Sitting at the Same Table: a cross-disciplinary ‘constitutional-institutionalist’ approach to the study of constitutions

Abstract

This article presents a cross-disciplinary approach to the study of constitutions: ‘constitutional institutionalism’. Conventional approaches in law, philosophy or political science tend to reduce constitutions either to their formal, factual or ideal aspects. The constitutional-institutionalist approach, by contrast, seeks to integrate these aspects into a more general perspective by focusing on the dynamic interplay between constitutional actors and constitutional norms. It understands constitutional norms as binding institutions that shape and constrain political action, but never fully determine it. Constitutional institutionalism furthermore asserts that constitutional norms, whatever form they take, only have meaning in relation to other constitutional norms as well as to constitutional actors, who impose meaning on these norms. Therefore, constitutional phenomena ultimately require interpretive explanations. This article concludes with a brief constitutional-institutionalist research agenda.

1 Introduction: ‘sitting at different tables’

Constitutions are massively important: they establish and regulate public authorities, the institutional relations among public authorities and the relationship between public authorities and the rest of society, including individuals and organisations. But how should we understand constitutions? What is the most meaningful way to account for their import and the way they work?

Scholarly literature roughly provides three answers to these crucial questions. (1) In studying constitutions, legal scholars tend to focus on what has been called the ‘large-c constitution’: the set of legal precepts that derive their authority from having been formally enacted by a constitutional assembly of some sort, a constitutional legislator, a court or by way of some other legally authorised procedure (Law 2010, at 377). (2) Political scientists, by contrast, are typically interested in the ‘small-c constitution’: the body of rules, customs and understandings that actually determine who holds political power, under what conditions, and subject to what (substantive) limits (Law 2010, at 377). Small-c constitutions can (partly) be embodied by formal norms, but they may also take the form of unwritten conventions that lack formal legal status. (3) Legal and political philosophers take yet another angle on constitutions. They focus on what we might call the ‘ideal constitution’: the set of ideas that suggest who should or should not have political power and to what end such power should be exercised. These ideas may or may not be embodied by the large-c or small-c constitution and can be used to justify or criticise their content.¹

Each of these approaches may be helpful, as they may allow us to shed light on important constitutional aspects. On the other hand, all three approaches tend to reduce constitutions to either their legal, factual or ideal aspects, while constitutions as we encounter them ‘in real life’ are comprised of all three of these aspects at the same time (Cf. Hirschl 2016, at 7). Legal scholars, political scientists and philosophers each therefore run the risk of providing only a partial understanding and incomplete account of constitutions. Moreover, as each of these disciplines carries with them their own assumptions as to what a constitution exactly is and how we can gain knowledge about it, mono-disciplinary findings may well be at odds and indeed contradict with each other. As a result, legal scholars, political scientists and philosophers often find themselves ‘sitting at different tables’, having separate conversations about the same or at least closely related constitutional phenomena (Cf. Almond 1988).

What kind of approach would allow us to recapture the complexity of constitutions and provide a more complete and integral account of their meaning and operation? How can we make the prospect of a unified – or indeed unifying² – field of ‘constitutional studies’ real? (Hirschl 2013). In this article, we merge some existing and new ideas, transform them into an integral perspective on constitutions and suggest how this perspective can be operationalised for conducting actual research.

Our central thesis is that a deeper and more comprehensive understanding of constitutions can be gained by connecting the legal, political and philosophical perspectives and focusing on constitutional change over time. This ‘constitutional-institutionalist’ approach, as we call it, seeks to integrate the formal, factual and ideal aspects of constitutions into a broader perspective and thus to lay the foundations for research that can study their interrelationship. Constitutional institutionalism focuses on the dynamic interplay between the beliefs and actions of constitutional actors and constitutional norms. It understands constitutional norms as regulative institutions that legitimate, shape and constrain political action, but never fully determine it. Constitutional institutionalism furthermore asserts that constitutional norms, whatever form they take, only have meaning in relation to other constitutional norms as well as to constitutional actors, who impose meaning on these norms. Therefore, constitutional phenomena ultimately require interpretive explanations.

The ideas behind constitutional institutionalism are not entirely new. MacCormick and Weinberger, for example, have developed an ‘institutional’ theory of law that explicitly seeks to avoid ‘idealistic’ or ‘realist’ reductions (MacCormick and Weinberger 1986, Weinberger 1991, MacCormick 2007). However, their theory is aimed at explaining law in general, while constitutional institutionalism is explicitly centred on the constitutional domain. Moreover, general theories of law such as those of MacCormick and Weinberger tend to remain rather abstract and fail to provide concrete instructions about how to actually conduct research. The same goes for calls for more cooperation between legal scholars and political scientists, and for cross-disciplinary and inter-disciplinary research more generally. Even when these pleas specifically address constitutional scholarship, operationalisation is often lacking.³ On the other hand, scholarship conducted in line with constitutional institutionalism typically does not aspire to explicate a general, reproducible research approach.⁴ As a result, general assumptions underlying the research tend to remain implicit. Thus, although the contours of a constitutional-institutionalist approach can be found in some strands of legal and constitutional scholarship, this approach is in need of explication. Constitutional institutionalism, as presented in this article, provides its own view on the constitutional domain (an ontology) and offers its own roadmap (an epistemology) for accounting for the key features that make up constitutional phenomena. It thereby explicitly lays the groundwork for integrated interdisciplinary research on constitutions.⁵

Section 2 of this article discusses the idea of a research approach in more detail and argues that scholars studying constitutions should pay serious attention to the ontological and epistemological claims and assumptions that underpin their work. Section 3 describes and evaluates three conventional approaches to the study of constitutions. While all three approaches offer a solid basis for meaningful research, we contend that in their extremes, all three approaches lead scholars to miss, or misperceive, vital aspects of constitutions and their interconnectedness.

