Constituent power and democracy ‘across generations’: A reply

Revisiting the opening question of political liberalism

Van der Walt starts his commentary by declaring his refusal ‘to consider Rawls a frivolous philosopher who concocted pseudo problems easily resolvable with readily available solutions’ (Van der Walt, 2024, 1437), by which he means turning divisive pluralism into ‘a straw man that can be put up and shot down with little ado’, namely, by invoking the panacea of ‘overlapping consensus’ and shared normative orientations. Of course, as Dworkin has written, major philosophical figures are so complex that each of us ends up having their own Kant, their own Hegel and – why not? – their own Rawls. I believe, however, that there are limits to interpretation, which are overstepped when too much content of the text interpreted is sacrificed to a peculiar, idiosyncratic thesis. Thus, I fully concur with Van der Walt’s sensible refusal to make Rawls a frivolous philosopher but wonder whether such qualification better applies to what he suggests. Before entering specific interpretive disagreement, here is a capsule rendition of his Rawls.

Rawls opens up Political Liberalism with the famous question on the problem of pluralism: ‘How is it possible for there to exist over time a just and stable society of free and equal citizens, who remain profoundly divided by reasonable religious, philosophical, and moral doctrines?’ (Rawls, 2005, 4). We now learn from Van der Walt’s reading that this is basically a rhetorical question: it states a problem for which there is no ‘institutional’ solution. Rawls, however, though fully aware of the lack of such an institutional solution, keeps writing the remaining 367 pages of the book, reading which amounts to a total waste of our time, except for two tiny sections: two pages that outline the nature and operation of the burdens of judgment, that is, the factors that explain why divisive, intractable pluralism is to be expected, and a few other pages in which Rawls urges us to face up to this predicament by not giving up our civility (217, 236, 253). On Van der Walt’s interpretation, Political Liberalism might have as well ended a few pages after that initial question, just those needed for advocating the acceptance of the burdens of judgment and civility. The rest of the book’s content – that is, the notion of a political conception of justice; the political conception of the person; the idea of a well-ordered society; the reasonable and the rational; the liberal principle of legitimacy, key to coping with pluralism through ‘legitimation by constitution’; overlapping consensus and constitutional consensus as distinct from modus vivendi; the five shared ideas of the good, that is, ‘goodness as rationality’, ‘primary goods’, the idea of permissible conceptions of the good, shared ‘political virtues’, and the good of political society; then public reason, the five principles of constitutionalism, and the role of the Supreme Court – all of these contents are irrelevant to Rawls’s main point as construed by Van der Walt: namely, the impossibility of remedying the fact of divisive, reasonable pluralism. Public reason has a special role in this construction: it is there to fail, specifically to fail at bridging the gaps, and thus to remind us by its failure to bow to the burdens of judgment and refrain from losing our civility.

Van der Walt wonders whether the hypothetical correctness of my reading wouldn’t ‘sound the death’s knell’ (Van der Walt, 2024, 1438) to Rawls’s pertinence for political and legal theory, by making reasonable pluralism a predicament unavoidable but nonetheless susceptible of being remedied without generating oppression. Me too, I wonder how Rawls’s contribution would look if, following Van der Walt’s suggestion, his teaching were reduced to the moralistic urging of a Stoic acceptance of pluralism and of a certain disposition of civility. Weber and Freud converged on a similar stance. Both advocated courageously facing up to a fractured, unreconciled (outer and inner) reality, without succumbing to the lure of re-enchanting the world or of one’s own fantasies. I doubt they would be remembered today if that was their main contribution.

Divisive pluralism, public reason, and self-constitution

In response to more specific objections raised by Van der Walt against my reconstruction of Rawls, I’d like to highlight four points.

First, Van der Walt wonders what remains of ‘the social fact of divisive pluralism and the intractable disagreements regarding the moral acceptability of coercive law’ if we assume ‘readily available principles of public reason and constitutional law that that all reasonable persons can endorse and must endorse if they are to be recognised as reasonable’ (Van der Walt, 2024, 1438). The interpretation advanced in SAG appears, to his lights, to reduce Rawls’s insight into ‘deeply divisive pluralism and intractable moral disagreement in contemporary societies to the marginal problem of unreasonable disagreement, unreasonable disagreement that raises no legitimacy concerns and simply warrants being contained like a disease’ (Van der Walt, 2024, 1438). That is an inaccurate interpretation of my text. No person minimally conversant with the paradigm of political liberalism would equate the problem of pluralism, aptly called of ‘reasonable pluralism’, with the clash of ideologically blinded unreasonable constituencies. Those clashes may exist but raise no theoretical challenge.

Second, in my fictive dialogue with Kelsen, wittingly set up by Van der Walt, to the question ‘Why would he [Rawls] have invoked this notion of divisive pluralism with regard to societies that can rely unproblematically on shared truth and normative claims?’, my answer is: because the shared truths available to us may not reach far enough. Shared beliefs in the truth of ‘2 and 2 is four’ and ‘snow is white’ won’t help much when it comes to such problems as taming inflation, reversing climate change, securing peace, and defining privacy. That is why both predicaments are true: namely, that we have divisive, persistent pluralism and we do possess some shared truths and values. The task of public reason is – over and beyond failing and reminding us to observe the duty of civility – to extend the reach of those little truths we share, in order to have them cover as much territory as possible. Incidentally, also normative content is shared across the divide. Let me humbly refer Van der Walt to the instructive reading of Lecture V of Political Liberalism, with the elucidation of five ideas of the good shared across the plurality of comprehensive conceptions. Among them we find even a set of political virtues, that can be presumed to cut across the cultural and moral divides of society. Interpretation can legitimately emphasize one aspect or another of the interpreted material, but should not ignore textual content that speaks against the thesis.

Third, Van der Walt’s remark about Kelsen’s ban on bringing ultimate truths into the political arena offers an interesting contrast to Rawls, unacknowledged by Van der Walt. Both Kelsen and Rawls converge on the incompatibility of democracy and the public pursuit of ultimate truths. However, their convergence ends there. They are no longer on the same page when you raise the question: if I should give up on my truths in the public arena upon realizing that they ‘are not going to get me anywhere’, what’s next? Kelsen’s answer is: to seek compromise. Compromise, or ‘modus vivendi’ can’t be Rawls’s answer, because compromise is unstable insofar as it relies on instrumental motivations. Pace Van der Walt (and Kelsen), you can’t have a just and stable society on the basis of unstable compromises. You need a deeper, more lasting, convergence, in which people agree to what they agree to not because it is convenient, but because they concurrently think that it is right. That is not my idiosyncratic interpretation of Rawls: that idea can be found in Lecture IV, Section 3, on modus vivendi and overlapping consensus. Should that entire Section also be dismissed as indicative of the bogus nature of reasonable pluralism as a strawman set up for being shot down?

Fourth, on my use of a modified version of Plato’s allegory of the cave, Van der Walt correctly identifies its centrality for my book’s argument but misses its meaning altogether. Why would the philosophers exposed to the world outside the cave agree and yet differ on enough aspects of it to warrant a pluralism-preserving principle of not backing up contested views through law? Van der Walt takes issue with the philosophers’ perception of the outside world as ‘essentially unitary and essentially singular, notwithstanding the plurality and severalty of the members of the group’ (Van der Walt, 2024, 1440). I can’t see how that combination of convergence and yet secondary divergence should count as an objection. I construe the philosophers’ perception of the normativity encountered outside the cave (for Plato the Sun symbolizes the idea of the Good, and constitutes a moral source of normativity) exactly along the tracks of Rawls’s normative rendition (quoted above in my reply to Rasmussen) of a society which, in spite of its irreducible pluralism, is nonetheless well ordered because citizens accept, and its basic structure reflects ‘the very same principles of justice’ (Rawls 2005, 35). If ‘there is no divisive pluralism in sight’ in my image of the cave, Van der Walt should explain why instead such pluralism exists in the above-cited passage by Rawls. If my account turns divisive pluralism into a strawman to be shot down without much ado, he should explain why in Lecture 1 of Political Liberalism pluralism is treated differently.

