently, there will be differences as to whether the functional orientation is either towards the will (in stricto sensu), the preferences or the interests of the represented. In particular, interest-based theories of representation may consider that the will of the individuals bound by the decisions of the representative will count as secondary; best interests and preferences do not necessarily, and often do not, overlap. Yet my point still stands, namely that the relationship between representative and represented is one where the latter’s preferences and wills of interest count as a gravitational point for the actions of the former; as a normative criterion of assessment of the actions of a representative. This normative relationship is present in the entire spectrum of theories concerned with representation in a democratic law-making context, from delegate to virtual representation, and thus is present in representative institutions. But, as I will argue, although it may be contingently present in courts, it is not something mandatory for them or something that follows from their functions.
This functional orientation is present in the traditional delegate/trustee dichotomy for which the central question is who has the right to decide how to cast votes, whether the constituency or the representative. The question is explicitly formulated in terms of these two alternatives, and it emerges because it is either the nearness or the distance of the represented that is the criterion by which the delegate and trustee extremes are defined. The link between representative and represented shall, of course, be tighter in the case of delegate conceptions of representation, and looser when representatives are given independence in judging what the best course of action is in a given case.Footnote 46 But the distinction and the different positions one might find in between those two extremes are always framed with regard to the will of constituents.
This functional trait does not disappear when one adds more layers of complexity to the delegate/trustee distinction, a distinction that, like Rehfeld shows, obscures additional categories and questions in theories of representation. He argues that the distinction clouds three underlying distinctions regarding aims (whether the representative-lawmakers aim at the good of all or the good of a part), sources of judgement (whether the representative-lawmakers rely on their own judgement or on the judgement of a third party to determine the good at which they aim), and responsiveness (the degree to which the representative-lawmakers are responsive to sanctions).Footnote 47 But it can be readily seen that this more complex story is also told in terms of the functional orientation I describe.
This even holds for so-called theories of ‘virtual representation’ where representatives have no electoral relationship with their constituents, but where nevertheless the relationship between their practice and the constituency to which decisions are imputed, remains. This includes, for example, Mansbridge’s ‘gyroscopic representation,’ where representatives are self-reliant and relatively unresponsive to sanction,Footnote 48 because, again, the question of what makes an action whose effects are imputed to others representative or not, is one about the relationship between the decision-maker and the will of those individuals affected by the decision and the degree to which the latter must be responsive to the former. Every conception of representation, irrespective of the degree of independence in judgement granted to the representatives, orients them towards their constituent in some way. Borrowing Brito and Runciman’s wording regarding trusteeship, “trustees may be thought of as representatives if there is some way of making the interests of those whom they represent present in their actions, independently of whatever the trustees themselves happen to say or do.”Footnote 49 Although there are sundry ways of establishing that presence, it is always necessary to have the represented in sight, the actual individuals who will be affected by the representative’s actions. Should the transfer of competences from represented to representatives be total—what Brito and Runciman refer to as ‘literal trusteeship’—then the relationship is no longer one of representation but of paternalism.Footnote 50 In short, independent judgement is not always representative independent judgement.Footnote 51
Theories of political representation thus share that the will of constituents is their gravitational point, the object they are supposed to interpret, and that divergences about the meaning of representation are relative to this functional trait. And this is irrespective of the degree to which that will binds the representative. Consequently, what Rehfeld thinks for democracy also goes for representation, namely that “we must always justify and explain cases in which law deviates from citizens preferences, whereas no such prima facie justification is required in cases when law conforms to the preferences and wills of those it governs.”Footnote 52 Those preferences and reactions become, in the end, vantage points from where the activity of a decision-maker is either conceptualised or evaluated as representative.
These reflections take us back to the rather simple—though not trivial—etymological point that representation involves making someone’s will present again.Footnote 53 It is useful then to close this subsection by remembering that every ideal view of representation already identified by Pitkin expresses this functional orientation: in delegation, representatives act as the represented; in trusteeship, representatives act for the represented; in descriptive representation, representatives stand as the represented; in symbolic representation, representatives stand for the represented.Footnote 54 But, are the represented the gravitational point of judicial decisions? Are courts functionally oriented in this way?
