1 According to one prominent view, lawyers’ comparative advantage over political theorists (shared with some political scientists) is found in their understanding of institutional design: how the formal rules defining institutions affect how they operate, how particular institutions can be combined into an overall institutional system that achieves desired goals more effectively than any single institution can, and – to some extent – how considerations of power affect how institutions operate and interact. This expertise in institutional design contrasts with expertise in conceptual analysis.
- 1 The “thin” (that is, undemanding) definitions of liberal and popular constitutionalism that animate (...)
- 2 None of the institutions are essential to implementing popular constitutionalism, but their use is (...)
- 3 Specifically, representatives will not be confined to merely “working out the technical details” of (...)
- 4 For a defense of that proposition, see Tushnet & Bugaric (2021: ch. 12) (which also develops other (...)
2 For that reason this Essay is not a defense of popular constitutionalism as a version of or alternative to liberal constitutionalism.1 It takes as a given the proposition that some version(s) of popular constitutionalism are compatible with liberal constitutionalism, and discusses how several institutions might be designed to implement popular constitutionalism within a liberal constitutionalism frame. The institutions are (1) forms of direct popular legislation such as referendums, (2) imperative mandates or instructions to representatives that the representatives must follow, sanctioned by automatically removing a noncompliant representative from office, and (3) modern communications technologies used to elicit citizen views as an alternative to voting (or polling).2 I assume that popular constitutionalism does not require direct popular participation in all forms of law-making including directly making fully enforceable laws, but allows for the possibility that representative government (of a deliberative rather than merely technocratic sort) will have a quite extensive domain.3 Implicit in these preliminary comments is the proposition that not all proponents of popular constitutionalism (or, in some versions, not all populists) are anti-institutional in principle.4
- 5 I use the term “referendums” to refer to all forms of direct legislation by the people. In U.S. law (...)
- 6 An additional feature – sometimes amounting to a “concern” – of referendums is that they can be the (...)
3Legislation by the people themselves may be the cleanest example of popular constitutionalism. Scholarship on referendums5 worries, though, about whether – or rather about the extent to which – direct popular legislation is compatible with liberal constitutionalism. I consider here what I believe to be the two most prominent concerns: (1) that referendums can oversimplify complex policy options in ways that sometimes produce outcomes that are indefensible in principle, incoherent, and inconsistent with what the people would prefer after the kind of deliberation that occurs in representative assemblies, and (2) that referendums systematically, though not inevitably, threaten rights of minorities that liberal constitutionalism guarantees.6
- 7 For an overview of the problems associated with complexity, see Morel (2018: 149). Constitutions th (...)
4The Brexit referendum illustrates many aspects of the first concern. The implications of one of the options – Exit – were unclear because Exit could occur in many ways. We can call this a problem of complexity: The people were asked to choose between a simple option and a complex one.7 Even a referendum cast in “yes-no” terms can be complex. Consider for example a referendum on this question: “Should the following complicated tax law be adopted?” Complexity cannot be eliminated by offering the people several options: “Remain,” “Exit on the following terms,” and “Exit on these alternative terms,” or the like. Formulating a complete set of options may be difficult or impossible. Even more, a referendum with more than two options makes the appearance of a Condorcet voting paradox possible, even likely if (as seems likely) the complexity is not characterized by those features of policy choices that make preference rankings single-peaked.
- 8 There is another sense of complexity that I explore below – issues that are complex because any sim (...)
5The idea of a voting paradox brings out one key problem with referendums on complex issues. Begin with the observation that policy proposals that are complex in this sense have numerous components. A person might prefer component A to component B only if the overall policy also contains component M. Some such preferences might be non-negotiable, so to speak. The “only if” is quite strong. Other such preferences might be weak. The voter’s view is that having components A and M is desirable, but a policy that contains component A but not M is tolerable if the policy contains a substantial number of other components unrelated to A and M (and similarly for other people and other components).8
6This phenomenon can occur with respect to ordinary legislation considered by a representative assembly. Sometimes it is said that a referendum presents people with a choice that cannot be amended or tinkered with to come up with a better option. In contrast, it is said, legislators can negotiate until they have before them a complex proposal that has a chance of being satisfactory to a majority. That contrast is overdrawn. Referendum proposers have an interest in allowing discussion and modifications that increase the proposal’s chance of adoption once submitted. A representative can refuse to vote in favor a proposal that fails to satisfy a non-negotiable requirement; a voter can vote “No” on any proposal that similarly fails to satisfy her non-negotiable requirements. Representatives can make judgments that the proposal before them, while not perfect, is on balance better than the status quo; so can voters.