On the philosophers’ deliberation sideways at the entrance of the cave, Van der Walt raises the following question:

How can any deliberation about the presentation of what one has observed be considered the definitive element of that presentation that renders the actual observation itself irrelevant? If this were an apt understanding of deliberation, deliberative reason may as well forego the trouble of careful observation. It could just make things up as it goes, provided the presentation of these made-up things is the outcome of some or other deliberative process that leads to agreement. (Van der Walt, 2024, 1441)

The notion that ‘deliberative reason may as well forego the trouble of careful observation’ reveals that the point of the allegory – the search for a normative standpoint capable of cutting across, as ‘justice as fairness’ does, the irreducible plurality of comprehensive views – has been lost on Van der Walt. Foregoing the trouble of careful observation would be equivalent to concocting a compromise, a modus vivendi, whatever formula placates the demon of conflict among the philosophers in whichever way. Rawls never tires of reminding us that a political conception of justice is not a ‘political mediation’ between rival normative orientations. The overlapping consensus is a (likely, not necessary) consequence of a well-done philosophical work of freestanding construction (Rawls 2005, 39–40), that is, of ‘careful observation’ in the vocabulary of the allegory of the cave.

However, the crucial misunderstanding concerning the allegory of the cave comes right next. Van der Walt questions the intent of SAG to ‘present constituent power as a vehicle of Rawlsian public reason that no longer draws its norms from celestial observation, but does so instead on the basis of a careful assessment of its own history, so as to determine a course of action that is “most reasonable for us”’ (Van der Walt, 2024, 1441). He rejects not just my interpretation of Rawls, but Rawls’s own ‘initial assumption that a conception of justice rooted in a historical tradition (as opposed to eternal principles of reason or natural law) would render it “political” and “not metaphysical”’ (Van der Walt, 2024, 1441) as ‘completely unwarranted’. Drawing on Kelsen, Van der Walt contends that

It makes no difference whether one considers the foundation of law written in the stars or written in one’s own history. The latter is bound to be as metaphysical as the former. Why? Well, because both foundations would require a cognitive appraisal of an existing state of affairs, it would turn on an episteme … In the case of both foundations, the celestial and the more or less fleetingly historical and terrestrial, will the assessment of the foundational source turn on an answer to the question ‘What is the case?’… Rawls’ and Ferrara’s ‘non-metaphysical’ foundations must in the final analysis lead to the questions ‘Who are we, after all?’, ‘What is the script that informs our irrecusable “here we stand”‘, as Ferrara puts it. And it is exactly here that the problem arises to which Kelsen already alerted us long ago: These questions can simply not be answered positively under pluralist conditions (Van der Walt, 2024, 1441-1442).

The problem here lies in the narrow cognitivist interpretation of what it means for a political conception of justice to be rooted in a historical tradition, as though the conception of justice were a reflection of the tradition, the more accurate the better. That interpretation would make Rawls a late and uninspiring epigone of the Savigny historical school of jurisprudence: the closer our political conception of justice to ‘who we are’, the better its normative credentials. Consequently, Van der Walt is misled to think that the benchmark is epistemic – how truly reflective of us one conception or other actually is – as attested by his contending that ‘any attempt to assert cognitive appraisal under these circumstances would need to claim metaphysical access to a hidden truth – be it about the stars, be it about our history, be it written in our hearts – that some see, and others don’t’ (Van der Walt, 2024, 1442).

The irrecusable ‘here we stand’–stance on which our endorsement of a conception of justice ultimately rests, instead, is not a cognitive or epistemic pronouncement. Drawing on theorists of self-constitution, we can think of this stance as a practical commitment with which we identify, as the philosophers sideways at the entrance of the cave do. That commitment reflects something of us without leaving us hostage to ourselves: it transforms us at the same time. It could not be expressed in better words than in Rawls’s:

What justifies a conception of justice is not its being true to an order antecedent to and given to us, but its congruence with our deeper understanding of ourselves and our aspirations, and our realization that, given our history and the traditions embedded in our public life, it is the most reasonable doctrine for us. (Rawls 1980, 519)

Cutting away the term ‘aspirations’ reduces that brilliant formulation to the banal idea that our view of justice should mirror who we contingently have come to be. Aspirations are transformative, possibly theory-driven, and are not epistemic but practical: they are the commitments that make us who we are. Our view of justice is neither dictated to us by a normative order ‘antecedent to and given to us’, as in all foundationalist philosophies, nor is dictated by who we happen to be, as in hermeneutical and historicist approaches. The originality of Rawls’s view of public reason, and of its standard of ‘the most reasonable for us’, consists of having balanced together these two moments in a unique paradigm, whose novelty, unfortunately, evaporates when Van der Walt reduces its core to the merely epistemic question ‘who are we, after all?’.

Making overlapping consensus virtual

In the second section of his contribution, Van der Walt offers his own interpretation of political liberalism, aimed at making sure that Rawls ‘does not go down in history as the raiser of a pseudo problem’ (Van der Walt, 2024, 1443). In his opinion, the proper way to rescue Rawls from this risk is to somehow ‘de-materialize’ overlapping consensus, reconfiguring it as purely virtual, ‘a critically necessary assumption that might allow one to deal with a problem that one cannot resolve’ (Van der Walt, 2024, 1443). This Kelsenian reading of Rawls posits that ‘Rawls never considered ‘overlapping consensus’ an observable fact amenable to cognitive appraisal. As a critically necessary assumption … one would have to argue that Rawls never doubted the sheer fictionality of this key notion in his work’ (Van der Walt, 2024, 1443).

In support of his claim, Van der Walt cites three loci: two are the already mentioned sections on the burdens of judgment and the duty of civility, and the third is an excerpt in which Rawls warns us that public reason cannot perform the miracle of silencing the burdens of judgment and reconciling all our divergences. In the end, argues Van der Walt, Rawls ‘urges’ us ‘not to abandon public reason because of its failure to resolve disagreement. More precisely: The ideal of public reason itself, he says, “urges us” not to give up on it. He “urges us” to stick to it, as if it has not failed, as if it is still intact’ (Van der Walt, 2024, 1444).

While those three loci certainly are crucial for the paradigm of political liberalism, it remains a total mystery how they provide evidence for the purely virtual, presuppositional nature of overlapping consensus. One would need an extra argument nowhere in sight. On the contrary, the idea that Rawls never considered overlapping consensus an observable phenomenon flies in the face of real textual evidence (detailed here below) and incurs a major theoretical difficulty.