4. Is the Judiciary Representative?
I have hitherto described three versions of the JRT and shown that the functional orientation towards determining what is the will of the represented pervades different conceptions of political representation.
It is now time to answer whether the JRT is about that, about representation. I mentioned at the outset that it is not clear to me that the JRT uses the term ‘representation’ in a way that captures the functional orientation of courts towards the law, and not towards the will of constituents. With the idea that representative institutions are functionally oriented towards the will of their constituents in the background, I can now show why. The reasons are two. First, what the JRT-endorsers refer to when they use the term representation is best described by using an array of other categories, with similar or concomitant, yet distinct semantic domains such as legitimacy, acceptance, conformity, rationality, justification, argumentation, interpretation, inter alia. As a result, judicial practices are not best accounted for by describing them as representative practices (4.1).
Second, the JRT is an expression of the traditional wariness that majoritarian institutions lack the incentives and the institutional structures to give the constitution the importance it has. This concern leads the scholars discussed in section 2 to look at the courtroom as the proper representation of something other than the short-termist, transient wills and preferences of the citizen of here and now. This concern is understandable, but the solution relies on a fallacious contrast between non-judicial and judicial institutions; between the worst features of the former and an ideal version of the latter. A fairer balance should acknowledge that although legislators often fail at representing what Rosanvallon calls the ‘people as principle,’ it does not follow that this role is performed by courts. The tendency is, however, to unwarrantedly downplay the value and capacity of parliaments and to extol the role of courts (4.2).
4.1. Calling Things by Their Name
I have argued that political representation involves a functional orientation towards the constituents’ will. Whether courts have such an orientation determines whether they are representative institutions.
The JRT says that there is a relationship between judicial decisions, the Constitution, and, borrowing Rosanvallon’s language, the will of the people as principle. It then concludes that courts can be representative institutions. But this move is not warranted, and it leads scholars to use the term representation to describe things that occur within the courtroom for which we already have other terms.
Let me begin by stating a non-controversial fact pertaining to the courts’ functional orientation: constitutions are legal and political documents. Qua political instruments, they embody principles whose instantiation is open for political debate among the different branches of government and among citizens. Any institution interpreting and applying a constitution will inevitably be judging matters with political content. Courts too.
But, of course, constitutions are also law of the land. Many of their provisions are judicially enforceable, and this power is given to courts on the grounds of the distinctive ways in which they frame and decide cases. Should this function be of the kind exerted by the other branches of government, then its existence would be either superfluous or redundant. If the judiciary is to be considered a separate, distinct body, the mechanisms and the language by which judges decide must be unique in their content and effects.Footnote 55 This is shown in cases with clear political content where judges strive to describe their function as part of the realm of the legal. Consider, by way of example, two recent decisions by the UK Supreme Court, informally known as Miller 1 and Miller 2,Footnote 56 where the Court explicitly recognised the political character of the issue at hand while also being explicit about the strictly legal task with which the Court is entrusted. Referring to UK government ministers’ announcement of their willingness to terminate the UK membership of the European Union, the Court said in Miller 1 that the case,
has nothing to do with issues such as the wisdom of the decision … the terms of withdrawal … or the details of any future relationship with the European Union. Those are all political issues which are matters for ministers and Parliament to resolve. They are not issues which are appropriate for resolution by judges, whose duty is to decide issues of law which are brought before them by individuals and entities exercising their rights of access to the courts in a democratic society.Footnote 57
In Miller 2, the court insisted on this point from the very outset:
It is important to emphasise that the issue in these appeals is not when and on what terms the United Kingdom is to leave the European Union. The issue is whether the advice given by the Prime Minister to Her Majesty the Queen on 27th or 28th August 2019 that Parliament should be prorogued … was lawful.Footnote 58
This bears on whether representation is what we should expect from courts. If, as I have explained above, representation in democratic law-making contexts involves making someone else’s will present again, and if judicial procedures are triggered by how facts relate to legal sources and arguments and not to anyone’s particular will, then judicial practice should be accounted for by using other terms, for example, adjudication, argumentation, justification, interpretation, inter alia. Every normative discussion about whether representatives should aim at the good of their constituents and be more or less responsive to their sanctions,Footnote 59 does not apply to the actions of courts. To the extent that courts can justify their decisions and describe them as the result of legal arguments, deviation from the will of the people says little about the validity of those actions.