- 9 Note that this solution is unavailable when the proposal fails because it contains some component o (...)
- 10 Note that this is different from the question of whether there should be in principle some limit on (...)
7But, one might say, legislators have an opportunity to take successive votes. If a complex proposal is defeated on the floor because it is not on balance acceptable to a majority, representatives can retreat to the committee room and modify the proposal, then present the modification to see if it is acceptable – and they can do so repeatedly.9 Referendums are one-shot events, with no opportunity for a do-over. This is not a necessary characteristic of referendums. In principle the people could be asked to vote repeatedly on variants of rejected proposals.10 In practice, though, successive referendums on variants are unlikely.
- 11 Experience in the United States with rules that limit legislation or referendums to a “single subje (...)
8This counsels in favor of imposing some conditions on referendums. Referendums on complex issues might be allowed only if the proposal put to the voters makes its components clear: When the ballot proposal says “Exit,” it means “Exit on the following terms.” In institutional terms, we would have to authorize someone to determine that a proposed referendum satisfied these or other requirements. In constitutional systems with a tradition of constitutional review or with a tradition of robust ultra vires review, the solution is simple: The courts, whether constitutional or administrative, decide whether a proposal satisfies the requirements of simplicity or clarity.11
- 12 Because referendum-adopted policies are not different in principle from representative-adopted legi (...)
9A clear policy can be a bad one, of course, just as legislation can be bad. The solution to a bad statute is to repeal it. The solution to a referendum that generates bad policy is the same. We can allow the legislature to treat a policy adopted through referendum as it treats any other piece of legislation, subject to immediate revision or even repeal. Or, we can allow the legislature to modify a policy adopted through a referendum only after a prescribed period such as two years. Or, we can allow modifications of referendum-adopted policy only by a second referendum.12
- 13 A person might reject a job offer that has a greater long-term payoff than does her current job but (...)
- 14 Here too referendum-adopted legislation is not all that different from representative-adopted legis (...)
10Complex policies, and some simple ones, have other characteristics. They have to be integrated into the existing corpus juris, and they might fit badly or cause unanticipated consequences. A distinctively U.S. example is this: A referendum strictly limiting a legislature’s ability to modify the existing property tax system may lead legislatures to increase sales taxes in ways that have troubling distributional effects, and it almost certainly will cause people to stay in the homes they own longer than is good for them.13 Again in principle there is a simple solution to this problem when it arises. A referendum-adopted law directing the adoption of a policy with these characteristics should impose an obligation on the legislature to adapt the corpus juris to this new feature.14 And, if and when unanticipated consequences appear, the overall policy-making system should be allowed to address those consequences as described above, by amending or repealing the enacted piece or legislation.
- 15 Two semi-technical notes: (1) Some referendum systems in the United States allow the presentation o (...)
11Suppose we do say that referendums should be allowed only if they offer a choice between two clear policies. Where a complex policy is placed before the public, the alternative should be “No” rather than another complex policy, because “forcing” a choice between two complex policies does not allow a voter to express a preference for the status quo or for another complex policy.15
12We might be concerned that clarity as a solution to the problem of complexity would create a different problem. In politics clear and accurate policy descriptions might be misleading. This might be particularly true of complex policies presented to the public as a list of clear components. Ordinary people might focus on one or two of the components on the list and not appreciate how those components might interact with other components in ways that alleviate (or exacerbate) the voters’ concerns. Even more, political leaders might make one or two components the focus of their campaigns for or against the referendum, again with misleading effect.
13One can imagine institutional responses to concern about misleading clarity. In the United States presumptively disinterested officials are sometimes charged with developing neutral and accurate descriptions of ballot propositions that are distributed to the public in advance (and that sometimes appear on the ballot as well). The British Electoral Commission did force a change in the Brexit referendum’s wording to make it more neutral and clear, though of course the outcome suggests that attempting to purify the questions put to the voters might fail.
14More significant, though, the concern about misleading clarity rests on assumptions about voter incompetence. And popular constitutionalism rests on the contrary assumption, that ordinary people are generally competent (or, perhaps more accurately, are at least as competent as the representatives they elect) in making political decisions. Perhaps critics of referendums could develop more precise accounts of specific forms of voter competence that would be consistent with the assumptions of popular constitutionalism, accounts that would allow some but not all referendums on complex policy questions.