As far as textual evidence is concerned, entire sections exist in which Rawls sets overlapping consensus apart from constitutional consensus and modus vivendi, and analyzes the pathways likely to lead from each form of stability and integration to the next (Rawls 2005, 158–168). Key to this reconstruction of the steps toward an overlapping consensus is the idea that ‘citizens have two views, a comprehensive and a political view; and that their overall view can be divided into two parts, suitably related’ (Rawls 2005, 140). The hope is that it be possible then ‘to ground the constitutional essentials and basic institutions of justice solely in those political values, with these values understood as the basis of public reason and justification’ (Rawls 2005, 140). But then in his ‘Reply to Habermas’ Rawls answers the question ‘How can public justification of the political conception of justice be carried out?’ by pointing to ‘the existence and public knowledge of a reasonable overlapping consensus’ (Rawls 2005, 392). Other textual evidence comes from the loci in which Rawls addresses the fate of those who are not parties to an overlapping consensus (Rawls 2005, 64 fn 19) and from the concept of a well-ordered society.

Similarly, Van der Walt’s claims that public reason cannot be built on ‘shared truth claims’ and ‘shared normative claims’ flies in the face (a) of Rawls’ insistence on the fact that it is reasonable and rational for citizens, when debating basic matters, ‘to appeal only to a public conception of justice’, by definition shared across their divides, ‘and not to the whole truth as they see it’ (Rawls 2005, 216) and (b) of Rawls’s more general point that whenever a public and binding decision must be reached, for example in a criminal trial, we should not appeal to the whole truth as we see it, but should argue from the more limited truth that is shared, for instance in the guise of court records. What better evidence could be adduced for Rawls’ conviction that public reason enjoins us to start our arguments from shared reasons and move from there to contested conclusions, in the hope of making them shared too? Van der Walt’s claim that ‘public reason is not a reality, but an easily abandonable, precarious ideal’ incurs the obvious difficulty of accounting for Rawls’s singling out the Supreme Court as an exemplar, or an institutional realization, of public reason. Public reason may also be an ideal, but its institutional embodiment testifies to the fact that it is also a thick, institutionalized ‘reality’.

On constituent power dissolved

Finally, Van der Walt’s account of constituent power is flagrantly incompatible with anything Rawls had to say on the subject and bears little resemblance with what is generally understood by this term. At the core of constituent power, according to him, we find a propensity to deal ‘with divisive pluralism in a civilised way’ (Van der Walt, 2024, 1446). The unity of intent of constituent power is best dissolved, as Van der Walt would have it, into an ‘unfathomable panoply of constituent forces’ (Van der Walt, 2024, 1446). So plural is this panoply of constituent powers, that some of them, like the Supreme Court (whose controversial shifts from Plessy to Brown, and now from Roe to Dobbs are depicted as manifestations of constituent power (Van der Walt, 2024, 1446-1447), are actually constituted powers.

Be that as it may, let me just recall that when Rawls addresses constituent power, contrary to Van der Walt’s elucubrations about it, he affirms that it belongs ‘to the people’ (Rawls 2005, 231), not to a panoply of constituent forces, and ‘sets up a ‘framework to regulate ordinary power’ (Rawls 2005, 231), in lieu of musing over the burdens of judgment and the duty of civility. The point of constituent power is not simply to create a basic structure, but to inscribe in a constitution ‘the political ideal of a people to govern itself in a certain way’ (Rawls 2005, 232) and ‘fix once and for all certain constitutional essentials’ (Rawls 2005, 232), to which all constituted powers and the amending power exercised by single cohorts of citizens will have to remain faithful unless they are prepared to change the regime. Furthermore, Rawls continues, ‘the aim of public reason is to articulate this ideal’ (Rawls 2005, 232), needless to say consensually, not solely or primarily by contemplating abysses of dissent and urging the duty of civility on the contenders.

These few quotes from Political Liberalism suffice to mark the distance between Rawls’s conception, condensed but clear, of constituent power and Van der Walt’s interpretive wishful thinking that would replace the constituent subject’s unity of purpose, unfolding across generations, with an unfathomable panoply of forces, some of which even constituted, that do little more than remind us that agreement is not to be expected and this predicament must be faced up to without going at each other.

In the end, I have sympathy for the gesture of visionary courage with which Van der Walt addresses Political Liberalism, so little constrained by textual evidence and pushing to the limit the unsettling, almost tragic message, antithetical to reconciliation, that emanates from the twin doctrines of the burdens of judgment and the duty of civility. However, I remain skeptical about the value of this interpretive line. While it leaves my own interpretation unscathed, too many of the conceptual tools that Rawls has contributed to contemporary discourse end up idling inoperative, the force of his paradigm – aimed at making divisive pluralism livable, when most other contemporary theorists merely repeat the hollow mantra of ‘living together in diversity’ – is dissipated, and in the end the pertinence of his groundbreaking work migrates from the desks of political and legal theorists to those of moralists. Van der Walt’s rendition exposes Rawls to the risk of becoming a well-meaning, inconclusive moralist who offers us two recipes, premised on civility and reciprocity, for coping with the condition of unreconcilable pluralism. Rawls deserves better.

Democratic sovereignty ‘across generations’ and the transgenerational people

Winter converges with me on the symptoms of the disease that is currently affecting many democratic societies – ‘seismic fissures in liberal democracies in the form of populist, ethno-nationalist, and authoritarian movements’ (Winter, 2024, 1451), but radically differs on the diagnosis of the disease and finds that the cure suggested in my book is ‘no better than the disease’ (Winter, 2024, 1451). Sequential sovereignty, misleadingly hailed as a solution in SAG, constitutes instead the problem.

As he puts it, ‘the problem of populism isn’t that, in claiming to speak for “the people”, a segment of the current electorate violates some coherent notion of [transgenerational] sovereignty’ (Winter, 2024, 1451) – which for him ‘has the same anti-democratic vice’ of imposing some external content, that is, ‘the value-judgments of the founders as interpreted by a constitutional court’ (Winter, 2024, 1451) to the current citizens’ will. The problem is rather that populists seek to impose their will on a more pluralist majority. What populists and liberal advocates of sequential sovereignty paradoxically share is the idea that ‘political questions are resolved by appeal to authority and not through democratic debate, negotiation, and compromise’ (Winter, 2024, 1451, emphasis added).

The opposition between solving political questions ‘by appeal to authority’ and ‘through democratic debate, negotiation, and compromise’ highlights the problematic quality of Winter’s view. While the democratic authority appealed to, in the case of constitutional democracy, is traceable to the collective author (inclusive of the living citizens) of a constitutional script designed to shape the life of the political community, ² it is totally unclear what Winter means by ‘democratic debate, negotiation, and compromise’ as alternative methods for solving political questions. As history shows, negotiation and compromise could easily generate monsters, in the absence of normative parameters about what can be negotiated and what cannot. Even neo-Hobbesian ‘modus vivendi’ liberals à la Gray and Horton do admit that negotiation and compromise, or ‘modus vivendi’ in their terminology, are ‘far from the idea that anything goes’ (Gray 2000, 20), and I take it that Winter would not disagree. Then the appeal to some kind of ‘authority’ comes right back into the picture. The parties may agree to whatever compromise, but that compromise must be checked, by some institutional actor, against the standard of ‘proper’, as opposed to unacceptable, compromises on the basis of some benchmark. Winter is then back to square one. All the appeal of democratic debate, negotiation, and compromise suddenly vanishes in front of basic, ineludible questions such as: who is to set the benchmark against which compromises are checked for acceptability? If it is set by the same parties who negotiate, then there is no independent benchmark at all, and we might as well openly avow that ‘anything (that pleases the parties) goes’. If, instead, the authority of the benchmark is independent of the parties’ will, then where does it come from and what are its normative credentials? Assuming the standard of fair, as opposed to rotten, compromise has normative traction, who is to apply it and enforce it by invalidating, if necessary, what the negotiating parties agree to?