Against this background, sections 4 and 5 introduce constitutional institutionalism as a research approach that rejects the extreme positions that reduce constitutions either to their formal, factual or ideal aspects and integrates these aspects into a more general perspective. Section 4 describes the central assumptions that constitutional institutionalism makes about the nature of constitutions, while section 5 deals with constitutional institutionalism's assumptions on how we can learn about them. Section 6 concludes with a brief research agenda: a number of topics for which constitutional institutionalism as a research approach looks particularly promising.

2 Research approaches: what they are and why they matter

The legal, political and philosophical perspectives on constitutions discussed in the introduction can be regarded as characteristic of distinct research approaches in constitutional scholarship. As such, they entail much more than just a particular focus on one of the various aspects of constitutions. By opening up the field of research in a particular way, these perspectives determine in advance, on an abstract level, how scholars perceive the nature and scope of their object of study and how they can, and should, proceed to acquire knowledge about it (Cf. Marsh and Stoker 2002, at 311–313). A research approach to constitutions provides an initial overview of the constitutional domain as well as a set of broad, abstract assumptions about the ways in which it is possible to get to know this domain. In technical terms, it presents both an ‘ontology’ and an ‘epistemology’ (Marsh and Furlong 2002, at 17).

Ontology refers to what is. Ontological claims or assumptions relate to something that exists, what it looks like, what components make it up, and how these components relate to each other and their environment (Hay 2011, at 462). Each research approach to constitutions incorporates – implicitly or explicitly – its own ‘constitutional ontology’: a set of claims or assumptions about the reality of constitutions; what a constitution really is (Green 2017, at 4–7).⁶ Different approaches to constitutions portray and delineate this reality in different ways. It makes much difference, for instance, whether scholars regard legal norms, actual practices and/or normative ideas as the basic components of the constitutional domain. And even if distinct approaches acknowledge the existence and relevance of the other components mentioned, they might give them a different place, position and significance in their overall scheme of constitutional reality. For example, one may attribute to legal norms an existence independent of their acceptance by actors to whom they are directed, or instead understand them as being merely a reflection of actual practices. Similarly, one may regard legal norms in complete isolation from the ideas they purport to express, or instead see them as being only meaningful in light of these ideas.

Next to a constitutional ontology, a research approach to constitutions also includes a ‘constitutional epistemology’ (Cf. Green 2017, at 4–7). A constitutional epistemology is a set of claims or assumptions about how we can gain knowledge on constitutional reality (Hay 2011, at 464). Different approaches to constitutions give different answers to the question as to how to account for constitutional phenomena. If one regards a constitution as the body of rules, customs and understandings that de facto determine the exercise of public authority, for example, the most appropriate way to establish the content of a constitution seems to be empirical observation. By contrast, if one views the constitution primarily as a legal artefact which derives its authority from its ‘writtenness’ (Cf. Albert 2019, at 137) and the way it has been adopted, one will likely favour methods of textual and originalist interpretation (Cf. Elkins and Ginsburg 2021, at 321–343). Different variants of the latter approach are themselves also based on different epistemological positions. One may claim, for example, that it is possible to know the original intent of the authors of a constitution, even if the constitution was written in times long past and its drafters disagreed about many things, while others may deny this possibility and argue that one can only – and therefore should instead – explain the meaning of a constitutional document in light of its present-day context.

Since ontological assumptions have epistemological implications, disagreement over the question as to how to recognise constitutions is often an epiphenomenon of a more ultimate disagreement about the ontological question of what a constitution is (Cf. Hay 2011, at 461).⁷ Richard Fallon, for instance, rejects the argument in defence of text-based approaches that ‘everyone knows that the written Constitution is the supreme law of the United States, binding on courts as well as other officials’ by arguing that this argument ‘begs the central questions in issue’, most prominently the question ‘what “the Constitution” is’ (Fallon 1998, at 545–546). In a similar vein, Sanford Levinson states that ‘[o]ne must first decide … what “the Constitution” is even prior to deciding what “it” means’ (Levinson 2011, at 29). At the same time, both authors do not really tackle the what-is-the-Constitution question (Cf. Green 2017, at 15–16). According to Levinson, ‘all suggested answers inevitably are circular’ (Levinson 2011, at 36). This statement reflects a more general insight: an ontology is ‘necessarily irrefutable in its own terms’ (Blyth 2002, at 294). As an initial overview of the constitutional domain, constitutional ontologies posit, but do not demonstrate, what the constitutional domain is made up of. As Mark Blyth argues, ‘[o]ne can posit ontologies all day long and there is no a priori way of sorting out which ones are more productive than others, other than producing actual research’. (Blyth 2002, 308). But that does not necessarily mean, of course, that one ontology cannot be more helpful, meaningful or indeed convincing than the other.⁸

By providing an organising perspective on the object of inquiry (constitutions), a research approach lays the ontological and epistemological foundation for constitutional research. Crucially, however, it does not automatically come up with particular research questions, research goals or research methods. As we understand it, an approach merely provides a set of broad and general assumptions about the reality of constitutions and the ways in which it is possible to gain knowledge of them. An approach typically allows for substantial freedom with respect to concrete research designs. Within the parameters set by their ontological and epistemological assumptions, scholars are free to ask any question and use any method that best achieves their purposes.⁹ Research approaches furthermore should be distinguished from theories. On its own, an approach explains nothing. It merely provides the basis for developing theories.¹⁰

A research approach to constitutions gives an answer – on a very general level – to the question what a constitution is and how we can gather insight into it. Because of the general nature of this answer, constitutional scholars often do not discuss their ontological and epistemological claims and assumptions,¹¹ but simply assume their validity or deem them self-evident.