Mind you, this is not to say that some of the standards of the kind judges apply do not in fact emanate from institutions which are clearly representative. The point is a different one: judges use those standards because they stem from legal sources, and it happens to be the case that in many contemporary democracies those sources are in fact representative. But the coincidence is contingent. Borrowing a Hartian distinction, judicial decisions may be described as representative seen from the perspective of an observer. For example, it is common among some authors to claim that despite their countermajoritarian character courts have historically remained close to public opinion. The argument goes that courts cannot systematically remain out of tune with the people’s preferences about constitutional matters. When they have deviated from the preferences of the majority of the public, they have not done this for long.Footnote 60 An observer could then say that this convergence between the people’s and the judges’ preferences shows that the latter take the former as normatively binding.
Yet even if judicial opinions were statistically aligned with the preferences of the public,Footnote 61 it would not be because the nature of their function mandates them to express the views of the public on what the law means. Rather, the practice is better accounted for by saying that judges interpret and apply the law, and that the stability of their interpretations is as a matter of fact dependent on their long-term acceptance by a large part of the public. But this does not mean that the function of the courts is in the first place to express the people’s will and that they are accountable or responsive to it; that this was an expression of judicial duty, derived from the nature of the institution. It is not. What these arguments show is that inasmuch as courts are part of a democratic joint project of governing,Footnote 62 they cannot perform their functions in a political vacuum. Yet this is a far cry from the contention that their practice is best accounted for as a representative one. Coming back to our borrowed Hartian distinction, such analysis fails to adopt an internal point of view.Footnote 63 The statistical correlation between popular will and judicial preferences does not imply that popular will is a source of law. Rather, we better describe those sorts of statistical coincidences between the popular will and the content of judicial decisions as contingent. And we may describe them by saying that in those cases judges represent the people as a contingent by-product of their job of applying the law. Yet things could have been otherwise. That is, judges could have decided in ways that clearly run counter to the popular will. And in those cases, it would be misleading to criticise those judicial decisions by claiming that judges failed at performing representative roles.
But these distinctions are collapsed by the JRT. Alexy, for example, explicitly acknowledges that the type of representation he attributes to judges is purely argumentative. But I find it difficult to understand why anything would be lost in Alexy’s argument if, instead of accounting for his thesis as one about representation one would just call it argumentation, or justification, or reasoning, or adjudication, or any other term that points at the distinctive ways in which judges decide cases.
I want to consider one argument by Matthias Kumm that could rescue the JRT. Kumm agrees with Alexy that “there is no place in a democracy for an institution that can claim only argumentative representation.”Footnote 64 He grapples with this problem by falling back on the procedures by which constitutional court justices are usually appointed, which involve the political branches of government. This, he thinks, “is at least a weak form of volitional/decisional representation of the people by courts that complements the argumentative dimension.”Footnote 65 Provided that this warrants the representativeness of constitutional courts, Kumm then claims that now “the point would be to require courts to be as representative as possible, ensuring the stronger possible volitional link without undermining the court’s specific function relating to argumentative representation.”Footnote 66
This argument has three problems. First, it ignores that although appointment procedures include the political branches of government, they are designed to prevent ideological bias, not to reflect the judges’ partisan affiliation.Footnote 67 Second, that to the extent that those procedures result in partisan courts,Footnote 68 this would be a deviation from the ideal of isolating the judiciary from party politics,Footnote 69 not a manifestation of it. Finally, Kumm’s argument explains too much. Following his reasoning, we could give the title ‘representative’ to agents and institutions with clearly no representative function at all. Take the case of the selection and appointment processes of the Governor of a Central Bank as an example.Footnote 70 This functionary is formally appointed by the Government after designation by the Head of State. Once appointed, the Minister of Economy and Finance must inform Parliament of such nomination, following a procedure created by Congress. Does this clearly political process of appointment give the Governor of the Bank and indirectly the Bank itself, representative functions? How can one square this with the autonomous and independent nature of this institution? It is certainly counterintuitive to call the Bank a representative institution. Kumm’s reasoning, however, leads to this conclusion.