In more general terms, what is missed by Winter is the notion that democratic debate, negotiation, and compromise don’t free his conception from the need to presuppose, as highlighted by Frank Michelman in his contribution, some standard of fair debate, negotiation and compromise antecedent to and independent from the single event under scrutiny and whose ‘fulfilment’ is assessed by actors impartial with respect to the parties – precisely, what a constitution (written or unwritten) supposedly offers and what an institutional referee (usually, but not necessarily a court), does.

Another manifestation of the same conceptual lacuna can be observed in Winter’s criticism of my rehauling of Plato’s allegory of the cave for highlighting the non-foundational normativity of Rawls’s notion of public reason. In line with his intimation that all political questions, including those concerning higher law and rights, be ideally solved through ‘democratic debate, negotiation, and compromise’, Winter overlooks Rawls’s point that a political conception of justice, prior to becoming the focal point of an overlapping consensus, needs to be ‘freestandingly’ constructed through philosophical argument. It is this constructivist, ‘freestanding’ moment, that sets justice as fairness apart from sheer political compromise and grounds the distinction between stability and ‘stability for the right reasons’. Rawls is very explicit on this point: ‘we do not look to the comprehensive doctrines that in fact exist and then draw up a political conception that strikes some kind of balance of forces between them’ (Rawls 2005, 39). In fact, ‘to do so would make it [justice as fairness] political in the wrong way’ (Rawls 2005, 39–40, emphasis added).

Winter’s claim that if justice as fairness ‘is political and not metaphysical, there is no outside the cave’ is either a misunderstanding of Rawls or a totally external criticism. If ‘there is no outside the cave’, then we’ve left the domain of normative political philosophy. Justice becomes equivalent to what the majority of the cave dwellers believes justice to be. What is missed is the ‘tertium’, the alternative offered by public reason’s standard of ‘the most reasonable for us’, between objective, self-imposing, ineludible out-of-the-cave principles and in-cave compromises that mediate arbitrary preferences based on subjective opinions. This lacuna is even more striking in that Winter uses the relevant quote – ‘the appropriate constitutive normativity 'is the one that realizes the best fit . . . between its free-standing principles and the historical, political, and cultural features salient for the to-be-constituted people’ (Ferrara 2023, 2; emphasis added)’ (Winter, 2024, 1453) – but misses the explicitly stated point that within political liberalism the ‘free-standing principles’, needed in order to keep justice distinct from what justice is believed, even by Nazi, to be – are by no means self-sufficient, as in all kinds of foundationalisms. On the contrary, they must be optimally fitted with, not turned into an expression of, the ‘cultural features salient for the to-be-constituted people’ as found inside the cave. In the polarity of transcendent principles and arbitrary belief or preference, what is missed is the novelty of ‘the most reasonable for us’ which partakes of both worlds, the one inside and the one outside the cave.

Legal principles and cognitive assumptions: Plessy and Brown once again

The reformulated allegory of the cave – in which not one single philosopher, but a group thereof, consult on how to report what has been diversely experienced outside the cave, and are led by their common public reason to exclude that legal coercion could ever be used to buttress contentious, ‘non-overlapping’ aspects of their plural reports – is meant to convey a sense of how the standard of ‘the most reasonable’ restrains the otherwise unfettered exercise of full constituent power. That picture, according to Winter, does not account for how temporality comes in. The ‘political ideal’ to govern the cave in a certain way – un-Platonically validated by consent inside the cave – should allow for adjustments and variations that reflect changed circumstances and should allow present generations of citizens, or cave-dwellers, to avoid ‘the grip of the dead’ by re-aligning the project with their historically changed sensibilities while not disfiguring it altogether.

Winter correctly identifies my strategy for defending the continuity of the ‘political ideal of a people to govern itself in a certain way’. Fidelity across generations must be directed, and can realistically be expected to be directed to the core principles of the constitution, its underlying view of justice, and its list of basic rights and liberties, but not to the specific applications that the founders of a political order made of those normative essentials, in the light of the beliefs they happened to share about the world.

Winter takes issue with that conceptual separation, but especially with my recourse to the transition from the application of the ‘equal protection’ clause of the Fourteenth Amendment in Plessy v. Ferguson to the implications attributed to the same clause half a century thereafter in Brown v. Board of Education. I will respond to his objection in two steps, first in general and then as it applies to Plessy and Brown.

As regards the more general point, Winter claims that my strategy for combining fidelity to constitutional essentials and necessary updating of the constitution ‘fails because it assumes that principles have meaning and content separate from the situated understandings that animate them’ and that ‘an ahistorical realm of “principle’’’ exists, that ‘stands above the social values and assumptions that give a principle whatever content it might have’ (Winter, 2024, 1452). He contrasts two very different conceptions of freedom – the neoclassical economist’s idea of maximizing preference satisfaction in a market and Kant’s view of freedom as a moral subject’s consistent self-legislation – and somewhat rashly concludes that ‘each conception of freedom is incommensurate with the other’. Neglecting to address the point, widely shared across major philosophical divides, ³ that for any two meaning-constructions to be perceived as incommensurable views of something, rather than as totally unrelated as carrots and parrots, they must share at least one point of contact, which provides then the basis for the opposition and the related claim of incompatibility, in our case Winter neglects to mention that both the Kantian and the standard neoclassical economists purport to account for the human capacity for free action. They are views are about freedom. As I observed in SAG, to claim that a principle changes its nature and object, as opposed to its implications, if interpreted according to different premises is as absurd as claiming that unless we agree on the specifics of an Aristotelian or a Newtonian explanation of bodily motion, we are not observing the same falling of a ripe fruit from a tree. We can of course construe perception in this way, by defining it as the total sharing of the theoretical (or practical) underpinnings of what is conveyed by the senses or practical experience, but the result is an unappealing view according to which communication is possible only within small circles of theory insiders, if not perhaps total solipsism.

Article 11 of the Constitution of Italy states: ‘Italy rejects war as an instrument of aggression against the freedom of other peoples and as a means for the settlement of international disputes’. The meaning of this provision, so important for Italian politics, pivots around the reference assigned to the signifier ‘war’. That assigned reference may make the difference between an Italian cabinet’s legitimately participating in several contemporary conflicts that fall short of the classical idea of war (most likely on the minds of the Italian members of the Constitutional Assembly in 1946–47) as a conflict between regular armies following a formal declaration of war, and the same cabinet acting unconstitutionally if a more updated definition of war is adopted, undeclared and including militias, terrorist groups and other sorts of agents. Should we conclude, along with Winter, that there are as many principles of ‘the rejection of war’ as there are interpretations of Article 11? That conclusion would make the Constitution’s meaning totally interpreter-dependent and thus – barring an unlikely convergence of all interpreters – also the regulatory power of the Constitution would disappear. How can then a bona fide Italian political actor abide by the duty of constitutional fidelity, if the meaning of that crucial principle wavers constantly on account of its key-term’s indeterminacy of reference? Is there a way of singling out a unitary normative commitment, to which fidelity is to be directed, underneath the frequently changing cognitive assumptions concerning what war is? Pace Winter, I think that possibility is provided by directing one’s loyalty to the normative commitment arguably imputable to the subject of constituent power. A possible rendition of the content of that commitment could be that the Constitution of Italy registers its author’s commitment to reject a bellicose spirit that accepts war as an instrument of politics and to reduce recourse to belligerence, whatever its present-time most sensible definition, to the historically possible minimum.