We believe, however, that scholars should not ignore ontological and epistemological issues, for at least four reasons. First, there is the fact that we cannot escape taking ontological and epistemological points of view. No constitutional scholarship is ‘neutral’ in this respect (Hay 2011, at 460). Helpful scholarship makes its reader aware and critical of the presuppositions on which its research rests, by making them explicit and reflecting upon them. Second, the kind of research questions and research goals scholars formulate, and what type of research methods scholars employ, will at least partially depend on the research approach to constitutions they take. If scholars are not clear about their approach, they will miss orientation and their questions, goals and methods will likely lack a clear direction. Third, making approaches explicit may be helpful in showing what constitutionalists are actually doing – or supposed to be doing – when they look into constitutional phenomena and reach their conclusions. This, in turn, may make constitutional scholarship better reproducible. Fourth, ontological and epistemological issues are the most fundamental source of divergence in constitutional scholarship, separating competing academic disciplines, schools and paradigms from each other.¹² For the sake of fruitful communication, all constitutionalists should acknowledge their own ontological and epistemological presuppositions and appreciate, even if critically, the views of others (Cf. Marsh and Stoker 2002, at 313).

3 Three conventional approaches, their strengths and limitations

In this section, we describe and evaluate the three main approaches to the study of constitutions, which we call the ‘formalist’ approach, the ‘idealist’ approach and the ‘realist’ approach, respectively.¹³ We present these approaches as ‘ideal types’: they outline particular constitutional ontologies and epistemologies in a pure form.¹⁴ In this article we aim to give a general classification of possible research approaches to constitutions mainly for explanatory and analytical reasons. The research of most constitutional scholars will of course not neatly fit within any of these approaches, but rather combine elements of them.

3.1 The formalist approach

Formalism draws on the idea that a constitution is primarily, or indeed entirely (Cf. Kelsen 1945, at 124–125), a set of discrete, autonomous legal norms, which derive their authority from having been formally adopted by some legally authorised actor, like a constitutional assembly, a constitutional legislator or a constitutional court. Such actors are sometimes being referred to as ‘the sovereign’ (Green 2019).

In the formalist approach, a constitution is a closed system in the sense that it can be changed only through a limited number of legally authorised procedures, the most notable being formal constitutional amendment. The limited number of ways to change a constitution contributes to the autonomy of constitutional law and the ‘closure’ of constitutional law from its context that guarantees its unity and coherence. Against a ‘moral’ understanding of the constitution, the formalist approach maintains that the validity of constitutional norms does not depend on the extent to which constitutional norms satisfy ideals such as justice, democracy or the rule of law (Green 2019, at 114).¹⁵ At the same time, the formalist approach detaches constitutional norms from actual social reality, that is to say, from the facts on the ground. The validity of individual constitutional norms is determined only by the constitutional order itself, that is, the whole of all legal-constitutional norms (Kelsen 1945, at 117). Constitutional norms remain valid as long as they have not been invalidated in a way that the constitutional order itself determines (Kelsen 1945, at 117–118).¹⁶ By consequence, the formalist approach has a minimalistic ontology. Its basic units are formal legal norms. Constitutional actors, political practices and normative ideas are only relevant to the extent that the law formally acknowledges their existence.

Even though formalist scholars might recognise that law is the product of socio-political processes, they emphasise that if we want to know what the constitution requires, we should resort to legally designated sources only. In accounting for constitutional norms, then, formalism centres on the identification and interpretation of constitutional precepts. While ‘classical’ formalists regard the master-text constitution as a comprehensive catalogue of valid constitutional norms, formalists that are more flexible may recognise a broader constitutional canon, including judicial decisions, organic laws (Wheare 1966, at 3), ‘quasi-constitutional’ (Albert 2017), ‘super-’ (Eskridge and Ferejohn 2001) or ‘landmark’ (Ackerman 2014, at 11) statutes and international law (Ackerman and Golove 1995). In any event, the interpretation of these sources should be guided by a limited number of interpretative methods; in general, the formalist creed is ‘fidelity to the text’ (Goldsworthy 2012, at 690–691).

3.2 The idealist approach

In the idealist view, a constitution is not so much a set of fundamental legal norms as it is a collection of ideas (Rhodes 2011, at 148–149). Ideas, in this context, refer to particular conceptions of how political power should be structured and used, and to what end. We therefore might also call ideas ‘ideals’ (Rueschemeyer 2006, at 228). Examples of ideas that are somehow part of most Western constitutions are democracy, the rule of law and human rights. Constitutions can also be seen as the embodiment of more elaborate ideologies, like liberalism, socialism or nationalism (Frankenberg 2006, at 440–441). Ideas may be embodied by law, but they do not necessarily coincide with it. Analytically, the large-c, small-c and ideal constitution are distinct.