- 16 The example is adapted from a referendum in Switzerland. A similar example can be found in a 2015 G (...)
15Should simplicity be another criterion for referendums? Consider a referendum on the question, “Should the legislature guarantee a minimum income of at least X – adjusted for inflation – to every person residing in the nation?”16 This proposal is clear. It is not simple, though. It implicates complicated policy questions: How much less (if at all) will people work if guaranteed a minimum income? What will the effects on economic growth be? It also implicates moral questions: What duties do we have to non-citizen residents, whether long-term or otherwise?
- 17 This formulation assumes that interests are categorically less important than rights, though not th (...)
16The same concerns about voter competence arise in connection with the policy dimension of this and similar proposals. The moral dimension is sometimes thought to raise different concerns. The problem is not that some moral questions are complicated. Rather, the problem is something akin to self-interest: Voters asked about the rights of non-citizens are, it is suggested, likely to undervalue those rights relative to their own rights or, more important, their own interests.17
- 18 For an overview of the problems associated with rights-related referendums, see Fatin-Rouge Stefani (...)
- 19 Matsusaka 2020. Drawing on a comprehensive data set of referendums in U.S. States, Matsusaka (2020: (...)
17This concern is brought home most clearly in suggestions that it is inappropriate to subject questions about individual rights to referendums.18 John G. Matsusaka’s conclusion from his review of referendums in the United States is that “[i]nitiatives do pose a threat to minority rights, but … the threat is not immense.”19 The most common example from recent history is the Swiss referendum approving a limitation on the construction of minarets. Yet, we should keep in mind the equally recent Irish referendums on marriage equality and abortion (and the Australian quasi-referendum on marriage equality) in thinking about the “no referendums on individual rights” proposition.
- 20 The issues taken up next raise difficult and deep questions about the proper analytic structure for (...)
18These examples suggest that we could benefit from a more differentiated analysis. I make a stab at providing one in what follows.20
- 21 I have thought about this distinction more in connection with freedom of expression, but I am reaso (...)
- 22 For accuracy, I add that both examples involve presumptively impermissible government actions but t (...)
19First, I have found it useful to distinguish between what I call the core of an individual right and that right’s specifications.21 The core consists of the general and typically quite abstract statement of the right – “freedom of speech” or “freedom of religion” – coupled with uncontroversial examples (paradigms, we might call them) of the right’s violation. So, for example, the core of freedom of speech uses the paradigmatic example of seditious speech – speech critical of government policies that is said to undermine public confidence in the government and, for that reason, to impair the effective implementation of the criticized policies. The core of freedom of religion uses the paradigmatic example of a government policy prohibiting religious believers from engaging in practices they regard as central to their religious commitments.22 Specifications arise when government policies trigger concern that the right might be violated in circumstances not covered by the core.
20I have two thoughts about the implications of the core-specification distinction. First, the government policies associated with specifications typically serve non-trivial public interests in a not entirely unreasonable way. Consider restrictions on tobacco advertising: In the language that prevails in comparative constitutional discourse, such restrictions impair the advertisers’ free speech interests in the service of important public health goals. One might reasonably believe that specifying the content of a right of free expression so as to allow such restrictions is reasonable – though of course one could reasonably take the converse view.
- 23 The simple answer – bias against historically despised minorities – is placed in question by the Ir (...)
21In these and similar cases, we can see how reasonable people could disagree about whether the government policy does indeed violate the right. It is not clear to me that ordinary people are incompetent or likely to be biased in evaluating whether a referendum threatens rights in a case of specification. Or, more narrowly, perhaps we can make progress in figuring out the difference between the Swiss minaret referendum and the Irish referendums by thinking about why ordinary people might have been biased in the first but not the second and third cases.23
22Second, I speculate that violations in the core result at least as much from self-interested decisions by political leaders – from representative government – as from self-interest among the general population. This is clearest with respect to sedition law: Political leaders who develop policy are likely to be especially sensitive about criticisms of their policies, perhaps more so than the voters they represent. For that reason they may be more likely to adopt a sedition law than would the people through a referendum. Similarly, John Hart Ely’s account of the proper bases for constitutional review refers to the fact that the “Ins” are sensitive to challenges from the “Outs.” And here he rather clearly had legislators in mind, not the voters they represent.
- 24 Will Partlett’s (2012) work on president-led referendums in central Asian post-Soviet nations is pa (...)