The problem with Kumm’s argument is that it fails to consider the functional orientation of courts towards legal sources. The example of Central Banks shows that just because an agent has competence to issue decisions with erga omnes effects it does not necessarily mean that it is mandated to make anyone’s will present again. Their functions are certainly of a different nature (e.g., issuing tender money, supervising credit institutions, etc.), which although changing over time, have always taken place in the context of relationships between the institution and government and between the bank and the financial system, not between the bank and the will of the people.Footnote 71 Even those criticising the increase in independence gained by Central Banks in recent decades from representative institutions express worries about lack of accountability, not lack of representativeness.Footnote 72 One could of course insist that the law is the instrument that governs these relationships, and the law has been democratically enacted by representatives, but this would lead us into the absurd conclusion that anything that is regulated by law could be an instance of representation. As Pitkin avowed regarding formalistic views of representation: “[a] judge may issue a judicial injunction binding on [a] union; yet we do not say that he is the union’s representative or represents it when he does so.”Footnote 73
Something similar goes for constitutional courts, whose orientation is foremostly towards interpreting and applying the constitution even if they lead to results which have nothing to do with the wills of those who are affected by their decisions. This orientation towards functions different from making anyone’s will present again makes it the case that although constitutional courts justices, like Central Banks’ governors, are indirectly appointed by elected politicians, it is a different thing to say that they are representative institutions. This is not to say that Central Banks or courts are not accountable. They are.Footnote 74 Accountability and legitimacy certainly constrain the independence of both Central Banks and courts. As Lastra avows, “[a] democratic society always retains the discretion to reinvent itself; and should the people of a country make a considered, deliberate choice to seek some other goal, the institutions of central bank independence should respect that decision.”Footnote 75 But should we call this representation? Why not legitimacy and/or accountability?
Pettit faces a similar problem. ‘Indicative representation’ of the kind that judges would allegedly incarnate, would require them to read a constitution and provide its provisions with meanings, with which constituents would then be in a position to feel identified. But I wonder then, how ‘indicative representation’ would be different from the notion of interpretation. I struggle to see the gains in calling this practice of meaning ascription ‘representation.’Footnote 76 Judges are not mandated to make anyone’s will present again; they interpret and apply the law, and the law is an instrument that by definition applies to everyone irrespective of anyone’s particular desires, wills, opinions, or preferences. It applies even if its addressees ignore its existence. Of course, this story is one step closer to a noble dream than to a nightmare,Footnote 77 but it holds true even if one considers that judges create law every time they decide. Law creation and change can happen through several avenues,Footnote 78 many of which will not be representative, if only because many of them—for example, those enacted in parliament—cannot be attributed to anyone’s particular intention.Footnote 79
It is by no means obvious that those goals courts point to are best described as objects of representation and not as objects of interpretation. Interpreters do what Malone in Samuel Becket’s Malone Dies affirms in his reveries: “I have to look long and fixedly and give things time to travel the long road that lies between me and them.”Footnote 80 Judges ought to look long and fixedly at the law, not at someone’s will. As a US Supreme Court Justice aptly put it, “[t]he Court’s power lies … in its legitimacy, a product of substance and perception that shows itself in the people’s acceptance of the Judiciary as fit to determine what the nation’s law means and to declare what it demands.”Footnote 81 It may well be the case that what the law demands is different from what the public would prefer. And in such cases, the analytical tools we may use to criticise such decisions will not come from theories of representation.