Concerning my example of the transition from Plessy to Brown, Winter’s fair reconstruction of my argument (Winter, 2024, 1453-1455) exempts me from the need to restate it. Let me address his three objections.

First, Winter contends that ‘neither the Plessy nor the Brown Court implemented or enforced the “principle” it espoused’ (Winter, 2024, 1454). In both instances the justices, or the majority thereof, were paying lip service to the principle they pretended to embrace. In apparent violation of the ‘separate but equal doctrine’ invoked to justify the Plessy decision, in Cumming v. Richmond County Board of Education (1899), just three years later, the Court upheld the imposition of ‘a school tax on all county residents’ which however resulted in providing a high school facility only to whites; and in Giles v. Harris (1903), the Court rejected ‘a challenge to the 1901 Alabama constitution which, it openly acknowledged, systematically disenfranchised black voters’ (Winter, 2024, 1454). The Brown Court did not fare better: although in the years following Brown it extended desegregation to public beaches, public baths, and city buses, in 1962 it upheld the decision, on the part of the city of Jackson, Mississippi, to close rather than desegregate its swimming pools, despite the racist message conveyed by it: better to have no swimming pools at all, than integrated ones. My response is that the objection is beside the point: insincere, hypocritical referees, who operate inconsistently or even in bad faith, do exist in all games and all human practices, and courts are no exception. Accurate though it might be, this observation speaks to the need for having better referees, not for eliminating the function of refereeing, devolving it to the players, or replacing it with a compromise between the contending parties.

Second, Winter contends that ‘the meaning of segregation was actively contested in 1896, as it had been for half a century’ and in fact ‘for over a century before Brown, abolitionists, Reconstruction Republicans, and early 20th-century civil rights activists fought for a broader understanding of equality’ (Winter, 2024, 1455). Again, this observation is factually faultless but irrelevant to the argument. That the formula ‘separate but equal’ was contested long before its official rejection, even decades before it was created by the Plessy Court, does not affect the merit of my argument because it was not officially instated or rejected by the institution entrusted with ‘saying what the Constitution says’. What other actors affirm or reject can enrich our sense of the debate in the public sphere, before and after the Court’s pronouncements, but what matters from the angle of the sequential view of democratic sovereignty is what the institutional interpreter of the constitution affirmed in a way binding for all and on the basis of which empirical assumptions.

Third, Winter cites several cases – Wickard v. Filburn (1942, 124–25), Lopez (1995, 557), and Sibelius (2012) – which evidence back and forth oscillations in the Court’s adoption and then rejection of certain cognitive assumptions, in this case concerning the tenability of the distinction between activities, mostly economical, of merely local or national significance. Furthermore, Winter mentions shifts not simply in the background assumptions, but in the very interpretation of a principle – for example, Katz 1967 redefining ‘the Fourth Amendment as protecting not property, but privacy’ and thus now, contrary to past practice, making evidence subject to search and seizure and requiring a warrant for wiretapping (Winter, 2024, 1455-1456) – as constituting ‘the strongest counterexamples’. I don’t dispute the reality of these and other oscillations – for example, from admitting (in Bakke (1978) and Grutter (2003)) to excluding (in Students for Fair Admissions v. Harvard (2023)) consideration of race in admission to university programs – but I deny that they count as counterexamples. In the sequential conception of constitutional democracy, the mandate of a supreme or constitutional court is to interpret, from the point of view of the transgenerational author of the constitution and not that of the present voters, ‘what the constitution says’. It is then clearly within its mandate to correct past pronouncements in light of new factual elements, to backtrack if these new factual elements at a later time seem more dubious, and certainly to reinterpret principles in the light of new sensibilities meanwhile arising – as exemplified by the rise of concern for privacy, or later for environmental protection. Taking these new sensibilities into account and allowing them to reshape the interpretation of constitutional principles inflicts no injury to democracy. The difference between democracy and juristocracy, which correctly Winter wishes to keep neat and sharp, lies somewhere else. As Rawls put it: ‘The constitution is not what the Court says it is. Rather, it is what the people acting constitutionally through the other branches eventually allow the Court to say it is’ (Rawls 2005, 237). Thus, the crux of the matter, to address Winter’s concern, is not whether the Court can or cannot change its mind on empirical and normative matters, whether it can or cannot correct previous pronouncements that from the beginning were ‘egregiously wrong’, but whether the institutional arrangements in place and their present pattern of implementation leave enough leeway for ‘the people [here, the citizens] acting constitutionally through the other branches’ to question, modify or radically reject the court’s interpretation of the constitution. The work of Bruce Ackerman documents how since the time of the New Deal, this leeway has shrunk. That is the problem that needs to be addressed, not the interpretive oscillations of the Court.

Serial sovereignty, the ‘wanton’ republic, and the ‘indistinct’ republic

My case for a sequential, as opposed to a serial, conception of democratic sovereignty, is built in the negative, as backed up by the need to avoid the very negative consequences that potentially follow from a serial model. Winter addresses only two and questions the likelihood of their occurrence. My example of a volatile, ‘wanton’ republic is presented as extreme, that is, unlikely, and as a ‘puzzling hypothetical’. I agree that the case of a polity keen on alternating at each generation a presidentialist and a parliamentary, a bicameral and monocameral institutional structure, combined with established and disestablished state religion, and constitutionalized or contestable social rights looks like a whimsical collective subject exercising arbitrary, ephemeral, unreflected preferences. And I also concur with Winter’s point that ‘real-world societies are situated’ and so are the political actors active in them. They respond to political forces that influence their action and prevent such conduct from reflecting just random, ‘wanton’ volition. Still, the object of our discussion is not the probability of a certain danger actually occurring, but the total impotence of a certain theoretical model to offer remedies once the black swan is unfortunately with us.

Similarly, the danger of the ‘indistinct republic’, that fails to anchor its generations to their predecessors through a political project for self-government and may induce them to resort to ethnocultural forms of self-identification is downplayed by Winter on account of the deep continuity, emphasized also by Michelman and Llewellyn, of constitution and culture, or ‘ways of living and doing.’ Two remarks are here in order. First, not all constitutions benefit from this continuity with the local political culture. Those born out of regime change – just think of post-war Germany and Italy – are somehow more transformative than reflective of the ethnocultural identity of the democratic sovereign. In these cases, it is then of great importance that the political values of the constitution sink their roots deeper and deeper into the citizens’ culture and do not remain empty formulas. Second, surely the probability that a future generation of Italians identifies more with the French (post-war) framers than with their own is almost as remote as that of their identifying with Bronze Age chieftains or medieval samurais, but once again the object of my critique is the impotence of the serial model of democracy to block a regressive, purely ethno-national form of popular self-identification.

Finally, the third danger of serial democratic sovereignty – that is, a polity in which the various generations are not equal in terms of rights and liberties not, as indeed happens, out of historical contingency, but out of intentional design on the part of one segment of ‘the people’– is addressed by Winter subsequently, in his discussion of reciprocity.