In the idealist approach, the focus is on ideas; the ideas that make up the ideal constitution and that may or may not be fully reflected in constitutional law or actual practices. The legal norms that give effect to these ideas and the way they operate in practice come, if at all, second. Actual practices play, at most, only a marginal role in the idealist ontology.¹⁷ When idealist scholars interpret a small-c or large-c constitution, they typically aim to make explicit the ideas they believe are enshrined in the constitutional document or that animate the rules, customs and understandings that actually govern the exercise of political power. By restating at a general level the ideas set out in the constitution, they can help to apply the constitution to concrete constitutional controversies.¹⁸

When idealist scholars reconstruct the ideational content or background of an existing constitution at a particular moment in time or across a particular period, their accounts tend to be based on a wide-ranging interpretation of the relevant historical, cultural, legal and political material.¹⁹

Alternatively, they may develop more abstract theories that claim to be valid more or less regardless of time and place, as in the case of John Rawls's theory of justice (Rawls 2003). Such theories are based less on the analysis of existing constitutions than on pre-given ideas like freedom and equality. Most constitutional theories, however, occupy a middle ground between both extremes; they are the product of a combination of normative theory building and interpretative analysis of existing constitutions (Dworkin 1977, at 149). An example of this approach is the work of Ronald Dworkin, who advocated ‘a fusion of constitutional law and moral theory’ (Dworkin 1977, at 149).

3.3 The realist approach

The realist approach relativises the autonomy and importance of constitutional law and ideas. It understands the constitution primarily as an empirical phenomenon: as the patterns, customs and understandings that actually regulate the ways in which public power is constituted and exercised (Law 2010, at 377) Realists do not recognise rules because of their origin (e.g. in formal procedures) or because of their moral substance, but because of their actual operation in the social and political world. As Karl Llewellyn puts it, ‘it is only the practice which can legitimize the words [of legal-constitutional provisions] as being still part of … [a] going Constitution’ (Llewellyn 1934).

Constitutional realists, then, focus not so much on ‘law-in-books’ as on ‘law-in-action’.²⁰ They concentrate on the ‘real’, ‘living’, ‘political’ or small-c constitution – the rules that in fact determine the exercise of public authority. In theory, the small-c constitution, the large-c constitution and the ideal constitution can perfectly overlap. In practice, however, the small-c constitution is likely to depart, at least in some respects, from its large-c counterpart and may even contradict it (Law 2010, at 377). New or changing practices may give rise to conventions that are not (yet) recognised as formal norms, but that effectively guide and constrain constitutional actors (Barber 2010, at 81-85 and 89-103).

Some realist scholars pursue quantitative research and seek to explain the causes and consequences of constitutional arrangements similar to how the natural sciences explain natural phenomena (Law 2010, at 378). They try to establish causal relationships between legal, political and social facts using theories to generate hypotheses that can be tested by direct observation.²¹ Other realist scholars instead employ qualitative research methods (Rhodes 2011, at 145–147). Their general aim can be to understand how constitutional patterns of behaviour shape and constrain the actual ways in which political power is exercised, or how political power in turn shapes constitutional structures.²²

3.4 Evaluation

Dependent on one's research goals, the formalist, idealist as well as realist approach can point to important insights. To a large extent, they do not exclude each other. However, they also incorporate different ontological and epistemological claims and assumptions. By conceptualising constitutions in their own way, they offer three different, partially competing views on the very nature of constitutions. Characteristic of the formalist approach is that it conceives of a constitution as a body of fundamental legal norms that is shaped by its own internal rationality and that can only be described and analysed in terms specific to the law itself. Normative ideas and political practices are not constitutive of the constitution. The idealist and realist approaches, by contrast, recognise moral and behavioural elements as constitutive for what a constitution is. Their definition of the constitution includes elements of a non-legal nature: in the idealist approach, normative ideas (moral and political ideals); in the realist approach, actual practices (social and political facts). The precise extent to which a constitution depends on non-legal elements varies, depending on the scholar one consults.²³

Regarding the question how to define the nature of constitutions, it can be doubted whether this question leaves room for a correct answer – ‘correct’ in the sense that the validity of the answer could be tested ‘empirically’ (Hay 2011, at 464). In the philosophy of science, there is a growing consensus that there are no firm, unchangeable foundations to knowledge (Baert 2005, at 147–150). If we accept that all knowledge is somehow socially ‘constructed’ and that scholars cannot step out of the history and culture in which they are embedded, knowledge in the sense of an adequate representation of the intrinsic nature of an ‘external’ world is impossible. Any inquiry must begin with basic concepts and assumptions that reflect our prejudgments, which we can revise but never escape (Baert 2005, at 151–152). Without an initial conceptualisation of constitutions, which inevitably depends on the perspective of the scholar, there would simply be no order in the complexity of things as they appear to us. The only way forward, therefore, is by adopting a sufficiently advanced research approach, by actually conducting research and by critically reconsidering – at all times – our presuppositions in light of the results.

The formalist, idealist and realist approaches all provide a solid basis for meaningful constitutional research. However, because they have a distinct and specific focus, the three approaches are also inherently limited. These limitations can make good sense, as extremely ambitious research projects are rarely feasible (Cf. Van Hoecke 2015, at 6–7). Nevertheless, these limitations are also problematic insofar as they leave out key variables that make up prominent features of constitutions as we encounter them ‘in real life’ (Hirschl 2016, at 7). If the general aim of constitutional scholarship is to gain more knowledge about constitutions, to understand their character, and to see how they work, an a priori reduction of the object of inquiry to just one of its aspects seems a bad start.²⁴ We should better begin from a more comprehensive perspective.

