Faux ami? Interrogating the normative coherence of ‘digital constitutionalism’

Abstract

This article interrogates the normative coherence of the label of ‘digital constitutionalism’. In particular, I argue that the use of the label ‘constitutionalism’ in digital contexts often conflates the practical realities of existing contractual governance models with the superficial appeal of constitutional structures. As a result, the label is misleading in both normative and qualitative terms as it obscures the true nature of the governance architectures to which it is applied, which are more appropriately understood as implementing a distinct genre of ‘private policy’.

I. Introduction

The internet and the actors who have proliferated within the digital spaces it has provided have long been subject to competing arguments about whether they should be subject to public or private regulation. ¹ Against this background, the immediate attention of academics, practitioners and governments examining the regulation of digital spaces has focused on how private actors function as governors of the technical infrastructures of the digital environment. This scrutiny has recently expanded to look beyond regulatory approaches to infrastructure to consider how private actors in digital spaces affect public constitutional ² values through their regulation of speech, privacy ³ and other values traditionally protected in Constitutional documents. ⁴ In response to this scrutiny, as well as the fragmentation of traditionally state-controlled functions, private actors have adopted increasingly formalised, private systems for adjudicating claims about constitutional issues as part of an attempt to maintain independent regulatory control of digital spaces, and in response to government demands that these spaces do not become wholly a-constitutional.

These attempts to ‘constitutionalize’ private self-regulation of digital spaces are particularly evident in Twitter’s decisions to police political speech and remove content it deems to infringe individual privacy. ⁵ Facebook’s Oversight Board though Google’s internal system of adjudication in relation to ‘right to be forgotten’ claims presents a similar, albeit less transparent, model of private actors assuming quasi-judicial roles in assessing and enforcing constitutional values. ⁶ Elsewhere, the peer governance mechanisms employed by Wikipedia have been proffered as an example of the ambiguity and ‘tyranny of structurelessness’ that opaque, private governance efforts can generate even when they are modelled on disparate rather than centralized power structures governing digital spaces. ⁷ This capacity to shape constitutional values online, and the desire to do so, have been articulated by some scholars as representing ‘digital constitutionalism’; as part of this, private actors have internalized public Constitutional values and integrated them within the regulatory architectures of private digital spaces. ⁸

These attempts to ‘constitutionalize’ private regulatory structures have not gone unopposed. In the face of private efforts at regulation which are perceived as insufficient or unpredictable in terms of enforcement, governments have pushed back – seeking to regulate or oversee the operation of such systems or requiring particular, parallel protections to those provided by traditional Constitutional protections to be put in place. ⁹ In the United Kingdom ¹⁰ and Ireland, ¹¹ for example, governments have proposed the introduction of legislation imposing a broad range of obligations on companies in respect of how illegal and harmful content is treated online. The proposals in both jurisdictions effectively grant the designated regulator (Ofcom in the case of the United Kingdom and the proposed Media Commission in the case of Ireland) powers to oversee the policies designed and implemented by private actors to ensure they perpetuate the appropriate balancing of fundamental rights and interests in digital spaces. Similar legislative proposals have been progressed in Germany through the amendment of the existing Network Enforcement Act ¹² and also in France. ¹³

Much of the regulatory debate between private actors and states has thus focused on the capacity (or incapacity) of private actors to define and enforce public, constitutional values, either on their own terms or at the behest of state actors, but has ignored the more fundamental questions raised by the invocation of an idea of ‘digital constitutionalism’. What precisely is meant by ‘digital constitutionalism’ and what normative, structural or qualitative characteristics must be present for privately led regulatory efforts to be considered to have become a ‘constitutional’ project? In this article, I interrogate the meaning of ‘digital constitutionalism’ and seek to offer some preliminary thoughts on what ‘digital constitutionalism’ means – and what it does not. ¹⁴ In particular, I argue that the regulatory conduct of private actors in digital spaces does not evidence the characteristics that would justify the application of the label of ‘constitutionalism’. While certain applications of the label are partly accurate, a preponderance of the measures to which the label is ascribed do not possess the normative or structural characteristics that would justify the label of ‘constitutionalism’ being applied to them.

The majority of the online governance structures that have adopted constitutionalist language to self-describe their efforts should be viewed not as constitutionalist, but rather as demonstrating the emergence of ‘private policy’ architectures. As part of these architectures, private actors rhetorically espouse a commitment to constitutional values that obscures the true contractual justification and function of the regulatory methods they employ, and in doing so benefit from the presumptive of normative legitimacy of constitutionalism without offering the equivalent protection that a constitutionalist system would ensure for those governed by it.

In this respect, the examples of the Facebook Oversight Board used in this article are illustrative of the broader mismatches between the normative claims of digital constitutionalism and its substantive form and function. Recognizing this disparity is important, I argue, on the basis of two factors. The first is its reorientation of the analytical focus in assessing the normative use of the term ‘constitutionalism’ towards a more fundamental interrogation of the ways in which current governance structures operate. The second is a more precise appreciation of whether these governance structures are indeed ‘constitutional’ in nature, which permits us not only to answer how online spaces are governed at present, but also how they ought to be governed in future. The analytical reorientation prompted by critically engaging with the coherence of digital constitutionalism as both a label and an idea thus leads us to a substantive engagement with the existence and quality of the protections that private regulation in digital spaces can provide to fundamental individual rights and constitutional values.