23This second thought might also lead to a more refined account of the circumstances under which referendums implicating questions of individual rights might be appropriate – roughly, in cases involving specifications rather than core violations. I note one difficulty, though: Political leaders might see political advantage in promoting a referendum on a specification even if ordinary people on their own would not put the referendum process in motion.24 This might suggest some sort of institutional response, through the development of some limits on the role political leaders can play in referendum campaigns. What such limits (consistent with ideas about freedom of expression) could be is unclear to me, though.
24As noted in the Introduction, popular constitutionalism does not require the comprehensive substitution of policy-making by the people directly for policy-making by representative assemblies. Put more directly: Referendums will deal with only some areas of public policy. In my view we are unlikely to develop substantive criteria for identifying those areas. Rather, the domain of referendums will be determined by the people’s views about which policies are important enough to be put to the people (and perhaps secondarily, the people’s views about which of those policies is tractable to adoption by referendum).
- 25 More precisely, such mandates are enforced by parliamentary institutions whose decisions either are (...)
25Some constitutions prohibit imperative mandates or binding instructions from voters to the representatives they elect. Such mandates, in their strongest form, are self-enforcing: A representative who votes against instruction loses her seat in the legislature.25
- 26 A line of decisions by the Supreme Court of Israel holds that coalition agreements are enforceable (...)
26Imperative mandates come in two forms. One is connected directly to ideas of popular constitutionalism. Constituents instruct their representatives to cast specific votes on identified policies. Call this a candidate-focused imperative mandate. Importantly, such mandates are binding (enforceable institutionally) while (ordinarily) party platforms are not: Party platforms typically identify policies that party leaders pledge to do their best to implement but departures from a platform by party leaders and as a result by party members are frequent and are subject only to sanctions by voters in succeeding elections.26
27The other form of the imperative mandate is less directly connected to popular constitutionalism. In this variant political parties develop platforms, candidates who run on the party platform promise to vote only as the platform requires, those who are elected are expelled from the party if they break that pledge and lose their seats. Voters support parties, which gives party leaders a base in the people that popular constitutionalism recognizes.
28The most common form of this party-focused variant prohibits “aisle crossing,” that is, the defection of as candidate from the party of which she was a member when elected, to some other party. Bans on aisle crossing raises complex questions about government stability, particularly in systems with proportional representation. A ban on aisle crossing, or more broadly imperative mandates tied to party platforms in toto, is generally thought to give party leaders too much power. I am generally persuaded by the arguments against party-focused imperative mandates. Those arguments are not as powerful against candidate-focused mandates, and the tighter connection between such mandates and popular constitutionalism justifies retaining them as part of the tool-kit of popular constitutionalism.
- 27 For an overview, see European Commission on Democracy Through Law (Venice Commission) (2009).
- 28 In addition, depending on how the rule allowing instructions is interpreted, the representative mig (...)
- 29 The concern about deliberation is sometimes phrased as a requirement that legislators represent the (...)
29Critics offer several reasons that even candidate-focused imperative mandates are undesirable.27 First, they are in principle inconsistent with the proposition that legislatures are deliberative bodies because they bar a representative from deliberative exchanges with respect to the instruction’s subject. Not only is the representative barred from entertaining compromises,28 the representative can respond to reasoned arguments against the policy at issue with the unreasoned assertion that her constituents dictate her vote.29
- 30 A related concern is that a representative under instructions might refuse to vote for a different (...)
30Second, critics argue that imperative mandates are undesirable on policy grounds. They prevent the legislators subject to them from bargaining over the issue, even when the legislator’s constituents would in fact be better off if – in exchange for abandoning the required position – the legislator secured some entirely independent policy (giving up a position on legalization of marijuana, say, in exchange for a large program of grants to support drug-treatment programs).30
31Many of the concerns about imperative mandates track those about referendums. They limit deliberation within the legislature rather than in the public discussion of a proposed referendum once it is placed before the people. They preclude or at least impair the adoption of more effective or more politically acceptable alternatives to the mandated policy. They can offer simple solutions to complex policies.
32Candidate-focused imperative mandates are a partial substitute for referendums. A referendum generates enforceable policy for the relevant jurisdiction when it obtains the required vote in the jurisdiction as a whole. In contrast, imperative mandates and instructions generate single legislative votes for policies, which become enforceable law only when a majority of representatives are bound by the mandate or when enough uninstructed representatives join those bound by the mandate.
- 31 I thank an anonymous reviewer for identifying the possible significance of the temporal gap.
