Rejecting Thin Constitutionalism

Guest Blogger

For the Balkinization symposium on Mark Tushnet and Bojan Bugaric, Power to the People: Constitutionalism in the Age of Populism (Oxford University Press 2021).

Tarun Khaitan 

The central claim that Professors Mark Tushnet and Bojan Bugaric make in their provocative new book ‘Power to the People’ is that populism is only contingently in conflict with constitutionalism. Much of their argument will turn on what populism and constitutionalism mean. Taming either of these two slippery concepts is ambitious enough, attempting to do so for both simultaneously would significantly advance current discourse in constitutional studies. This is likely to be the case even if they get it wrong: one learns a lot even from the mistakes of clever and insightful scholars. The first step in their carefully argued book is a description and defence of what they characterise as ‘thin constitutionalism’. It is this first step that I will subject to close inquiry in this post. 

What is ‘thin constitutionalism’ (hereinafter, ‘TC’)? The authors give a “least-common-denominator definition: its components are elements that nearly everyone committed to a normative idea of constitutionalism would agree on, and it doesn’t include elements that some people sincerely committed to such an idea wouldn’t think essential to the project of constitutionalism” (2), namely majority rule, constitutional entrenchment, judicial independence, and politicians and political parties (12).

The Normativity of Thin Constitutionalism 

The authors’ understanding of constitutionalism is normative, not (merely) conceptual inasmuch as it is (for them) a benchmark for what institutional arrangements are acceptable, rather than merely a descriptive concept or a pedagogic tool. The authors are not entirely clear on the normativity of TC though. On occasion, for example, they claim that “it is mostly a heuristic device” (11). Even so, I think the best reading of TC is indeed normative. After all, they define it as made of elements that ‘nearly everyone committed to a normative idea of constitutionalism would agree on’ (2, emphasis added). Further support to the proposition that the authors view TC as a (normative) guide to constitutional design is found in this claim: “you say that constitutionalism requires a specific thing, but I can point you to constitutional systems that seem to me to do just fine without it” (11). Presumably, if no such system could be pointed to, in the authors’ view, constitutionalism would indeed require that specific thing. 

Indeed, sometimes it seems that they view constitutionalism not just as a normative benchmark, but as the exclusive normative standard in relation to a state’s institutional setup: for the authors, asking whether something “must be permitted” is the same question as asking whether “constitutionalism requires that they be permitted” (10). In other words, the only relevant question one ought to ask about the (moral) permissibility of an institutional setup is whether it conforms to constitutionalism. This is a logical fallacy of the following sort:

All good persons are truthful.

Nikolai is truthful.

Therefore, Nikolai is a good person. 

For all we know, Nikolai could well be a truthful bully. Personal as well as institutional characters admit to value plurality: their goodness or badness cannot be judged along just one or two normative axes. Indeed, insisting that Nikolai must at least be good qua his truthfulness is dangerous, for it could be used to mask or paper over a serious flaw (bullying). Goodness of persons and institutions must be an all-things-considered judgment. 

Perhaps this criticism mischaracterises the normative essence of constitutionalism (as the authors see it). It may be that they think of constitutionalism as akin not to singular personal/institutional virtues (like truthfulness or efficiency), but rather a more general appraisal of the overall character of an institutional setup. Thus, a constitutionalist regime may itself be akin to a good person rather than an truthful person. But then, the concept of constitutionalism does no work, or—at least—no more work than what ’good’, ’desirable’, etc could do without invoking a vague, heavy-duty, -ism. 

Unfortunately, the authors don’t tell us about what kind of normative benchmark their conception of constitutionalism really is. I suspect the most promising clue lies in their frequently juxtaposing it with ‘politics’: “The idea of thin constitutionalism brings out the essentially political—rather than constitutional—nature of objections to specific populisms” (11, emphasis added). It seems that whatever constitutionalism requires, it is something other than ‘political’ (so, … judicial?). Further confusion is caused by their use of the adjectival form of the term ‘constitution’ (constitutional) rather than that of ‘constitutionalism’ (constitutionalist)  in this quote. Surely, an arrangement can be constitutional (in the context of a given constitution) without conforming to constitutionalism, and vice versa. If the substitution is telling, have the authors allowed their (US-inflected) equation of constitutionality-determination with judicialization to infect their understanding of what is at stake when we ask ‘Is this arrangement in conformity with constitutionalism?’ Perhaps they think the attractions of TC lie in carving out a narrow institutional role for courts. If so, TC would have little attraction for anyone who denies—as I do—any necessary connection between judicialization and the determination of what is constitutional (or, constitutionalist, for that matter). 