II. Defining digital constitutionalism

The first scholar to deploy the language of constitutionalism in relation to the digital environment was Brian Fitzgerald, who argued for the recognition of ‘informational constitutionalism.’ ¹⁵ In particular, FitzGerald’s argument was that the decentralized and globalized design of the digital information society required a mixed regulatory model in which both private and public actors participated and through which a new constitutional order would be determined by ‘a blend of intellectual property law, contract law, competition law, and privacy law’. ¹⁶ Lessig later picked up the threads of Fitzgerald’s understanding of private actors as de facto regulators who should be regarded as such in his arguments about the four pillars of regulation in the digital environment. ¹⁷ It was Berman, however, who first engaged substantively with whether the ‘code-based power’ of private actors to regulate digital environments should justify the extension of constitutional principles to their activities. ¹⁸

In particular, Berman argued for the extension of constitutional adjudication through traditional judicial review to a broader range of cases in order to foster constructive societal debate about difficult social and political issues, and permit courts to perform an ‘educative function by articulating the values that help constitute our national identity.’ ¹⁹ In particular, he argued that private law was insufficient as a regulatory mechanism and that by extending the application of constitutional values through judicial review, the state could ensure that the Constitution as a touchstone for articulating the constitutive values of the social and political lives of its citizens was not diminished by digital contexts. ²⁰ These accounts do not claim that the nature of the governance of digital spaces is constitutional, but rather that constitutional values or Constitutional principles should be applied to these spaces and the activities that take place within them. More recently, in describing ‘digital constitutionalism’, ²¹ Suzor – similarly to Berman – has identified a need to extend the provisions of a ‘constitutional’ structure to the digital environment ²² as a result of the potential for privately imposed regulatory structures to infringe constitutional values and Constitutional principles.

Suzor’s account echoes the model that Fitzgerald might ultimately have endorsed, arguing for the adoption of a more coherent law of contract that internalizes and thus extends the role of constitutional values. ²³ Crucially, Suzor views this meshing as necessary, not on its own merits but a result of a system in which public law has lost its pre-eminence and in which attempts to reassert it have largely failed. ²⁴ This view is, in many respects, aligned with the one presented by Teubner as part of his model of ‘social constitutionalism’. ²⁵ Teubner specifically argues that the idea of the constitution is now projected beyond the state and is no longer anchored to the national dimension from which it emanated. In that context, Teubner argues that a transnational, digital constitution is emerging through a series of ‘civil constitutions’ – sets of constitutional norms developed by non-state actors that are gradually transmuted into positive law through a process of mutual influence. ²⁶

While Teubner’s account of societal constitutionalism challenges traditional state-centric definitions of the constitutional question and of constitutionalism itself, ²⁷ his articulation of the precise mechanism by which disparate, private efforts can generate constitutional effects does not assume a capacity for self-designation as constitutional in the manner that other recent accounts do – notably those offered by Gasser et al. ²⁸ Rather, Teubner argues that these non-state forces must be understood as one part of a pattern of globalized social redefinitions of constitutional content in which state and non-state actors are parties and that may generate an independent transnational constitutional regime. ²⁹ In contrast, Gasser et al. use the term ‘digital constitutionalism’ as a descriptive label to refer to a diverse group of codified rights documents that seek to impose variously defined principles (broadly allegiant to general constitutional values) on private actors in the digital environment. ³⁰ Gasser et al defend the use of the label of constitutionalism on the basis that the models and documents identified in their work ‘undeniably’ display the values, problems and principles of constitutionalism inasmuch as they afford primacy to constitutional rules within a ‘hierarchy of legal norms’. ³¹

While these accounts vary in both the manner in which they understand what constitutionalism is, and how a constitutionalization of digital spaces could (or should) be achieved, what does unite them is a clear recognition that private actors in the digital environment were (and are) generating a system of norms that variously reflects, modifies, reshapes and even over-rides existing Constitutional principles and constitutional values. ³² In this respect, all of the accounts examined (with the exception of that presented by Gasser et al, who use the label descriptively rather than in evaluative terms) are broadly concerned with seeking to transmute vertically enforceable Constitutional guarantees to private actors in a manner that mirrors debates over the relative merits of doctrines of direct and indirect horizontal effect with which various European appellate courts (among other constitutional orders) have long grappled. ³³

In these accounts, digital constitutionalism is driven by two distinct concerns. The first, evinced by critiques such as those offered by Teubner, is the fragmentation of the traditional sovereignty and associated social orders that once informed law. In this respect, digital constitutionalism is concerned with integrating private actors into a more broadly drawn conception of the sovereign or Republican constitutional project in order to ensure the values of that project emanate into areas in which the traditional state enjoys little – or certainly less – control. ³⁴ The second concern, which perhaps underlies the accounts of Gasser et al most distinctly, is arguably tinged with cyberlibertarian themes inasmuch as it views digital spaces as distinct jurisdictional areas with sovereigns of their own, whose regulatory choices act as equivalent systems to traditional state-led constitutional models but may be influenced to echo the values and principles of traditional constitutionalism within the confines of their private spaces. ³⁵

Interestingly, what emerges from accounts driven by these distinct concerns is a range of markedly similar responses that view digital constitutionalism as effected by a range of privately defined standards that variously work independently of the state or displace state institutions and structures, ³⁶ and adopt only the descriptive rhetoric of constitutionalism, seek to integrate constitutional values into private law regulatory mechanisms or adopt a posteriori constitutional models, which measure the activity of private actors in digital spaces against constitutional values through limited judicial review. ³⁷ Precisely this diversity in the use of the label, but in particular the divergence between those using it as a rhetorical label (such as Gasser et al) and as a descriptive one, could be considered to cause the accounts to fall foul of Waldron’s warning that constitutionalism can be used, too easily, as a flag of convenience. What is more problematic, however, is that these accounts largely fail to satisfy the substantive requirements for designation as ‘constitutionalist’ efforts. It is not necessarily the absence of a state-centred definition that is problematic in this respect – although they are at odds with traditional state-oriented accounts of constitutionalism that position the state as the source of constitutional development and governance. Yet in a digital landscape which is increasingly polycentric – or perhaps multipolar – in its governance, this is not the primary shortcoming of these accounts, which justifiably seek to capture how various, and sometimes competing, sources and objectives of regulatory power can be reconciled as part of a mutual determination of constitutional objectives by private and state actors. What these accounts do not necessarily unpick, however, is whether and to what extent the normative use of the label ‘constitutionalism’ accurately describes the character or practical impacts of the activities to which it is applied.