4.2. Fallacies and Neglects
Of course, the concerns behind the JRT are understandable. In describing the judiciary as representative of the people, the thesis paves avenues for coping with the deficiencies of democratic, majoritarian, collective, and election-based decision-making procedures. The argumentative force backing judicial decisions and the stringent criteria judges must meet when justifying their decisions, especially at the constitutional level, supplement majoritarian volitional procedures. Absent these institutional safeguards, constitutions would fail at fulfilling their function of constituting a distinctive political identity, different from the particular wills and preferences of their framers, elected politicians, and citizens.Footnote 82 In such a scenario, citizens are abandoned to arbitrariness.Footnote 83 Individuals at the receiving end of judicial decisions are given actual reasons that they may or may not share, but that they can understand and accept because the reflexive process leading to their enactment was rational, fair, impartial, etc. But the move from this concern to the JRT incurs two fallacies: moralistic and nirvana.
4.2.1. Moralistic Fallacy
A moralistic fallacy is a reasoning stating that the world is as one thinks it should be. In the context of my discussion, it goes as follows: we need representation at the constitutional level, but political institutions fail at achieving this purpose. Therefore, the judiciary is a representative institution.
Pettit, for example, shifts his focus from representation to democracy when he tells us that “[d]emocracy is too important to be left to elected politicians alone.”Footnote 84 One can only agree with this. But it is a very different thing to say that, because representation is too important, courts (can) do it. Rosanvallon’s narrative is, in this sense, problematic as well, for it assumes that because legislatures are concerned with keeping the electoral people happy, the people as principle escapes them. If, the argument then says, we want to secure that the people as principle are considered, we should then look at courts. But the need for finding spaces for constitutional politics that are not driven by whims and short-termism, does not in itself ground that courts are in fact the institutions where the constitutional people are represented.
4.2.2. Nirvana Fallacy
The nirvana fallacy, in turn, is manifested in the JRT in that it involves an unfair or inadequate institutional comparison. A nirvana fallacy is one by which real things are compared or contrasted with ideal alternatives, and by virtue of those discrepancies it is deduced that the real is inefficient.Footnote 85 In terms of courts and legislatures, the JRT assumes that because the ideal features of constitutional representation are seldom found in real legislatures, one should look at courts and find those features there. That initially is a reasonable move to make, if it wasn’t that the features in the JRT are ideal, not real.
I am not the first to argue that there is a tendency in legal theory to portray the courtroom in such ideal colours and contrast it with a gloomy picture of legislative assemblies.Footnote 86 But this comparison takes a unique shape when it comes to representation. Alexy’s confessed necessity to complement ‘argumentative representation’ with majoritarianism can only arise from the assumption that the institutional features of the judiciary prevent it from deciding on purely majoritarian grounds. It fails, however, to consider judicial decisions that have been either poorly argued for or not argued for at all.Footnote 87 Imagine that the decision is never contested by the public, neither in the formal nor the informal public sphere, and that, in fact, the outcome (not the reasoning) is shared by the majority of the population. Should we call this decision a ‘representative’ one? Imagine the same thing about a legislature. Would we call this a completely non-representative decision?
These alternatives suggest that Alexy’s ‘argumentative representation’ works under idealised conditions where the judges’ opinions about a certain legal question largely coincide with a public that is willing to take their arguments as its own, combined with an image where legislators simply voted without properly debating what they were legislating about and for. The public to which ‘argumentative representation’ appeals, Alexy’s ‘constitutional persons’ and courts, are highly idealised, but legislators are not portrayed in such charitable terms.
Rosanvallon falls into the same problems by stating that legislatures represent the electoral people, assuming away the possibility that they may also be reflexive institutions. As with ‘argumentative representation,’ this institutional comparison neglects the instances in which legislatures have in fact reflected upon constitutional principles. There is nothing in the institutional setting of a legislature impeding it from engaging in dialogue about the constitutional implications of its decisions.