To sum up my first criticism, it seems that while the authors intend TC to work as a normative benchmark for constitutional systems, and want it to be a relatively undemanding benchmark to leave sufficient space for ‘politics’, they do not tell us much about its point or purpose. The authors do raise the issue of the point of constitutionalism briefly (22-23), only to dismiss a possible candidate (i.e. to limit governmental power). They offer no positive account of what the raison d’être of constitutionalism is. To put the point differently, we cannot know the content and scope of a normative phenomenon like constitutionalism without knowing what constitutionalism—thick or thin—is for. 

Reasonable Disagreement and the Least Common Denominator 

Instead, they prefer to give content to TC using what they describe as the least-common-denominator approach by stipulation: basically, if there is ’reasonable disagreement’ on whether a specific feature is or is not part of constitutionalism, it is automatically ruled out as a part of the concept. What then is the case for using the ‘reasonable disagreement’ approach to determine the content of a given controversial concept? Again, the authors do not seem to offer a defence for preferring this approach, except to say that “Whatever else you think constitutionalism requires, you’ll agree that it requires at least thin constitutionalism” (11). That may well be true—if the ‘reasonable disagreement’ test is indeed applied correctly—but all that can be discovered through this approach is some sort of conceptual core of the phenomenon under study. Chris McCrudden, for example, used something like this approach to determine the minimum core of dignity. While such a semantic least-common-denominator approach can help uncover the conceptual core of a phenomenon, and therefore aid understanding, it is incapable of determining its normative content and scope. 

To see why, consider that we are interested in determining what counts as ’murder’ (in the moral, rather than the legal, sense of blameworthy killing). As a moral concept, murder is an evaluative concept that is a normative benchmark for determining what types of killing another may be acceptable. Folks disagree, dare I say reasonably, about whether the following acts count as murder:

(i) killing A to protect the lives of B, C, & D (assuming all parties are innocent),

(ii) killing A upon her request to end her suffering (especially when A is not terminally ill),

(iii) doing something that one knows will result in A’s death, but intending to do something morally permissible (such as save one’s own life, avoiding a greater evil etc) and wishing there was a way to achieve this other objective without resulting in A’s death. 

Most of us will agree that none of these three scenarios constitute the conceptual core of what ‘murder’ is. But, that does not mean that the mere fact of reasonable disagreement on these three counts should—without more—exclude them from the normative scope of what amounts to murder. Countless similar examples can be given with respect to other normative phenomena, including the status of ‘indirect discrimination’ (disparate impact) as ’discrimination’, ‘consent obtained by lying’ as ’rape’, ‘hate speech’ as ‘protected free speech’, and so on. The least-common-denominator approach, when applied to any of these normative concepts, will inevitably result in permissive, libertarian, outcomes. The authors know this, for they “note one problem associated with thin accounts. … thick accounts make it too easy to find fault with actual practices. Thin accounts, though, may make it too difficult to do so” (11). One would have expected them, having noted this shortcoming, to defend their use of the approach. Unfortunately, no such defence is provided. 

To avoid misunderstanding, let me clarify that reasonable disagreement can sometimes give us reasons to act (or refrain from acting). In the context of deeply divided societies, they may give pragmatic reasons to refrain from deciding either way,^? and in democratic contexts give institutional reasons to allocate the relevant decision-making power to legislatures rather than to judges.^? The authors do not tell us that either of these reasons apply in the context of constitutionalism. So, this is the second criticism: either TC has no normative value, or its scope and content need determination by substantive arguments rather than the reasonable-disagreement method. 