III. The normative core of constitutionalism

The normative appeal of constitutionalism is almost instinctive. It imports into the contexts in which it is used a vision of public goods and popular protections backed by state guarantee. This appeal is evident in the broad adoption of constitutional models across a range of legal systems and disciplines. ³⁸ We need think only of the ‘constitutionalization’ of EU law, ³⁹ international economic law ⁴⁰ or contract law ⁴¹ to see evidence of this. More generally, constitutional democracy remains entrenched in public opinion as the aspirational form of governance, and while constitutions themselves are subject to frequent amendment and reinvention, their proliferation and endurance nevertheless indicate their majoritarian appeal as both political and legal structures around which public conceptions of legitimacy and the protection of individual rights can be structured. ⁴² The normative appeal of digital constitutionalism is rooted to no small extent in its invocation of these popular conceptions, which connect constitutionalism with the protection of individual rights, the integration of public, constitutional values within regulatory structures and the restriction of unchecked, centralized power. Yet it is not clear that digital constitutionalism secures all that its label implies.

An analysis of whether the label of digital constitutionalism is substantively coherent in the sense in which it is being currently used must necessarily begin with an understanding of the content and meaning of constitutionalism simpliciter. The word ‘constitution’ itself is defined by McIlwain as referring to those laws of the state (whether that state is democratic or monarchical) that result from enactments rather than custom and that provide the national framework of the state and the public law of the realm. McIlwain traces this definition through Whitelock to Cicero ⁴³ before advancing this basic definition a step further. Drawing on Bollingbrooke’s understanding, McIlwain defines the constitution as the assembly of laws, institutions and customs derived from certain fixed principles of reason directed to certain fixed objects of public good that compose the general system according to which the community is governed. ⁴⁴ While this state-centred origin of constitutionalism rather neglects the influence of private actors on the generation and recognition of Constitutional documents, ⁴⁵ it does reflect the normative presumptions that rest at the heart of the constitutional project – that it is directed towards centralizing and structuring disparate sources of rules as part of a legal hierarchy that is used to govern for the benefit of that community to which it applies, restrains the central authorities who exercise those rules so they do not become tyrannical and affords the governed community a central role in assuring accountability of the rules hierarchy itself.

While this tells us what constitutions are, and to some extent what they seek to do, it does not tell us what constitutionalism is. This, of course, is the central issue in any interrogation of constitutionalism - the term itself is rarely substantively interrogated with the same enthusiasm that characterizes its deployment (at least by those not actively engaged solely in this pursuit). ⁴⁶ Waldron is alert to this potential for constitutionalism to degenerate into an empty slogan, a tendency that he notes is evidenced by the word sometimes being used in a way that conveys ‘no theoretical content at all’. ⁴⁷ Moreover, even where constitutionalism is deployed in a substantively coherent manner, it is variously used as a description of the study of constitutions themselves, to indicate an area where normative rights concerns are now being considered, ⁴⁸ or used as shorthand for a tendency to codify existing obligations or restraints into written documentary form. ⁴⁹ To avoid the degeneration averred to by Waldron, the label of ‘constitutionalism’ must do more than merely indicate a superficial allegiance to something that could generally fall under one of these headings. Rather, constitutionalism should be understood as referring to these patterns as part of a broader tapestry that is characterized by fidelity to constitutional values (most often evidenced through fundamental rights guarantees) and Constitutional principles (most obviously exemplified in the structural requirements imposed by the rule of law and a traditional tripartite separation of powers model). ⁵⁰

Moving from McIlwain’s ideas of the definition of a constitution, and thus of constitutionalism, through the later accounts (and indeed critiques) offered by the likes of Barber and Waldron, a substantive concept of constitutionalism begins to emerge. This concept views constitutionalism as descriptive of theories that broadly concern themselves with the congruence of a particular area or approach with constitutional values and Constitutional principles. Constitutionalism thus describes contexts in which a central authority (not confined to the state) that seeks to centralize, organize and exercise power:

• • is fundamental, in the sense that it forms the axis around which the legal system of which it is constitutive turns, and

• • is directed toward securing a conception of the public good through,

  • ○ the restraint of State power (negative constitutionalism) where such restraint is accomplished either,

    • – by normative means e.g. the imposition of rights-based restrictions, or

    • – through the specific abrogation of powers e.g. the precise delimitation of the competences of discrete actors of categories of same, and

    • – the creation of effective and competent institutions with the capacity to use their power to promote the public good (positive constitutionalism).

Constitutionalism thus does not just have a negative dimension – effected through the imposition of restrictions on the power of a central authority ⁵¹ – but also a positive dimension – requiring the presence of institutional capacity in order to pursue constitutional objectives. ⁵² In these broadest terms, constitutionalism can, as Barber conceptualizes, be tightly connected to the state. ⁵³ However, it may also include those who act within and upon the structures of the state, to whom state powers are delegated or outsourced, or who exist outside the vertical constraints of traditional constitutional structures but are placed in the position of the state, as a de facto central authority, as suggested by accounts of modern constitutionalism, such as Teubner’s. ⁵⁴

It is not for this article, or indeed this author, to advance a new account of the meaning and content of constitutionalism. Rather, the account given above is necessarily a composite that offers a general picture of the characteristics of the theory and its normative content. Building on these foundations, however, it is possible to identify a normative core of constitutionalism (whether applied in traditional state-emanating models or more ‘social’ models such as those advanced by Teubner) that provides the mandatory minimum of a political community in the form of its customs, values and institutional ordering. ⁵⁵ In seeking to extrapolate a universal normative core that might apply to constitutional efforts that are transnational (as digital constitutionalism necessarily is), the structural features of constitutionalism (the Constitutional principles) will be central not least because constitutional values in the form of fundamental rights to freedom of expression, for example, may enjoy meanings and scopes that differ significantly across jurisdictions.