This section of his commentary concludes with the observation that while these dangers of serial democratic sovereignty are all hypothetical and indeed unlikely, ‘the danger of a 'wanton' constitutional court is actual’ (Winter, 2024, 1458), and ‘the theory of transgenerational sovereignty has no institutional resources for responding to a court gone wrong’. Both points are problematic in my opinion. While ‘bad’, corrupt, incompetent, partisan referees, are always a concrete possibility, such possibility by no means undercuts the necessity and desirability of the function of impartially refereeing between contending self-interested actors – regardless of whether these actors contend over the gravity of a foul in a soccer game or over the constitutionality of a statute. Second, it is incorrect that the sequential view of democratic sovereignty has no conceptual resources for responding to a court gone wrong. The problem of a court that unduly crosses the line between interpreting and transforming the constitution is addressed in a special section (SAG, Chapter 6, Section 3b, ‘The red line between interpretation and transformation’), and so is the need for possibly correcting a court’s pronouncement deemed disputable or totally to be rejected (SAG, Chapter 6, Section 3c, ‘Correcting the highest interpreter: author and interpreter of the constitution in conversation’). Admittedly, the proper, desirable institutional implementation of those conceptual resources is not really addressed in the book, which revolves around other priorities, but the conceptual juncture where those much-needed reflections could be grafted onto is clearly there

On reciprocity and freedom across generations

Steve Winter credits SAG for raising ‘the profound question of what the generations owe one another’, but my notion of vertical reciprocity would prove inadequate, for lack of symmetry: ‘We cannot deliberate with long-dead founders to reach decisions that reflect the unforced force of the better argument. Nor can we ask later generations about their circumstances and needs’ (Winter, 2024, 1458). Furthermore, unintended consequences complicate the picture. What is meant today as fully respecting the equal freedom of future generations may turn out to limit it: the right to keep and bear arms, designed to secure the viability of civic militias a couple of centuries thereafter may ‘hamstring[s] modern majorities faced with endemic gun violence and record numbers of mass shootings’ (Winter, 2024, 1459). As Winter points out, the problem of the ‘underdetermined’ republic – the third flaw of serial conceptions of democracy – ‘is baked into the very idea of a constitution’ (Winter, 2024, 1459). That is undoubtedly true, and I find Winter’s account of how vertical reciprocity relates to recognition perfectly sensible. In elucidating the meaning of reciprocal recognition in the intergenerational context, he writes,

We start by recognizing the degree to which our predecessors are ‘constitutively present’ to us as culture, worldview, and political memory. We listen considerately to what they have to say – that is, without reducing them to an anachronism that mirrors how we see the world today. In turn, an earlier generation treats us with recognition and respect by bequeathing to us the best system it can manage. It entrusts us, however, to exercise the same freedom that they enjoyed. We honor our predecessors by accepting both the constitution and our authority to remake it. And so on with our successors. Recognition is an act of the imagination that extends empathy and authority diachronically in both directions (Winter, 2024, 1459).

I endorse Winter’s point that ‘law can endure across generations. But democratic law can exist across generations only through negation. In affirming the choices of its predecessors, a current generation implicitly affirms its right to alter, amend, or abolish what came before’. Every generation retains sovereignty in the sense that its members can break out of the communal project. They can make a revolution, which is a historical fact. But insofar as they freely wish to remain in the same boat as their predecessors, citizens must understand their power to amend the constitution to stop short of subverting the constitutional essentials, the constitutive rules of the political game. One might be tempted to say that this sequential view of democratic sovereignty leaves them less free than a serial one, but that claim is as absurd as complaining that as chess players we’re not free to move the castle diagonally. We’re not forced to play chess, but if we play chess those are the constitutive rules of the game, that make chess chess and constitutional democracy constitutional democracy. If for some reason we don’t care about sharing an ongoing political project, then we’re free to use our constituent power in a totally unrestrained way. What is incoherent is to have it both ways: to pretend to be ‘under a constitution’, as opposed to amid a revolution or regime change, and to treat the constitution we’re under as a mere projection of our will alone, entirely at our disposal. Our freedom remains unaffected: we’re free to leave the shared constitutional project and, I fully agree with Steven, that if we the living citizens endorse the inherited constitutional project, that’s still an expression of freedom, not an unconditional obligation. If we want to have a political identity anchored in constitutional essentials, then we must preserve some of the planks of the constitutional boat while we replace the others. Our deference is not to an abstraction, but to a concrete project from which we could break out if we so desired. We freely renew our affirmation of a concrete project: which means, translated into normative language, that you can’t alter the defining core of the project and still claim to be affirming it.

The people and the population

Concerning the first issue, Niesen adroitly observes that my account of sequential democratic sovereignty neglects a crucial factor. In terms that are also more generally valid, but addressing specifically the cases of Italy and Germany, he points out that ‘“a people” as transtemporal actor is not just replenished by descendants of those that lived through or died by the hand of Fascism and Nazism, but also by immigrants and emigrants’ (Niesen, 2024, 1470). On becoming citizens, why would these new constituencies have to feel bound by the commitments of predecessors who are not their own? Wouldn’t a serial, Jeffersonian conception of democratic sovereignty appear ‘better geared to describing the moral and political situation that immigrants, having obtained citizenship, find themselves in as far as their share in constituent power is concerned’? (Niesen, 2024, 1470). After all, Niesen concludes, ‘immigrants can hardly be asked to share the historical-practical imprint of the original constitutional moment’ (Niesen, 2024, 1470).

Two ways of responding to this deficit of inclusion are then outlined by Niesen as alternative remedies. The first consists of biting the bullet and imposing ‘a historicist understanding of the legitimate exercise of constituent power on immigrants, leaving them free to mobilize for future re-interpretation, perhaps along the lines of a broader account of rejecting historical injustice wherever they find it relevant’ (Niesen, 2024, 1470), that is, by bringing memories of the injustices they suffered into constitutional discourse and the national narrative. The other way is ‘to let immigrants radically challenge the particularist ‘never again’-identity of the receiving country, and opt for a less historically impregnated, and less protectionist, constitutional setup’ (Niesen, 2024, 1470).

The second alternative is believed by Niesen to be more in line with the ideal of undiminished political freedom and necessitates ‘compromising in the future exercise of constituent power, since the historical features of its liberal constraints cannot be shown to be binding on immigrants’ (Niesen, 2024, 1471).

I thank Peter Niesen for having raised this fundamental question, peripheral in SAG because it was one of the main foci of attention in The Democratic Horizon (2014). The factual background that raises the question as an urgent one is straightforward: according to the World Health Organization, in the year 2019, 3.5 % of the world population (about 272 ml individuals) migrated to another country. Highly-developed countries, which include (albeit not exclusively) democratic regimes, evidence a migrant-to-resident population rate of 12.5% (with UK, US, Canada, Germany, Sweden, Switzerland, Belgium, and Norway all above 15% and up to over 20 %).

This new trend, susceptible to increase as new climate-change-related migratory trends accrue to the extant ones, calls into question the ‘methodological sedentariness’ that undergirds most accounts of justice and legitimacy – that is, the tacit assumption that the justice-reflecting quality and legitimacy of political institutions is a function of their relation to a presumably sedentary citizenry. From a normative standpoint, if we address political legitimacy after shedding away ‘methodological sedentariness’ wide-ranging new questions need to be addressed: what reciprocal political, as opposed to moral, obligations bind citizens and (various types of) residents? What do citizens owe to residents and vice versa?

As far as constituent power is specifically concerned, I would frame Niesen’s point in the following terms. Reciprocity-honoring democratic polities supposedly pivot around a political project for self-government, inclusive of rights and ‘political values’, which is usually scripted in a constitution, shared among the citizens across the political spectrum, and accepted for principled, as opposed to mainly prudential, reasons. Should migration-related modifications in the composition of the population be allowed to affect that picture? Notice my intentional use of the word ‘population’, which includes citizens and residents (legal and non-legal). My answer is then differentiated.