When scholars start from a narrow definition of constitutions, which highlights only one aspect of them, references to other aspects often creep implicitly into their accounts. This becomes clear if we look, for example, at strict formalist accounts of constitutions, which try to depict constitutions as legal artefacts whose content and meaning are completely closed off from politics and morality (Kelsen 1945). Actual practices and normative ideas tend to enter such accounts ‘through the backdoor’. To start with, it seems impossible to understand the basis of validity of actually existing constitutions without reference to social and political facts (Schauer 1994). Even for the supposedly ‘pure’ formalist Hans Kelsen, the foundation of the constitutional order lies in how people ‘actually behave’ (Kelsen 1945, at 119). The validity of constitutional law ultimately depends on what he calls ‘the principle of effectiveness’ (Kelsen 1945, at 118). In the case of a successful revolution, for example, the old constitution loses its validity not because it has been nullified in a manner anticipated by that constitution, but because it no longer corresponds to socio-political reality (Kelsen 1945, at 117–119). Under the new regime, people behave, by and large, in conformity with the new constitution, and this fact is an essential criterion for its validity.

Similar observations can be made with regard to the intricate ways in which the constitution – as a body of fundamental legal norms – is connected with ideas. If the validity of a constitution presupposes efficacy, this efficacy in turn depends on the beliefs political leaders (officials) hold. A constitution has to be accepted from ‘the internal point of view’, as Herbert Hart calls it (Hart 2012, at 102). For political actors to accept a constitution as worthy of obedience – and thus to regard it not just as a coercive instrument of power – there have to be ideas in light of which they can determine whether the exercise of political power has the quality of law (Cf. MacCormick 2007, at 16–18 and 56). These ideas are normative in nature. In assessing constitutions, different actors may refer to different ideas. In any case, it seems impossible to account for the acceptance of constitutions (and, hence, for their validity) without considering the normative ideas from which people take directions.

Taken in isolation, the formalist, idealist and realist approaches easily overestimate the significance and impact of the particular aspect of constitutions they focus on and downplay or overlook the other aspects – possibly leading to rigid forms of legalism, idealism or realism. Since the ontological and epistemological assumptions they make are all hard to prove (as well as to disprove), it seems wise to instead acknowledge and centre on the dynamic interplay of interaction, adaptation and conflict between the various aspects of constitutions, without fixing a priori their respective significance (Cf. Hay 2011, 474). Legal norms, normative ideas and actual practices are different phenomena that are somehow related but nevertheless are fairly recognisable and distinct. What seems most promising for a general research approach to constitutions is a perspective that rejects the extreme positions that reduce constitutions to either their legal, ideal or factual aspects and that instead covers all three aspects without establishing in advance their precise relationship and significance (Kumm 2009, at 401 and 408).

Offering such a perspective is the ambition of constitutional institutionalism, which we introduce in the next two sections. Our claim is not that constitutional institutionalism perfectly ‘mirrors’ the objective reality of the constitutional domain. We merely propose that constitutional institutionalism is a helpful and meaningful perspective that can pave the way for research into the various aspects of constitutions and the intricate ways these aspects relate to each other – aspects that until now have far too often been studied in relative isolation.

4 The ontology of constitutional institutionalism

How can we gain a deeper and more comprehensive understanding of constitutions? The answer we mean to present rejects the extreme positions that reduce constitutional phenomena either to their formal, ideal or factual aspects and seeks to integrate these aspects into an overall account. It provides a way of integrating the formalist, realist and idealist approaches, so that we can understand how different constitutional aspects relate to one another across time.

We propose to call this the ‘constitutional-institutionalist approach’. This approach draws upon historical institutionalism in political science.²⁵ Historical institutionalism centres on the development of different kinds of ‘institutions’, which it defines as patterns of thinking and doing that are, in the words of James March and Johan Olson, ‘relatively invariant in the face of turnover of individuals and relatively resilient to the idiosyncratic preferences and expectations of individuals and changing external circumstances’ (March and Olsen 2008, at 3). Institutions are, in this view, persistent social rules that shape, limit, and channel human behaviour and thought over time.

The core proposition of historical institutionalism is that institutions are the variables that explain action in the most direct manner (Peters 2005, at 164). At the same time, it claims that institutions depend on the behaviour and thoughts of actors. Actors are constrained by institutions, yet they also create, shape and reshape them. In other words, actors are both ‘rule takers’ and ‘rule makers’. Actors enact new institutions, but they take existing institutions as the starting point for defining their own identities and interests. Conversely, actors are constrained by existing institutions, but may modify or even overturn these institutions in the course of time. In understanding institutional continuity and change, historical institutionalism stresses that institutions are the legacy of concrete historical processes (Conran and Thelen 2016, at 60–61).

The school of historical institutionalism emphasises, moreover, that it is more enlightening to study political interactions ‘as life is lived’, rather than at only one point in time and in isolation of their wider institutional context (Sanders 2008, at 39). The ‘study of development’, as historical institutionalism is also called, does not merely offer an expose of individual events, conflicts or policies (Fukuyama 2014, at 6–7). It centres around the multifarious processes by which institutions emerge, evolve, and eventually decay. Historical institutionalism emphasises that institutions can and do evolve over time and that history does not necessarily develop in a straightforward and linear fashion. By consequence, historical institutionalism rejects accounts that explain behaviour merely in terms of rational choices based on given sets of preferences (as in rational choice theory) as well as accounts that explain institutions only through backward induction from apparent functions that institutions may serve today (as in functionalist theories) (Graber 2018, at 17–28). The basic precept of historical institutionalism is, as Karen Orren and Stephen Skowronek explain, that ‘because a polity in all its different parts is constructed historically, over time, the nature and prospects of any single part will be best understood within the long course of political formation’ (Orren and Skowronek 2004, at 1).