Despite such divergences, I would suggest that the normative core of constitutionalism requires, at a minimum:

• • structural restraints on the power of the central authority (whether this is the State or a private actor)

• • binding rules that compel the central authority to abide by these restraints and that enforce compliance where necessary, ensuring minimal accountability

• • a minimum requirement that the rules promulgated by the central authority are clear, accessible, prospective and enforced in a non-arbitrary manner, and

• • that the subjects (citizens in a state-bound model, or users in a digital model) are entitled, as of right, to rely on the central authority’s ongoing compliance with those rules in place, and to challenge the authority where such compliance is absent.

A final requirement for the normative core is that the central authority observe and respect certain minimal constitutional values, commonly in the form of foundational fundamental rights triggered in digital spaces, namely the right to privacy and the right to freedom of expression. This final requirement, while subject to fluctuation in terms of the meaning attributed to such rights between jurisdictions, is nevertheless explicitly averred to as a central feature of digital constitutionalism by a majority of the accounts to which the label has been applied. It has thus arguably assumed a position within the normative core of constitutionalism – certainly in a digital context – on the basis of the role attributed to it by those invoking the label and applying it to their efforts. Despite this invocation of constitutional values – in the form of fundamental rights standards – the accounts grouped under the heading of digital constitutionalism to date are notable for their failure to satisfy the other criteria that form the normative core of constitutionalism. ⁵⁶ In this respect, the label ‘digital constitutionalism’, while it capitalizes on the perhaps reflexive normative appeal of the constitutional project, offers false reassurance.

IV. Digital constitutionalism: A faux ami?

In part, this false reassurance stems from the ambiguities in and conflicts between the descriptions and invocations of digital constitutionalism offered to date. In seeking to identify a single articulation of digital constitutionalism from among these accounts, Celeste has defined digital constitutionalism as a blanket term referring to an ideology that seeks to establish and guarantee the existence of a normative framework for the protection of fundamental rights and balancing of powers in the digital environment. ⁵⁷ This is, of course, broadly correct. However, I would argue that this commonality is not sufficient to justify the use of the label ‘constitutionalism’ equally in respect of each of the models proffered in a manner indicating that they represent a coherent body of work speaking to the same intellectual endeavour.

Fitzgerald, for example, presents what is most accurately described as a model that views constitutionalism as referring to a new body of norms internally generated by private laws in the digital environment. Berman, meanwhile, presents a mechanism for securing greater social discourse about constitutional values in digital spaces as part of an account that is broadly aligned to digital constitutionalism as achieved through limited judicial review. Suzor comes closest to articulating a model that both seeks to secure digital constitutionalism and articulate those substantive values that are constitutive of it, viewing digital constitutionalism as resulting from a projection of fundamental rights and the rule of law onto actors in the digital environment through private law mechanisms and associated legislative processes. Against these accounts, Gasser et al use the term as a blanket descriptor for a diverse range of documents that describe rights-based standards that may be applied to the digital environment. The initiatives these authors variously seek to include under the umbrella of ‘digital constitutionalism’ thus, in reality, have little in common in terms of their core normative content. Inasmuch as a common theme can be discerned, it is limited to the fact that each of the approaches recognizes a need for the extension of certain fundamental rights standards to private actors and/or the state in the digital environment. ⁵⁸ However, this alone does not satisfy the requirements of constitutionalism.

The use of the label by Berman is generally unobjectionable – albeit his model would more accurately be described as seeking to secure constitutional dialogue rather than being constitutive of constitutionalism itself. Fitzgerald’s model is minimal and broadly drawn, articulated as descriptive of the emergence of a new set of values as determined by the interaction of private laws in digital spaces. This account problematizes the misleading nature of constitutionalism as a label. While private law may operate as a mechanism for constitutionalizing the digital space, it is cannot autonomously generate, by reference only to fundamental rights-inflected standards, a ‘constitutionalized’ landscape. In this respect, accounts such as Fitzgerald’s fail to satisfy the normative core of constitutionalism because they focus on broadly drawn fundamental rights while neglecting the structural and institutional controls required by the minimum content of constitutionalism.

This failure runs through even the more traditional ‘constitutionalist’ type accounts. Berman, for example, displays a clear commitment to securing constitutional values (in the form of fundamental rights) not through an integrated a priori model based on private law mechanisms, but rather through an a posteriori system of judicial review. While this, of course, implies an a priori extension of justiciable constitutional standards to private actors, it simultaneously leaves open the question of how rights balancing would operate in such a context, and what constitutional standards or principles would be considered justiciable. Moreover, it does not clarify how the structural requirements of the normative core are imposed on such settings. Indeed, it is implicit that they would not be. It is hard to foresee an occasion where, for example, a private actor would be deemed to be subject to the kinds of institutional restraints placed on the state to prevent it from accruing disproportionate power. While fundamental rights standards might fulfil this role to a degree, the requirements of legal certainty, accountability and non-arbitrariness are less easily accommodated (and are not foreseen) by Berman’s account. Ultimately, Berman leaves unanswered the questions of what the constitutive content of constitutionalism is and what the structural design of a constitutionalist model requires, in favour of a model that would seek to debate and secure fundamental rights, however defined.