33The first form of imperative mandate differs a bit from referendums: it is adopted at one time, enforced at another, and circumstances might have changed between those times. How significant that difference is in light of the fact that referendums are proposed and often voted on well before they are implemented (as was true of the Brexit referendum), is unclear to me.31
- 32 The possibility of non-electoral consultation is raised in Rawls (1993).
- 33 The qualification “generally” is important, because some liberal democracies will surely adopt poli (...)
34Referendums produce enforceable laws that reflect popular preferences. They identify those preferences with a binding vote. Other ways of identifying preferences provide valuable information without resulting in enforceable law. I have described elsewhere two requirements of what I call “modest constitutionalism”: that policy be made through generally reliable methods of eliciting consent from citizens, either through reasonably free and fair elections or through other methods of consultation,32 and that policy generally be responsive to the preferences of a wide range of citizens.33
- 34 Levy (2018: 339), provides an overview of mini-publics, citizen assemblies, and similar mechanisms. (...)
35Over the past generation scholars and practitioners of institutional-design have explored mechanisms for identifying popular preferences consistent with a popular constitutionalist version of these requirements of modest constitutionalism. These mechanisms are good candidates for inclusion in the set of institutions for implementing popular constitutionalism. I describe a few here and then identify some problems with them.34
36Political scientist James Fishkin developed a technology he calls deliberative polling. In ordinary polling the pollster asks a large number of randomly selected people a series of relatively short questions, sometimes preceded by a tiny bit of information to set the context. A typical question might be, “Do you think President Trump’s policies are helping the nation's economy, hurting the nation's economy, or aren't making much of a difference?” Answering such questions does not take much time.
- 35 There is some evidence, though, that these deliberations can sometimes push people to polarized pos (...)
37 Deliberative polling brings together a smaller group of randomly selected people for a longer period. A typical deliberative poll covers only a handful of issues, sometimes even just one issue. Deliberative pollsters give the respondents a packet of materials developed by experts describing the issues in some detail from a variety of political and policy perspectives. Then the respondents sit down to talk about the issues. Experience with deliberative polling strongly suggests that people who start out disagreeing with each other can hash out their differences and end up generally agreeing on how to deal with the policy questions they’ve been asked to consider.35
- 36 See Center for Deliberative Polling (n.d.) . It may be worth noting that juries consist of randomly (...)
38 Deliberative polling has been used outside the United States to generate proposals on constitutional matters: Korean unification, schooling in Northern Ireland, whether Australia should become a republic. Perhaps because of the scope of the issues such polls have dealt with, their immediate impact on policy has been relatively small, though the results may have had some modest effect in shaping public debates.36
- 37 Neblo, Easterling & Lazar (2018), describe these town halls.
39Advised by some political scientists, a few members of Congress have conducted “internet town meetings.” Instead of physically going to some city council room, the congressmember invites a randomly selected group of constituents to a virtual meeting over the internet. The format appears to foster substantially more thoughtful discussions of contentious issues than in-person town meetings do. Many more constituents can participate, since they do not have to travel to a town hall’s physical location, and the congressmember can hold many more internet town meetings than physical ones because the logistics are much simpler.37
40
- 38 The literature on participatory budgeting is extensive. For discussions with references to the lite (...)
- 39 For a description of participatory budgeting, see Baiocchi (2003).
41Participatory budgeting begins with neighborhood assemblies, usually of volunteers (but other modes of choosing members of these assemblies exist).38 Each assembly discusses the members’ priorities for spending their portion of the city’s budget and can recommend not only priorities but amounts to be spent. The recommendations from the neighborhood assemblies are sent forward to a city-wide citizen assembly, composed similarly. That assembly takes the recommendations and determines priorities and budget allocations for each neighborhood and for the city as a whole. Typically these last determinations are simply advisory to the city’s governing authorities, although in principle they could be binding, at least if the assemblies also could determine where the revenue to support spending programs would come from.39 The model of participatory budgeting can be extended to regional and even national budgets.
- 40 My presentation relies on Reuchamps (2020).
- 41 The citizens are selected in two stages: a group selected at random from the entire population is i (...)
42In late 2019 the parliament of the Brussels region in Belgium adopted regulations embedding a combination of deliberative polling and participatory budgeting in its structure.40 The regulations authorize the creation of a “parliamentary committee” consisting of fifteen members of parliament and 45 randomly selected citizens to develop policy proposals on specified topics.41 The committees are to meet for four days of hearings and deliberations, on topics proposed by citizen petitions. The parliamentarians and the citizen-members can generate separate policy proposals, and if so both are voted upon separately – the citizen proposals in secret, the parliamentary ones in public. Proposals that receive majority support are reported to the parliamentary members, who have six months to produce a report describing what has been done with the recommendations, offering “detailed reasons” for its actions. Among the possibilities, of course, is rejection by the Parliament as a whole.