I do not think that (various types of) immigrants – often forced to migrate and settle in the host society by circumstances beyond their control – should be cornered between the requirement to endorse these constitutional essentials for the same principled reasons as native citizens or alternatively be labeled ‘unreasonable’ and be destined to ‘containment’ (a theme which will be discussed below). The paradigm of political liberalism is hospitable and capacious enough to allow for an intermediate terrain of accommodation, consisting of two basic strategies for integrating well-disposed immigrant constituencies. First, as far as ‘political justification’ is concerned, under-theorized ‘conjectural arguments’ could be offered to these sectors of the population, that aim at hermeneutically engaging their comprehensive conceptions in order to offer them internal reasons for endorsing the liberal constitution (Ferrara 2014, 73–76). Second, should these efforts prove unsuccessful, a last-resort remedy is to allow members of these constituencies to endorse one or more of these, for them alien, constitutional essentials for prudential reasons, in the hope that over time the process of integration in the host society might – on a voluntary basis – lead them to a change of perspective and motivate them to pursue citizenship on the basis of a voluntary endorsement of the same constitutional essential(s) for principled reasons (Ferrara 2014, 105–108). Societies with close to 20 % immigrant population may then be described, from the angle of a renewed political liberalism, as ‘multivariate democratic polities’ in which the overlapping consensus, though still undergirding the legitimacy of the exercise of authority, encompasses some sectors of society for reasons of principle and other sectors for prudential reasons, with those who entirely reject the compact still remaining marginal fringes. The Rawlsian ‘stable and just society of free and equal citizens’ cannot thrive if the gap between the people and the population is too wide.

On the other hand, when immigrant residents become citizens, they certainly have the same rights and partake of amending power, subject to the limitations discussed in SAG, to the same extent as all other citizens. The idea or perhaps just the hope is that if being initially integrated as residents, then spoken to in culturally congenial terms by the host institutions and not penalized in any way for accepting the constitutional essentials for merely prudential reasons, then immigrants might take the step of seeking citizenship when they truly feel like endorsing those essentials, and the project they spell out, for internal reasons. Hope of course is just hope, and no guarantee can exist – as it will be emphasized again below in my response to Schupmann – against what a majority of the voters, if its determination lasts long enough, can actually make the constitution say, or against a lasting majority’s capacity to actually start a new compact.

Constituent power and supra-national self-constitution

In the final paragraphs of his paper, Niesen gestures toward an interesting comparison with Habermas’s take on constituent power. Certainly, Chapter 3 of Between Facts and Norms can be interpreted as a discourse-theoretic reflection on the limits that the citizens’ ‘public autonomy’ – a term partly equivalent to the never mentioned ‘constituent power’ – must respect when establishing the rights that ‘the citizens must mutually grant one another’ (Habermas 1996, 118). However, this discourse-theoretic justification of indispensable rights is general, binding on any process of constitution-making, just as classical natural law notions are or as if Rawls had posited the outcome of the original position as a limit on all constitution-making anytime anywhere. I concur with Niesen on the fact that although he never directly addresses ‘implicit unamendability’, Habermas too ‘is concerned with upholding the ratchet effect of constitutional innovation’ (Niesen, 2024, 1471) and I also consider Habermas’s dual form of constituent power, at work simultaneously in the construction of EU institutions and in the member-states, as a very interesting contribution. ⁴ Finally, I agree with Niesen’s claim that ‘in giving national constitutional orders a functional task to guarantee their survival in supranational orders (upholding historical achievements of liberty and justice), he [Habermas] is unconcerned with saving particularity’ (Niesen, 2024, 1472, emphasis added) – the point of constituent power being more to ensure that the ‘system of rights’ required by the discourse-theoretic derivation of a legitimate constitutional democracy be correctly instantiated in the case at hand than to express some unique ethical substance, though in full respect of the requirements rooted in discourse.

Niesen’s question of whether an elective affinity links my authenticity approach to constitutionalism on one hand, and the primarily domestic focus of my reconstruction of the legitimacy of exercises of constituent power on the other, prompts me to give a tongue-in-cheek negative answer. Authenticity has no boundaries. Nothing prevents a supranational constitution-making process, on the scale of the European Union, or even on the global scale of the United Nations (before the present so unpropitious times), from inscribing in a constitution-like document, over and beyond those rights that the discursive presuppositions, the original position or any other philosophical construction urges on us, the very substance which forms the raison d’être, or ‘la finalité’ (to use the EU jargon) of our committing ourselves to live politically under such terms and to do so either in consonance with, or transforming, previous local commitments. If the EU has so far refrained from pushing such a line – of singling out the core political values that the Union stands for in the form of constitutional essentials scripted in a document called ‘the Constitution of the European Union’ – that is only due to the aversion of several national élites to heeding that sort of constitutional pathway and to their preference for a functional, quasi-written constitution embedded in Treaties that only lawyers can decipher (Ackerman 2015, Ackerman 2019, 21–23; Ferrara 2018).

Similar considerations apply to the intelligibility of ‘global constitutionalism’ and the attendant global constituent power. I can’t engage here the enormous literature on this issue, to which Niesen has himself egregiously contributed (Niesen 2023). I’ll just offer two considerations. First, instead of complicating it, global constitutionalism alleviates the thorny, ubiquitous problem of identifying the subject of constituent power. Quite simply, no single human being can be legitimately excluded from partaking of such constituent power. Philosophers, dressed in the robes of Rousseauian ‘legislators’, can then freestandingly construct alternative normative renditions (including the ‘alter-globalization’ views of the World Social Forum) of ‘the good for humanity’ – the object of a Rousseauian ‘general will of humanity’ – and submit them to this subject. The question is rather if only the rights of humans should be protected by a global constitution.

My second consideration is that the mainstream approach to global constitutionalism has extracted the relevant ‘essentials’ from the existing practices of global institutions, from the UN to the WTO or the ICJ. This fact inclines global constitutionalism towards monistic political constitutionalism, with all the attendant difficulties in overcoming a mere ‘summary view’ of constitutive rules. However, the potential exists for applying the paradigm of political liberalism – including the role it assigns to constituent power – to supranational and potentially global instances of constitution-making if the necessary adjustments are introduced. ⁵ Again, the ‘political ideal’ of a collectivity ‘to govern itself in a certain way’ knows of no scale restrictions: given a fair system of representation, when the number of constituents prevents direct participation, that ideal could be the authentic expression of a local community as well as of the EU or humanity.

Three interpretive queries

First, concerning the relation of the people to its temporal segments Schupmann starts from the correct premise that ‘a present generation is not and does not own the identity of “the people”’. Then he adds that ‘in its role as steward, a present generation has been invested by past generations with a fiduciary duty to manage, protect and enhance it for future generations. That duty is to uphold “the people’s” fundamental identity over time’ (Schupmann, 2024b, 1476). Because of this fiduciary duty, ‘the electorate does not have a valid claim to exercise “the people’s” constituent power. At least, not unilaterally’. Then, Schupmann concludes, ‘the electorate is a constituted power…. Its power is derivative and conservative rather than original and creative’ (Schupmann, 2024b, 1476). While I completely agree with the idea that single generational segments of the people, acting as electoral bodies, are subject to normative restrictions in the exercise of their power to amend the constitution, I find Schupmann’s characterization of the role and powers of the electorate somewhat over-restrictive. The living generations of the people have a dual status. On the one hand, as a collection of voters, the electorate is a constituted power bound to play within the rule of the democratic game. On the other hand, the same citizens are not just representatives or stewards of the transgenerational people. They are also part of the people, and in that capacity, they possess a fraction of the constituent power of the transgenerational democratic sovereign. Because it is only a ‘fraction of’ the transgenerational constituent power, the living citizens’ amending power must not be misused for disfiguring the overall political project that they are entrusted with carrying forth: for this reason, it is constrained power. However, to the extent that the constitutional essentials are safeguarded, the living citizens are not bound to simply represent someone else’s will nor are they prevented from expressing their own will in the form of higher law. They are free to make their own unique voice heard and to constitutionalize novel sensibilities and unprecedented values, as history may give them occasion to care about: for example, the values of privacy, environmental protection, sustainability, a balanced budget, protecting biodiversity, and so on. Thus, the living citizens may indeed be ‘original and creative’, provided that they operate within the limits of the constitutional project that they partake of. Were the electorate simply a constituted power, it would logically be totally bound by the constitution in all its details and not in a position to amend it at all. If specific provisions allow voters to amend the constitution, directly or through their representatives, then they must be attributed some kind of not-constituted but somehow constituent power.