Of the accounts offered, Suzor’s is the most substantively coherent, evincing a vision of digital constitutionalism that seeks to integrate the requirements of the rule of law into a model of digital constitutionalism that emanates from the state on an a priori basis and that can be enforced through the laws of contract in a way that is supportive of fundamental rights. Suzor’s particular focus on securing legal certainty, prospectivity, due process and equal application of the laws grounds a vision that coheres with constitutionalism’s normative core as part of a model of constitutionalism directed towards securing the public good. A critical examination of Suzor’s model could inquire about the particular way in which his model is or should be differentiated from the standard practice of using constitutional values to frame legislation and statutory regulation. ⁵⁹ A legal realist might, equally, note that the contract law mechanisms used to achieve Suzor’s aim are not fit for the task, given their inherent tendency towards individualistic, rather than communal, goods. ⁶⁰ The account given by Suzor is, nevertheless, the most substantively coherent, and the lack of novelty in focusing on the law of contract must be balanced against the specifically defined substance of the model and its objective in extending not simply fundamental rights protection but also the structural restraints of the rule of law to private actors, a feature that arguably carries the account over the line in meeting the requirements of the normative core of constitutionalism.

Or does it? The fundamental criticism of digital constitutionalism in its current form – that it is unable to articulate precisely how it is constitutional rather than constitutionally inflected – lingers at the heart of Suzor’s account. The rhetoric of constitutionalism emphasizes the legitimacy of governance structures as the central justification for the exercise of power. Accounts of digital constitutionalism – including Suzor’s – seek to capitalize on that association with legitimacy but without engaging in an in-depth analysis of precisely why such legitimacy is present within the proposed (or actual) governance structures. While adherence to broad fundamental rights standards or the requirements of the rule of law go a significant distance towards ensuring this, they do not impose broadly structural restraints on the power of the central authority, nor do they provide rules that compel the central authority to abide by these restraints into the future or that enforce compliance.

This final point begins to lead us to the heart of why constitutionalism is a normatively misleading descriptor of the activity of private actors in digital spaces. The label of ‘constitutionalism’ presumes a mutual accountability between the central authority (whether the state or a private actor) and the subject (whether citizen or user), which is not present in digital settings. The relationship between the central authority and subject in the contexts described by digital constitutionalism is governed not by mutual accountability but by contract. In this respect, while the governance model employed by private actors in the digital environment may incorporate constitutional values and Constitutional principles – may even, as in Suzor’s account, incorporate a majority of the requirements of the normative core of constitutionalism – the basic structure and content of the relationship itself will not be characterized by the process of dialogic, mutual definition in which subjects are recognized as, prima facie, power holders who may enforce accountability and impose limitations upon the central authority. Rather, in a digital context, the central authority determines the extent and limits on its power without any formal (and with variable informal) reference to subjects. Moreover, its choice to depart from previous models and rules of self-regulation or to be bound by the rules it has previously established for itself is neither constrained nor controlled by the approval of the subjects, but instead by the requirements imposed by private law.

In this model, while a subject may object to a breach of contract, or may enforce a right afforded by consumer protection law, the foundational dynamic of the relationship between the parties is driven by a self-empowered and minimally constrained central authority and a consenting subject. That subject has certain minimal entitlements dictated by consumer protection law, but it cannot be said that these are ‘constitutional’; rather, they are largely drawn in informational terms – providing limits on what must be communicated to the subject . ⁶¹ While private law (including consumer protection law) is thus minimally constitutional inasmuch as its provisions do not breach constitutional rights, it cannot be said to be presumptively aligned with, nor to further, the normative core of constitutionalism. Informing parties about the powers of the central authority is not, after all, the same as restraining that power. More broadly, where private law constraints intervene in the form of competition law or consumer protection law, they do not weave some broad supranational constitutional construct. The qualifications these laws impose can modify the manner in which power is channelled within the contractual relationship, or they can ensure a greater number of actors can establish themselves as central authorities within a market, but they cannot, by some process of transubstantiation, transform a contractual relationship into a constitutional one.

This conceptual ambiguity over what can and cannot be characterized as part of a constitutionalist model is symptomatic of a broader misunderstanding about the nature of regulatory ordering and sources of power in the digital environment. While there is nothing to prevent the use of a fundamental rights-centred self-regulatory model – and such models are indeed to be encouraged where they ensure greater practical protection of fundamental rights in digital spaces – the nature of the label ‘constitutionalism’ is that it denotes a controlled expansion of constitutional structures as well as values. The requirements of the normative core of constitutionalism must be present within an architecture characterized by the mutual assent and interdependence of the central authority and subjects of that system.

Celeste has argued (echoing Fitzgerald) that digital constitutionalism represents a new era in the development of constitutionalism, ⁶² but shares the foundational values and the overall aims of constitutionalism writ large and represents merely an extension of those aims to the digital context. ⁶³ Certainly, where digital constitutionalism is used as a descriptive moniker for one of several models through which states may extend the jurisdictional capacity of constitutional values to private actors in the digital environment, this could be correct. Indeed, it is also correct in a conceptual manner – the normative core of constitutionalism, after all, is just this – an attempt to identify the requirements of constitutionalism and apply them in a new digital context. But Celeste’s description cannot accurately describe the relationship that subsists between a more broadly defined idea of a central authority and its subjects in digital spaces – and when we consider the normative core of constitutionalism in a context beyond traditional state-centred visions.