Constitutional institutionalism draws on historical institutionalism in political science, but focuses exclusively on constitutions: the sets of norms that establish and regulate public authorities, institutional relations among public authorities and the relation between public authorities and the rest of society, including individuals and organisations. Constitutional institutionalism rests on three basic ontological claims: (1) Constitutional norms are regulative institutions: binding rules that legitimise, channel and constrain the actions of constitutional actors. Constitutional norms can take the form of formally enacted legal prescripts, but also of informal conventions and understandings that are accepted by the relevant actors as normative standards for the exercise of public authority. (2) Constitutional norms cannot be understood in isolation: they only have meaning as part of an interconnected whole, a complex ‘constitutional regime’, encompassing the body of all constitutional norms – both formal and informal – of a political community as well as the actors that are constituted by, and interact with, these norms (Tushnet 2003, at 1).²⁶ And (3) constitutional norms are constructed historically: they are shaped and reshaped by interpretation and application in legal, political and societal arenas over time.

Taken together, constitutional institutionalism puts the legal, ideal and factual aspects of constitutions in their historical context. It appreciates that constitutional norms neither originate nor operate in a vacuum, but that they emerge, operate, decay and eventually wear out and die within a broader setting in which a wide range of actors, each with their own views and interests, continuously interact with each other. The constitutional-institutionalist approach thus reflects, for example, the fact that however solemn the forms legal constitutional norms might be casted in, and however firmly entrenched they might be, the content and meaning of constitutional norms may be the object of meaningful contestation. At the same time, the constitutional-institutionalist approach incorporates the formalist awareness that constitutional norms embodied by special legal forms, such as a master-text constitution, may indeed have a relative firmness of autonomy and enjoy special normativity within a political order.

5 The epistemology of constitutional institutionalism

So far, we have presented constitutional institutionalism mainly as an ontology: a generic understanding of what a constitution is. However, in order to be truly useful, constitutional institutionalism also needs to become an epistemology: a methodological way of acquiring knowledge of constitutional norms and their significance.

An explicit constitutional-institutionalist epistemology is not yet available. Some authors (mostly from the US) have suggested approaches comparable to our constitutional-institutionalist approach.²⁷ They have given fascinating accounts of actual constitutional developments.²⁸ But, to our knowledge, not one of them has explicitly and systematically answered the epistemological questions raised by historical-institutionalist-like approaches to constitutionalism. This makes much of the historical institutionalism-inspired studies of constitutions quite hard to verify and indeed to reproduce. Moreover, although these studies take a general perspective that includes formal, factual and ideal aspects of constitutions, the lack of an explicit epistemology still precludes constitutional scholars from ‘sitting at the same table’ and having a constructive conversation with each other about the same and closely related constitutional phenomena.

In this section, we seek to remedy this shortcoming by presenting constitutional institutionalism as an epistemology as well. In the view of constitutional institutionalism, the general aim of constitutional scholarship is to account for constitutional change over time. Such a developmental account of a constitution is more than just a description of the evolution of a relevant body of law. It also identifies the main factors that have shaped its formation and operation and that have determined its effects. Constitutional institutionalism thus advocates constitutional scholarship as an independent academic inquiry that is more theoretical and less practice-oriented than what is commonly called ‘constitutional theory’ (Fallon 1998, at 537; Strauss 1999, at 582).²⁹ Instead of trying to influence or predict the way judges decide difficult constitutional cases, by prescribing how the constitution should be interpreted and applied, the main ambition of constitutional institutionalism is to understand and explain how and why the constitution has evolved as it has.³⁰

The constitutional-institutionalist epistemology involves, on a general level, two steps that need to be taken to acquire meaningful knowledge of constitutions: (1) identifying constitutional norms and (2) describing and explaining their origins, operation and consequences. We discuss both steps in turn.

5.1 Identifying constitutional norms

In accordance with the constitutional-institutionalist ontology, constitutional-institutionalist epistemology seeks to account for constitutions by focusing on the formation and evolution of constitutional norms over time. A crucial, first step in this process consists of identifying constitutional norms, that is, in detecting their existence. How to do this?

Recall that constitutional institutionalism recognises that constitutional norms may be the product of legally authorised constitutional law-making as well as other, non-legal dynamic forces. But, however constitutional norms have come about or what form they take, what matters for their existence is whether they have been effective and accepted by a proportion, at least, of leading constitutional actors in the real world over time. The constitutional institutionalist who seeks to identify constitutional norms should therefore start by (1) identifying actual patterns of behaviour and thought, or ‘empirical’ institutions, relating to the basic ways in which a system is being governed (Llewellyn 1934, at 22). Identifying such patterns requires empirical observation of some sort. As the next step (2), she has to determine which of these patterns have staying power. Incidental patterns that exist for a short period and then disappear again can hardly be considered as having a regulative quality. Variations in behaviour and thought of people in government may be coincidental, the whims of an individual leader or part of the ordinary ebb and flow of politics, while a change in a rule is more structural; that is, it persists beyond the duration of single events or the tenure of individual leaders.

In order to acquire normative quality, a pattern of behaviour must furthermore (3) be accompanied by the opinion that the rule to which it gives expression ought to be observed (Barber 2010, at 85). Actual patterns – also referred to as ‘customs’, ‘habits’ or ‘traditions’ – may influence the actions of people, but they do not necessarily guide them. Pure customs are not backed by social threats. Whoever breaks with them is not subject to social criticism, let alone sanctions. Only when patterns are binding – that is, when actors care about their observance and will criticise others when they fail to respect them – can we regard them as ‘norms’.