- 42 For descriptions of the Icelandic process, see Ólafsson (n/d) and Gylfason (2018).
43For about a decade, a “crowd-sourced” constitution has been on Iceland’s political agenda. Outraged at the political elite’s failures, which culminated in a disastrous financial crisis in 2008, Icelanders developed an innovative process for drafting a new constitution.42
44 The first step was an assembly convened by a nongovernmental organization. The assembly brought together 1,200 people chosen at random from the national census list and 300 representatives of Icelandic companies and other groups, to discuss national problems. Ultimately the assembly recommended adoption of a new constitution. The national legislature then set up elections for a “constitutional assembly.” Anyone could nominate himself or herself to sit in the assembly, but people currently serving as political party officers or holding political office were disqualified from serving. More than five hundred people ran for the twenty-five seats.
45 The constitutional assembly operated as a standard constitutional convention, except that it was internet-accessible. Everything it did was almost immediately available on the internet, and it accepted suggestions from the public for constitutional provisions – the crowd-sourcing part of its design. Some of these suggestions were of course lunatic, and some bolstered ideas that were already on the agenda, but the assembly took some crowd-sourced suggestions seriously.
- 43 Ireland has used a similar process, though one slightly more controlled by the government in place, (...)
46 The public endorsed the constitution the assembly came up with, in a referendum that was technically only advisory to the legislature. The legislature in turn stifled the proposal, partly because Iceland had recovered well from the financial crisis and partly because the legislature’s political parties had opposed rewriting the constitution from the beginning. Elections in 2017 revived the idea of adopting the new constitution, because of the new Pirate Party’s electoral success. Though that party was ultimately left out of the coalition government, its advocacy of constitutional reform kept the issue alive.43
- 44 For a more detailed discussion of drawbacks associated with participatory institutions, see Tushnet (...)
- 45 I note, though, that some versions of participatory budgeting have been implemented in cities and p (...)
47 As solutions for our constitutional ills, all of these techniques have obvious drawbacks.44 People have questioned, for example, the neutrality of the expert briefing material presented in deliberative polling, and the possibility of getting a truly representative group of people to take the time to participate. The Icelandic constitutional process generated a document that, according to some constitutional experts, had some technical flaws. It has not yet been concluded because, by design, it bypassed the political parties whose support was important in moving the project forward. Notably, though, a new party saw political advantage in backing the proposal, and the new party’s success has kept the project alive. No matter what, though, Iceland is an unusual case because the nation’s population is under 350,000.45
48 I address in a bit more detail one objection to these techniques. Almost by definition they do not give special weight to those thought to be most knowledgeable about specific matters. It might be thought undesirable for the institution that makes binding policy on matters requiring special knowledge to identify and elaborate good policy to lack guarantees of access to such knowledge.
- 46 A common example here is the implementation of military strategy: Once the people have chosen the g (...)
49Popular constitutionalists have several plausible responses to this objection. First, it overestimates the degree to which specialists actually have specialized knowledge and underestimates the degree to which such knowledge is available within a population of ordinary people. Second, it underestimates the degree to which the results counseled by specialized knowledge should often be modified in light of concerns that ordinary people can bring to the table at least as effectively as specialists can. Third, sometimes domains in which specialized knowledge really is required can be identified in advance and exempted from these mechanisms.46
50 Perhaps most important, though, the objection from specialized knowledge is in deep tension with the premises of popular constitutionalism and so with the enterprise of describing institutions that can implement popular constitutionalism.
51Larry Kramer’s discussion of popular constitutionalism in the United States stresses the role of what was known as “the people out of doors” – groups of people demonstrating (as “mobs”) in support of their accounts of constitutional meaning, coordinating their actions with party leaders.47 The work of Reva Siegel and Robert Post stresses the role of social movements in generating constitutional meaning.
52Neither of these important accounts of popular constitutionalism in action deals with other institutions – relatively persistent modes of cooperative behavior – that can be the vehicle for popular constitutionalism on a regular basis. Political practices such as those discussed in this Essay set out ways in which institutions can be designed to implement popular constitutionalism. There may well be others already operating, but I hope that this paper gives some flavor of the enterprise of designing institutions for implementing popular constitutionalism.