Rosanvallon (and Kumm for that matter) too has to grapple with the problem that constitutional judges use sundry interpretive methods, grounded on different principles. Such plurality of approaches does everything but guarantee that courts will always have Rosanvallon’s representative goals in mind. Textualist judges will be happy to say that they do not care about the intention of a constituent or of the general public, or of the legislature when it enacted a certain law, but only what the words of the constitution mean. Should the general will run in a direction different from the one allegedly inbuilt in the text, then a textualist should decide against that will. Textualist judges have no authority to pursue any other broader purposes.Footnote 88
In turn, constitutional judges following Ely’s structuralist approach might choose to read the constitution as a document primarily concerned with clearing the channels of political change and with safeguarding the rights of discrete and insular minorities.Footnote 89 It should be admitted that this approach is prima facie not hostile to my account.Footnote 90 Yet, it is by no means obvious that, first, this approach is the only one legitimately available for them to follow and, second, that as critics of Ely have shown, structuralism needs to be grounded in some features of the constitution that is being interpreted. And it could be the case that some constitutions do not provide the materials for a judge to interpret in the way Ely thought the US constitution should be read. Or it also be the case that a constitution does provide the materials for such an approach, but that other approaches, depending on sundry factors, are better than a structuralist or proceduralist one.Footnote 91
Moreover, judges committed to Dworkinean integrity might interpret and apply the constitution thinking that it instantiates moral principles about political decency and justice, which do not necessarily have to be linked to or emerge from the general will. Put differently, the best possible judicial Dworkinian interpretation of the scheme of moral and political expectations built up within the law does not have to be a representative one; it could derive from a commitment to moral principles that has little or nothing to do with the preferences of anyone.Footnote 92
Perhaps endorsers of the JRT could find refuge in originalism, namely the view that judges are bound by and should therefore express the original intentions of the framers of a constitution? This could count as a version of the will of the constituents. Perhaps, but as with Ely, this would be one among many possible methods of interpretation available to them and it is an ideal that admits sundry versions, some of which give pride of place to the framers’ intentions, some preferring more textualist versions of the original meaning, among others.Footnote 93 Hence, there are no reasons for buying into the idea that the reflexivity of constitutional courts is necessarily oriented towards realising anyone’s will.
Similar reflections apply to accounts that take legislative intent as a paradigm of constitutional interpretation. In Ekins’ paradigmatic account, for example, it is clear that there is a difference between attributing to the judiciary a representative function and the judges’ duty to decide on the basis of legislative intent. The difference lies in something akin to my idea of functional orientation. That is, in Ekins’ view, that the legislature “is an institution that aims to act responsibly for the common good.”Footnote 94 This, which Ekins presents as an axiom of interpretive practices, “responds to the rationale for legislative authority,”Footnote 95 which in turn provides reason for interpreters to pursue the determination of legislative intent.
This appeal to the rationale of the legislature indicates that the centrality of legislative intent as an interpretive method is predicated upon the notion that the legislature grounds its powers in reasons specific to it.
One could affirm that the authority of the legislature to represent its constituency derives from these functional features. But these features are, as mentioned, specific to the legislature and though they place interpreters in a position to apply the law in a way that tracks the former’s intention, it does not follow that such alleged duty derives from the same functional rationale (alleged because in this paragraph I am accepting the prominence of legislative intent as an interpretive method for the sake of the argument). To follow the law because it emanates from a representative institution is not tantamount to say that the application is an act of representation.
In conclusion, these different methods of interpretation and the differences in how different scholars understand them show that courts may or may not be concerned with determining what the will of the people as principle is. If that is the case, diversions from the preferences of the public are not better accounted for as failures in the way they exercise the practice. They cannot be attacked for not fulfilling representative functions. On the contrary, representative institutions failing to consider the will of their constituents do have to provide reasons for this digression.
5. Conclusion
My argument has been a negative one, namely that what judges do is not best accounted for as an instance of representation. Judges perform several important tasks, but they do not represent.
In a different context, Harel has recently made what I think is the right move. He claims that the court “is owed deference [by the legislature] precisely because it is a non-representative body.”Footnote 96 I think the degree to which he trusts on courts is not fully warranted,Footnote 97 but for sure he is hitting the nail on its head by calling things by their name. Questions about representation will be better addressed by looking at institutions and agents whose functional orientation is towards make someone’s will present again. We would do well in looking at representative institutions if representation is what we value.