Exposing this allows us to refocus on why adopting the language of constitutionalism may be harmful. The most basic and abstract problem is, of course, conceptual – that the label of constitutionalism is not appropriate to the structures and relationships to which it is being applied in a digital context. This, in turn, leads to confusion over the source of power in digital spaces, and thus clouds our understanding about the legitimacy of such power, its justifications, the motivations that underlie it and the objectives it serves. This opacity in combination with the label ‘constitutionalism’ can cause users to assume they are subject to protections equivalent to those afforded to them in traditional, constitutionally governed settings. The label ‘digital constitutionalism’ can thus operate as a ‘faux ami’ – a label that offers a superficial reassurance that is not borne out in practice. Indeed, I would argue that, given the failure of many accounts of digital constitutionalism to satisfy the requirements of the normative core, and the practical governance models used in digital spaces, it is more productive to think about the regulation of digital spaces in terms not of public and private law but private and public policy.

V. From digital constitutionalism to private policy

Moving beyond the language of digital constitutionalism, but in accounts that are still coloured (to some extent) by a constitutionalist tone, both Jack Balkin ⁶⁴ and Kate Klonick ⁶⁵ argue that private actors, and specifically online platforms, have emerged as ‘new governors’ of rights. Rory von Loo also argues, in a more general account, that private actors in the digital environment now act as ‘gatekeepers’ for public policy enforcement – creating ‘an expansive area of unaccountable authority’ within modern regulatory states. ⁶⁶ This shift away from the language of constitutionalism and towards a focus on governance deals more critically with the regulatory models adopted by private actors, and is a step towards recognizing that such actors, while they may indirectly influence constitutional values, are not doing so as part of a constitutional project but are instead engaged in regulating through independently – and privately – developed policies.

Goodin et al define public policy as the system through which officers of the state exercise authority ⁶⁷ and by which the state seeks to contribute to the betterment of the life of its citizens. ⁶⁸ It is notable that public policy in Goodin’s account includes within its scope the informal, customary patterns of action and practice extant within the state and not only legal standards. Public policy thus looks beyond the purely legal architectures and de jure powers of a system and its architects, which are encompassed by constitutionalism, and includes the de facto power being exercised to secure certain objectives. Public policy thus includes not only the normatively driven question of what states ought to do, but also the practical reality of what they do. The other definitional feature of public policy in Goodin’s account is its orientation towards the resolution of political problems in the service of the public good. ⁶⁹ Lasswell and Kaplan, two of the founders of the field, have argued that public policy is dedicated toward the provision of ‘intelligence pertinent to the integration of values realized by and embodied in interpersonal relations’ as part of a model of governance that ‘prizes not the glory of a depersonalized state and the efficiency of a social mechanism, but human dignity and the realization of human capabilities’. ⁷⁰

In this respect, public policy can be understood as an integrated component of constitutionalism, in that it is explicitly value laden and prescriptive – recommending to those state actors to which it is directed a specific approach that dictates what the state should do and on what normative basis within the structures imposed by law as well as embracing what the state in fact does. ⁷¹ Thus, even where it pursues commercial or economic ends, public policy is intended to have as its ultimate aim securing the public good through integrating social values into the decision-making structures, rules and regulatory architectures imposed by and on the state. ⁷² In this understanding, as in theories of constitutionalism, private actors are appropriately enlisted to improve performance in the creation of public value but are so invoked to support rather than supplant the works of the state.

The project of digital constitutionalism claims that it has similar ends – striving to integrate public, constitutional values into its regulatory structures in a manner that promotes a generally defined empowerment and protection of users. Yet, in practice, the professed public aims of these private structures is less clear. Previous research has established that, despite expressing a desire to promote user privacy, private actors in digital spaces operate in ways that maximize profit rather than the fundamental rights of users. ⁷³ Concerns have also been raised about the manner and efficacy with which speech is policed on online platforms and the extent to which market concerns rather than user rights drive concerns about permissibility. ⁷⁴ Inasmuch as there is a policy that is being pursued in such contexts, it is more accurately described as one based on ‘private policy’.

In this respect, it is worth emphasizing that, as with its public counterpart, private policy includes not only formal mechanisms – for example, contractual rules and corporate governance standards – but also the practical behaviour and operation of private actors in securing and enforcing such rules and standards. While formal expressions of private policy may thus be influenced by the regulatory demands of state actors and their public policies, and may be constructed in a way that integrates certain constitutional values, we must look beyond this to understand the character and composition of private policy in practice. Experience has shown that, at its most fundamental level, private policy is organized by the law of contract, which provides the practical mechanism for the implementation of the rules and standards governing digital spaces. ⁷⁵ The relational character of this ordering mechanism, as well as the practical conduct of private actors in ‘empowering’ users, indicates that the character of the interests pursued in digital spaces is particularly private not only inasmuch as it is effected through private law mechanisms, but also as a result of the specifically private interests it seeks to attain. ⁷⁶

As such, private policy cannot be considered an integral component of constitutionalism in the same manner as its public counterpart. While adopting constitutionalist rhetoric or apparently constitutionalist features may form part of the design of private policy inasmuch as it is one of the mechanisms by which private actors may seek to secure consumer or user confidence in the furtherance of their objectives, it is separate from constitutionalism itself. These governance structures should thus be understood as a deliberately private endeavour with associated privately oriented objectives and interests. Unravelling the accounts offered by Gasser et al, and reading them in the light of this understanding, and the rearticulations of online governance models offered by Klonick and von Loo, thus begins to expose the false reassurance often generated by the label of digital constitutionalism – and the private interests it can obscure.

Private policy in practice

Perhaps the most explicit and ambitious project of digital constitutionalism to date has been evidenced by Meta’s (previously Facebook) repeated attempts to impose a gloss of constitutional governance upon its self-regulatory efforts. Meta’s rebranding of its terms of service as a ‘Statement of Rights and Responsibilities’, ⁷⁷ which referred to the ‘rights’ and ‘freedoms’ of users, was an early indication of this desire to mimic the trappings of constitutionalism. In these terms, Meta was clearly using constitutional language to offer superficial legitimacy to the company’s contractual terms. The Statement of Rights and Responsibilities was, in turn, structured to be drawn from the values embodied in the company’s Facebook Principles, also drafted in the universal tone of Constitutional documents – swapping the individualistic contractual language of ‘you’ and ‘user’ with the more solidarity laden terms of ‘people’ and ‘every person’.