Steps 1–3 allow us to identify norms – binding institutions – for people in government. Yet, the group of people in government does not necessarily exert all public authority. In constitutional democracies, for example, relevant constitutional actors typically also include the people and leaders in political parties (Ackerman 1991). Where the ‘constitutional community’ and the people in office do not coincide, institutions can only be regarded as binding if (4) they are accepted, implicitly or explicitly, as valid, or as ‘constitutional’, by a sufficient number of leading constitutional and political actors (officials), whether inside or outside government.³¹

Note that the standards which leading actors use to assess the constitutional validity of norms – also called ‘rules of recognition’ (Hart 2012, at 100–110) – may differ from place to place. In countries such as Germany, the master-text constitution is a prominent and authoritative measure (though not the only one). In the UK, by contrast, the constitutional debate mostly centres around such an intangible concept as the supposed will of the Queen in Parliament. A constitutional institutionalist should foresee that the rule of recognition may be ambiguous and indeed contested (Hart 2012, at 100–110). She should expect that the rule of recognition may not be explicitly stated, certainly not fully (Hart 2012, at 101). And she should be particular sure to anticipate that the rule of recognition which surfaces in legal doctrine may not fully coincide with the actual rule of recognition. As constitutional institutionalists should understand Herbert Hart, the rule of recognition ‘is shown in the way in which particular rules are [actually] identified’ (Hart 2012, at 101).

From the viewpoint of the constitutional institutionalist, in sum, constitutional norms emerge or change where (1) new basic ways of government emerge or where existing ways evolve; (2) these ways prove to have staying power; (3) these ways are regarded as binding by legal and political actors; and (4) these ways are accepted, by explicit consent or in silent acquiescence, as valid by constitutional and political leaders.

Some constitutionalists who have thought about the formation of constitutional norms have suggested that a pattern of behaviour, in order to become a valid constitutional norm, must be based on an adequate normative reason, for example that it must ‘enable the machinery of the state to run more smoothly’ and that it must be ‘desirable in the circumstances of the constitution’ (Jennings 1967, at 136).³² Constitutional institutionalism takes a cautious stance here. It recognises that constitutional scholars may – and sometimes do – judge that a norm that has widely been accepted by the constitutional community is nevertheless invalid because the reasons advanced for its justification are mistaken. When making such judgements, however, scholars turn into political actors, as they no longer merely seek to explore and describe constitutions, but also to affect and change them. The ultimate consequence could be, as Nick Barber warns, that there will be ‘one set of rules governing the functioning of the constitution, and another set in the writings of constitutional scholars’ (Barber 2010, at 84).

5.2 Describing and explaining constitutional norms

Constitutional institutionalism not only provides a basis for identifying constitutional norms, but also for their description and explanation. In particular, it seeks to account for (1) the origins, (2) operation and (3) consequences of constitutional norms. It does so by viewing constitutional norms in relation to the constitutional actors that interact with them, and by studying how the beliefs of these actors shape the emergence and evolution of constitutional norms.

This particular focus is inspired by interpretivism in political science.³³ The basic precept of interpretivism is ‘that to understand actions, practices, and institutions, we need to grasp the beliefs – the intentional meanings – of the people involved’ (Bevir and Rhodes 2016, at 12). In other words, rather than seeking to discover an ‘objective reality’ (as positivism does), we should seek to examine the meanings that human beings themselves impose (Ferdinand et al. 2018, at 21). ‘Interpretative explanations’, as we suggest to call them, account for constitutional actions (including the creation, interpretation and application of constitutional norms) by asking how constitutional actors regard their own behaviour, that is, by asking what this behaviour means to those actors. Such explanations differ from ‘functionalist’ explanations, which are based on the idea that we can account for the existence and operation of constitutional norms by referring to the ‘functions’ they supposedly fulfil, as well as from ‘behavioural’ explanations, which seek to account for constitutional actions by relating them to other features of the relevant actors (for example, their membership in a given elite) (Cf. Closa 2006, at 583). In answering ‘why-questions’ – for example, why is a particular norm adopted, why is that norm interpreted in a particular way, and why have its addressees resisted its application? – interpretative explanations appeal to the beliefs of the different actors involved. Since these beliefs are themselves interpretations of the world in which actors find themselves, a reconstruction of these beliefs takes the form of ‘interpretations of interpretations’ (Hay 2011).

Accounting for (1) the origins of a constitutional norm requires answering the question why that norm exists, and how is has come into being. An interpretative explanation will focus on the relevant constitutional actors in their capacity as ‘rule-makers’ and study the particular ‘actor constellation’ that has made possible constitutional change (and suppressed other possible courses of (in)action). By reconstructing the beliefs of the actors involved, we can find out their motivation for adopting or accepting a new or changed norm. As Howard Gillman et al. explain, constitutional developments cannot be understood without taking into account ‘the most important partisan coalitions that fought for electoral supremacy, the main interests that supported those coalitions, the positions those coalitions took on the most important issues … and the extent to which one coalition was more successful than others’ (Gillman et al. 2016, at xxviii). Reconstructing the beliefs of the actors involved may of course also suggest that norms emerged ‘silently’, that is, without a (demonstrable) intent or awareness of the change on the part of constitutional actors (Passchier 2017A).