Is Thin Constitutionalism Really Thin? 

My third, and final, criticism in this blog post is that the authors apply their least-common-denominator method incorrectly. There are four elements that the authors identify as constituting TC’s content: (i) majority rule, (ii) constitutional entrenchment, (iii) judicial independence, and (iv) politicians and political parties (12). Each of these elements are stated at a very general level, to allow very different ways of approximating to these requirements. But if the test of a feature being admitted to TC’s scope is an inability to “point … to constitutional systems that seem … to do just fine without it” (11), it is not clear that all of these items pass muster. Take majority rule for example: even if one considers a disagreement based on China’s remarkable ability to lift millions out of poverty as nonetheless below the threshold of ‘reasonableness’, it is harder to say that those who disagree with majority rule in the context deeply divided societies are also being ‘unreasonable’. Constitutional theory and practice have recognised consociationalism for too long and too clearly to dismiss it as unreasonable disagreement with the ideal of majority rule. Consociationalism is not some imperfect mechanism for approximating to majority rule; at least certain forms of consociationalism are designed to ensure that ’the majority’ does not rule. The same can be said of constitutional entrenchment: the authors insist that TC requires that there be at least some constitutional provisions that are entrenched against ordinary majorities (19). This would exclude constitutions like that of the United Kingdom, which rely on constitutional culture, rather than procedural entrenchment, to protect its fundamental norms. Whatever its merit, is the British position unreasonable? The argument from reasonable disagreement proves too much. In fact, if properly applied, it may well result in a null—rather than thin—content for constitutionalism. 

These problems arise because the authors stipulate but do not tell us why these four, and only these four, items qualify as contents of TC. The list is presented as so self-evident that not much is offered by way of justification. Nor do the authors say very much about the methodology employed by them to discern what disagreements exist, and how they are determined to be ‘reasonable’. Is it disagreement between framers of different constitutions? Or opinions of constitutional scholars? An overwhelming majority of actual illustrations—both positive and negative—of constitutional phenomena in the relevant section (titled ‘The Elements of Thin Constitutionalism’, pp 12-32) are drawn from the United States (pp 14, 18, 19, 20, 24, 27, 28, 29, 30, ). This is somewhat ironic, as the authors (correctly) warn against a ‘selection effect’ because ‘the conclusions you draw about populism and constitutionalism depend upon the populisms you include in your study’ in the introductory chapter (4)—surely this is true not only of populism but also of constitutionalism. The outsized role that the concept of ‘reasonable disagreement’ has played in American constitutional thought and political life—thanks perhaps to John Rawls—is well-known. TC applies this very-American methodological legacy beyond its Rawlsian usage in the context of disagreements over comprehensive conceptions of a good life to disagreements over institutional forms. It also employs a parochially American look-and-feel lens to determine what is ‘reasonable’.

 In conclusion,

There is something attractive about using the lens of ‘disagreement’ to think about constitutionalism. After all, it is a plausible normative claim that constitutions ought to be framed with the general agreement of all, or almost all, sections of the polity. But accepting this normative proposition entails neither the predictive claim—that people are most likely to agree to very thin constitutions only containing items they agree upon—nor the normative claim—that this is how constitutions should be designed. The predictive claim is belied by the evidence of remarkable thickness of constitutions in the postcolonial world, as well as case studies that show that folks often prefer constitutions that include their thick claims (including those they disagree with) as a package deal over those that include none.^? As a normative claim, TC may be worse than not employing constitutionalism at all: just as Trump employed his supposed ‘authenticity’ to paper over his numerous personal failings to gain the confidence of the American voters, the thoroughly undemanding normative standard of TC may well help legitimise decidedly bad regimes. 

Tarunabh Khaitan is the Head of Research, Bonavero Institute of Human Rights (Oxford), Professor of Public Law and Legal Theory, Faculty of Law (Oxford), Honorary Professorial Fellow (Melbourne Law School), and the Walter V. Schaefer Visiting Professor of Law (Chicago Law School). He is grateful to Prof Nick Barber for commenting on a draft of this post. Email: tarunabh.khaitan@law.ox.ac.uk.