The use of terms such as ‘rights’, ‘responsibilities’ and ‘freedoms’, but also the categorization of contractual provisions and entitlements in the language of constitutional obligation, as ‘principles’ and ‘rules’ premised on ‘values’ drawn from a hierarchical series of sources, allowed Meta to co-opt the gloss of constitutionalism but does not reflect the normative core of constitutionalism nor the relational structure of constitutional projects. It cannot be stated definitely on the basis of publicly available information whether this is the result of an opportunistic attempt to deliberately co-opt the language of constitutionalism in order to obscure the true nature of the power being exercised, or merely an unintentional by-product of the commercial ecosystem that seeks to assuage public scepticism through espousing a commitment to ‘public’ values. What is certain, however, is that the resulting attempt to echo constitutionalism’s mesh of the language of rights and restraint and public law forms within a privatized system of enforcement does not display the substantive characteristics necessary to attach the label ‘constitutionalism’. It does, however, confuse the practical capacity of that term to act as a signifier of certain protections on a popular, and normative, basis.

More recently, Meta has moved away from this linguistic affect and has reasserted its contractual provisions as ‘Terms of Service’. ⁷⁸ However, in its place the company has adopted a new model of ‘constitutionalism’ in the form of the ‘Facebook Oversight Board’. The Board, which was at one point in its development referred to informally as a ‘Supreme Court’ for the company, has progressed Meta’s rhetorical commitment to Constitutional principles further than its previous textual efforts, instituting an internalized appeal system for disputes over freedom of expression on the platform ⁷⁹ and suggesting that, in future, the mechanism might also offer an appeals system for other platforms owned by Meta. ⁸⁰ Comprehensive examinations of the Oversight Board’s governance and operation (as provided for in its Charter) have been undertaken elsewhere. ⁸¹ The significant feature of the board from the perspective of this piece is its attempt to position itself as a constitutionalist mechanism that evinces a commitment to an ambiguously defined Constitutional principle of freedom of expression/free speech while simultaneously declining to be bound by those same principles as they are applied by state actors and democratically sanctioned Constitutions.

Inasmuch as the Oversight Board is committed to values or principles that overlap those provided in Constitutional documents, it is committed only as part of a model that implicitly endorses the idea that constitutionalism can be achieved by private actors, through selective commitments to certain values or principles defined, enforced and overseen in a contractual system. ⁸² Far from evidencing a substantive commitment to the normative core of constitutionalism, the Board’s adoption of constitutional language and a privatized judicial review structure functions as a ‘legal talisman’ ⁸³ that conveniently appropriates signifiers of constitutional allegiance while failing to evidence the substantive features necessary to constitute a constitutionalist model and the values, and objectives that constitutionalism seeks to secure. ⁸⁴

It is a truism within public policy studies that the actor who defines a problem controls not only the perception of its contours but also the design, and thus the manner, of its resolution. ⁸⁵ In this respect, constitutional governance is, as Levinson notes, the project of creating, allocating and constraining power ⁸⁶ – or, as Mashaw puts it, identifying and bringing to bear the law in formulating policy problems and their solutions, and doing so before other actors can offer competing resolutionary claims. ⁸⁷ Where constitutional models of governance are present, in other words, there is necessarily an understanding (indeed, a guarantee) that power is held and distributed in a particular manner. ⁸⁸ Where the label ‘constitutionalism’ is invoked, it is an indication that similar decisions as to how and where power is concentrated – and how it will be used – are also taking place, although not that they are being made within the constraints the label implies. ⁸⁹ Thus, when we look to Meta’s creation of the Oversight Board and its attendant governance architecture, even if it cannot satisfy the requirements of constitutionalism, the Board’s constitutionalist ambitions and the structures and language it has adopted dictate how power is exercised and by whom no less than in a truly constitutionalist model.

The much-touted paradox of constitutionalism is that the restrained central authority that it seeks to secure can, in fact, become more powerful that an unrestrained one. ⁹⁰ This critique holds that the constitutionalist features that restrain power, when viewed in isolation, can also serve to expand power when viewed in broader temporal or topical frames. ⁹¹ In the case of digital constitutionalism, the risk is not only that no effective restrains parallel to those offered by traditional constitutionalism are present but also that the trappings of constitutionalism – which obscure the true justifications and concentrations of power within digital spaces – will permit the same outcome feared by critics of constitutionalism. Adopting the language of constitutionalism permits private actors such as the Oversight Board to determine what practices are permitted to them – just as much as, if not more than, those that are prohibited – and implies that such self-determination (which is, in fact, extra-constitutional) is infused with the legitimacy of a constitutionalist project.

The Oversight Board, for example, is empowered by its Charter to decide whether content posted on Meta’s platforms constitutes of a violation of the company’s terms of service and in light of its attendant contractual and corporate governance standards. ⁹² The Board is also delegated the authority by the Charter to provide general policy guidance on foot of one of these decisions or where Meta requests such guidance. ⁹³ This appears to be a relatively narrow jurisdictional remit. However, the constraints the Charter imposes in fact permit the Board to commit itself to a legitimized process of adjudication of what are, in essence, rights claims that are otherwise beyond their reach – acting to ‘generate credible commitments that induce others to behave’ in the desired way. ⁹⁴ The discrete jurisdiction established by the Charter endows the Board with a policy-making power justified and legitimized by its constitutional appearance and design. Crucially, and in addition, the independence of the Board ⁹⁵ permits both it and Meta to each distance themselves from the activity of the other. Meta can thus disclaim the Board’s decisions and responsibility for them where they are unpopular, while the Board can equally disclaim the activity of Meta that has generated the disputes before it. ⁹⁶ The result is an architecture in which power is delegated to the Board, not with altruistic intent but in a manner that allows both actors to avoid the attachment of a full measure of accountability for the mechanisms and decisions they enforce and consider.