By (2) the ‘operation’ of constitutional norms, we mean the ways in which they are interpreted and applied by the relevant constitutional actors, including, for example, judges and civil servants. Constitutional institutionalism proceeds from the assumption that most constitutional norms are ambiguous; the meaning of norms is seldom self-evident and always in need of interpretation, even where they are casted in plain, written language. Different actors can see and interpret the same norms in different ways. Such differences may give rise to contestation and conflict over the meaning of the norm and can even lead to its gradual modification over time (Heller 1996, at 1180). Instead of suggesting the one right answer to questions of interpretation, constitutional institutionalists take a more distanced perspective and take stock of the various interpretations that are suggested, rejected or adopted, and inquire how the beliefs held by the relevant constitutional actors, as ‘rule-appliers’, motivate their interpretations and the way they apply the relevant norms.

Finally, constitutional institutionalist also study (3) the consequences of constitutional norms: they want to know the consequences of different constitutional arrangements on the ways in which public authority is exercised. While formalist scholars sometimes simply assume that constitutional norms ‘matter’, constitutional institutionalists acknowledge the fact that we do not yet know much about the actual performance of constitutions (Law 2010). Ideally, constitutional norms motivate actors to do their constitutional duties (Fallon 2009, at 995–997). In fact, however, ‘compliance’ is far from self-evident, even in constitutional democracies. Hence, we must also observe the actual behaviour of the actors to which constitutional norms are directed (the ‘rule-takers’) and investigate the reasons they give for their actions, as well as their own interpretations of the norms (Schauer 2012, at 212).

6 By way of conclusion: a constitutional-institutionalist research agenda

This article has presented constitutional institutionalism as an historical and cross-disciplinary approach to the study of constitutions. One of its contributions to constitutional literature is that it provides an ontology of constitutions which may allow students of constitutions with different disciplinary backgrounds to (finally) ‘take a seat at the same table’ and have a constructive conversation about what a constitution is and how different insights into it relate to one another. The explicit epistemology of constitutional institutionalism may then allow us to operationalise the study of constitutional development and show what constitutional institutionalists are actually doing – or are supposed to be doing – when they look into actual constitutional regimes and reach their conclusions.

The constitutional-institutionalist approach may provide constitutionalists the possibility to take up some new research challenges. It may, in the first place, be helpful for anyone who seeks to understand the significance of constitutional norms, understandings and conventions within the broader context of a political order. It may help legislators and judges, for example, to identify path-dependencies in political orders, and hence, to better determine what kind of decisions are sensible and practical in terms of what they can actually achieve in the real world. Any law or judicial decision that does not account for the environment in which it is supposed to take effect and operate, may at best be irrelevant and in the worst-case scenario undermine the authority of its authors. Our constitutional-institutionalist approach thus suggests new and more meaningful ways to read and study laws and judicial decisions and draft them effectively.

Our constitutional-institutionalist approach furthermore may be helpful in providing empirically informed nuances to optimistic claims regarding the possibility of constitutional ‘design’.³⁴ Contrary to what constitutional ‘engineers’ sometimes assume, institutionalist scholarship points to the fact that there is only relatively little knowledge about the extent to which and the conditions under which constitutional norms are likely to produce the institutional change that generates intended substantive effects (March and Olsen 2008, at 15). That does not mean that legal forms cannot shape reality at all. Rather, it means that constitutional designers need to learn how to ‘embrace’ (Cf. Boulton et al. 2015) the complexity of our high-speed age in which their frames are supposed to operate (Cf. Scheuerman 2002; Hirsch Ballin 2020, at epilogue). The constitutional-institutionalist approach may be particularly helpful in taking up this challenge, as it may allow us to account for the interconnectedness of different constitutional phenomena and developments, and the dynamic and unexpected ways in which context and history may shape and reshape constitutional norms.

Our approach may also shed a fresh light on how constitutions change, that is, the multifarious ways in which constitutional norms emerge, operate, decay and eventually wear out and ‘die’. Constitutional institutionalism may thus suggest remedies to what Tom Ginsburg has labelled ‘the 7^th inning problem in comparative constitutional law’: the habit of constitutional scholarship to focus exclusively on the 7^th inning of constitutional law-making, thus tending to disregard the 1^st to 6^th and last two innings, let alone relating all ten innings to each other (Ginsburg 2016). Constitutional institutionalism, we believe, has the potential to liberate the topic of constitutional change from the dogmatism of legal scholarship, the rigid idealism of legal and political philosophy and the relative superficial realism of political science. Instead, it is capable of revealing long-term stories of interplay between the existing norms that govern government and the underlying structure of societies.

Finally, we believe that our constitutional-institutionalist approach may be helpful in reconfiguring important debates in constitutional theory. We provide three suggestions in this regard. First, following the distinction between ontology and epistemology, constitutional institutionalism may assist debates over originalism in constitutional interpretation and put such debates in their proper perspective. One hypothesis in this regard is that the historical interplay between original meanings and originalist interpretations on the one hand, and non-originalist constitutional interpretations as well as real-world practices on the other, cannot be evaded, whether one likes it or not. Second, the debate over unconstitutional constitutional amendments might benefit from a recognition of how different ontologies are being related. Taking a constitutional-institutionalist perspective might reveal, for example, that including entrenchment clauses in written constitutions is nothing less than an attempt ‘to limit the illimitable’ (Loughlin 2010, at 305). Third, debates over informal constitutional change might benefit from the constitutional institutionalists ability of showing how emerging informal constitutional norms relate to the legal aspects of a constitution and vice versa, appreciating that different kinds of legal and non-legal forces may shape and reshape a nation's basic normative structures over time (Passchier 2017B).