Of course, it may be argued that any central authority may advantageously use constitutionalist characteristics to concentrate their power and minimize or distance accountability. Indeed, there is a basic risk in any context in which a central authority creates standards that bind them that they will advantageously construct their de jure regulation to render the standards susceptible to waiver. The question is often framed in constitutional theory as one of ensuring that these internally defined standards — ‘parchment barriers’ ⁹⁷ in Madison’s telling – can function as effective systems not only of de jure but also de facto restraint. ⁹⁸ In traditional constitutional settings, this is accomplished both through internal institutional mechanisms (such as the separation of powers) and through external pressure inherent in the relational structure of constitutionalism itself.

In a digital context, no such parallel check exists. Meta can, of course, create the Oversight Board, but it might equally choose to ignore the Board’s decisions without breaching anything beyond an agreed contractual standard or a code of corporate governance. Users might, equally, choose to leave a platform or service (as a corollary to a democratic mechanism in traditional constitutional settings). Yet, in practice, it has repeatedly been shown that the possibility of such choice on the part of users is largely theoretical. ⁹⁹ The result is not a transposition of traditional constitutional restraints to digital settings but a false reassurance that this has occurred unsubstantiated by practical mechanisms or guarantees.

It is not that private actors are, in definitional terms, incapable of identifying or securing the kinds of public goods that lie at the heart of constitutionalism and public policy; however, the ‘constitutional’ model presented by the Oversight Board, as well as many of the models to which the label is more generally applied, does not serve this end. The Oversight Board, for example, addresses only those ‘hard cases’ referred to it for consideration ¹⁰⁰ and not only solves these cases on the basis of public good objective (however coherently defined), but directs its efforts towards ameliorating speech conflicts generated by popular criticism and consumer dissatisfaction to maintain financial stability and corporate growth. Inasmuch as private actors are ‘constitutionally’ oriented towards the development and maintenance of a particular set of values, it is those contained in their own corporate frameworks that serve shareholder-driven and defined goods, not public ones.

In this respect, it must be acknowledged that public goods can be defined in numerous ways. From a minimal libertarian or neoliberal perspective, the maximization of market freedoms is sufficient for securing the public good. ¹⁰¹ More expansive understandings of the public good move from ordoliberal views ¹⁰² through to social democratic or social market ¹⁰³ views, which understand the public good as including market freedom but also a supplementary (and perhaps overriding) social character. Even within these ideological groupings, there is debate about what the precise content of the public good is and how it should be achieved. As such, these understandings of the public good are neither uncontested nor stable. However, even from a minimalist neoliberal view in which the public good is achieved through maximal market freedom, there is reason to doubt the capacity of private actors (specifically in the digital context) to secure it. ¹⁰⁴

Even if we accept that the Board enjoys an institutionally independent character, ¹⁰⁵ and that its composition reflects a broad range of views, ¹⁰⁶ it lacks the capacity and jurisdiction to identify and enforce constitutional values in the form of fundamental rights standards in a coherent manner. Moreover, the nature of the contractual relationship between central authorities in the digital environment, and the users subject to the rules they impose (including those that create bodies like the Oversight Board), as well as the lack of structural limitations on power and accountability mechanisms, mean that these efforts cannot be described as constitutionalist. What they represent is an at best fragmentary idea of privatized governance and appeals systems guided by adherence to rights standards drawn in terms that are reconcilable with the commercial objectives of the private actors themselves. This is without an inquiry into the consistency with which these fundamental rights standards are defined or enforced, or their coherence when measured against their equivalent provisions in ‘traditional’ constitutional contexts. In this respect, the activities that are being undertaken by private actors are neither de jure nor de facto constitutive of ‘digital constitutionalism’, but are more appropriately described as activities that form part of a broader pattern of ‘private policy’ formation.

VI. Conclusion

The label of digital constitutionalism can, of course, be both useful and meaningful. For this to be the case, however, we must adopt an analytical lens that inquires beyond its use as a flag of convenience, and engages with its requirements. In the absence of a precise, coherent and consistent understanding of the normative and practical characteristics of digital constitutionalism, its features have become the ideal mechanism for obfuscating the activity of private actors as they concentrate power and conceal the location of institutional accountability. The result is that digital constitutionalism conceals more than it reveals, using a normatively reassuring label to signal an allegiance to uncertain standards.

Thomas Paine, reflecting on the newly drafted American Constitution, remarked that the document was ‘to liberty what grammar is to language’. ¹⁰⁷ Shortly afterwards, Arthur Young, commenting on the French Constitution of 1792 – which had drawn heavily on the North American example – noted that it had been constructed as ‘if a Constitution was a pudding to be made by a recipe’. ¹⁰⁸ The observations of both men are apposite in the case of digital constitutionalism and the struggle to assess both its substantive coherence and capacity to meet the requirements of the label to which it aspires. Too many of the instances in which the label has been used to date, however, fail to situate their description within an understanding of the structural and normative requirements of constitutionalism – treating it as a label that can be justified by grammatical manoeuvring or the addition of only a selection of ingredients. In doing so, these uses of the label ‘constitutionalism’ fall victim to the trap identified by Waldron: deploying constitutionalism as an empty slogan rather than a substantive concept. More fundamentally, the result is that these efforts are less a matter of constitutionalism and can more accurately be described by reference to, and understood through the lens of, private policy. This can expose not only the true character of the governance structures in place within these digital spaces, but also what constitutional values and principles are – and are not – protected within them.