Second, on the role of a constitutional court as the enforcer of vertical reciprocity, another caveat is perhaps useful. Schupmann observes that ‘because it has a rather different motivational set, the constitutional court represents “the people” differently than the electorate’, and that ‘the constitutional court’s conservatism may make it a better steward of the interests of other generations of “the people”’ (Schupmann, 2024b, 1478). I would incline to emphasize that what makes a supreme or constitutional court a (putatively) better steward of the interests of past and future generations is the nature of its mandate. While legislators, prime ministers, or presidents who privilege the interests of their electoral base are just considered mediocre, relative to historical figures – called ‘statesmen’ in patriarchal terminology – who truly pursue the common good of the entire polity, a high court that becomes the mouthpiece of public opinion is not merely mediocre: it has blatantly failed its mandate. Of course, as Steven Winter keenly reminds us, this view of the court’s conduct is pure ideal theory. Not differently than in a soccer game, the referee is supposed to impartially evaluate the actual occurrence and gravity of fouls. Then, on the ground, there might be corrupt, incompetent, biased, unreliable referees and perhaps coaches, or team-managers or players who display more fairness, just as in politics we might observe legislators who protect the interests of future generations better than the appointed justices. Thus I would connect a court’s ability to represent and safeguard the interests of generations other than the present one less with the mindset of its members – history offers plenty of evidence of partisan courts – than with the nature of the mandate that the constitution assigns to it.

Third, let me briefly comment on Schupmann’s claim that

when unamendability is merely implicit and the amendment clause is supreme, any law may be revised or even abrogated legally. Even the most basic constitutional law sits under the reservation of the exercise of the amendment clause. The electorate thus retains the legal power to use the amendment clause and amend or abrogate any aspect of its constitution – including democratic constitutional essentials. It seems to legally retain control over the exercise of constituent power, even as we recognize it does not have a legitimate claim to that power (Schupmann, 2024b, 1479 emphasis added).

This contention is crucial for Schupmann’s more general claim that my excessive reliance on implicit unamendability sets the stage for a tension between legality and legitimacy which my political liberal model of constitutional democracy had better stay clear of. Before addressing the objection, here it’s worth clarifying that for Rawls the term ‘legal’ cannot be understood in a formalistic sense, as ‘compliant with the letter of the law’. Evidence for this broader meaning comes from passages in which Rawls uses several times the adjective ‘valid’ as synonymous with ‘legally binding’. Let me quote them:

This [the fact that a particular understanding of the constitution may be mandated to the Court via a formal amendment] raises the question of whether an amendment to repeal the First Amendment, say, and to make a particular religion the state religion … must be accepted by the Court as a valid amendment. It is truistic to say … that if the people act constitutionally such amendments are valid. But is it sufficient for the validity of an amendment that it be enacted by the procedure of Article V? (Rawls 2005, 238)

Equally clear seems to me the fact that Rawls is not using the term ‘valid’ as synonymous with ‘legitimate’, as if repealing the First Amendment could be a fully legal act, assuming that the provisions of Article V were respected, albeit one of questionable legitimacy. By arguing that the Court should refrain from considering such amendment ‘valid’, Rawls is clearly meaning that the Court should declare it ‘legally void’, not binding. To conclude on this point: Schupmann’s claim that entrusting constitutional continuity on the notion of ‘implicit unamendability’ leaves the door ‘open legally for the electorate to cannibalize its constitution – even if on a different normative level it would be illegitimate for the electorate to do so’ (Schupmann, 2024b, 1479-1480), may well apply to some views of implicit unamendability, but not to Rawls’s or mine. This clarification is relevant for the discussion, upcoming next, of Schupmann’s claim that ‘implicit unamendability’ is an intrinsically weaker instrument – if compared with eternity clauses – for protecting constitutional democracy from its internal enemies.

Three reasons why implicit unamendability is not weaker than explicit unamendability

Moving on now to the more substantive question – is implicit unamendability a weaker tool than explicit unamendability for the purpose of ensuring the continuity of the constitutional essentials? – I would defend the view outlined in SAG (especially Ch 7.3) on the basis of three considerations.

First, as the passages quoted above indicate, Rawls’s view of implicit unamendability, as well as my own, that builds on his, in no way can be said to pave the way to a theoretically dubious and politically dangerous tension of legality and legitimacy, because amendments that violate the constitutional essentials are also ‘illegal’ – presumably after being so declared by a constitutional court.

Second, Schupmann maintains that because ‘they are nowhere enacted explicitly through positive statutes’, implicitly unamendable constitutional essentials are somehow ‘immaterial’ (Schupmann, 2024b, 1480). I am not sure what additional meaning the term ‘immaterial’ adds to the fact that a normative prescription is not part of positive law. ‘Immaterial’ seems to suggest negligibility, indeterminacy, and a smaller degree of bindingness – all characteristics that bear at best a loose, if any, connection with the absence of ‘positivization’. It’s hard to grasp how being ‘implicit’ correlates with being ‘immaterial’. In a section of Chapter 7 entitled ‘What can amending power not change?’ (Ferrara 261–264), five distinct kinds of ‘structural unamendables’ are listed. The fifth type illustrates our point. All human practices, including law and amending a constitution, take place against a shared background of assumptions, that everyone assumes that everyone else assumes as inviolable, simply not at the disposal of our human will. Take the law of gravitation. Certainly, no human law can be ‘valid’, in the sense of ‘legally binding’, if it enjoins its addressees to perform acts that violate the law of gravitation, even though nowhere in any constitution it is positively stated that statutes cannot be valid if they oblige people to act contrary to the law of gravitation. There is simply no need for such a specification because the inviolability of gravitation is a universally known fact that no one would question. Does that make gravity ‘immaterial’?

Lack of positivization leaves constitutional essentials up for authorized constitutional interpreters to fix, with ample margins for error, and the danger looms large – as Schupmann emphasizes – that populist forces may colonize those judicial benches and populate them with their acolytes. Surely the judicial authority to determine what is essential to a constitution ‘can be exercised arbitrarily’ (Schupmann, 2024b, 1480). Once again, errors are possible in refereeing any human activity, and that ubiquitous possibility does not detract from the sensibleness of conceiving of the function of refereeing as completely separated from that of playing the game. The point of dissent reaches philosophically deeper than it seems. Of course, there are ‘super-legal [i.e., non-positive] constitutional norms’ that ‘hold normative priority over positively enacted articles of the constitution’ (Schupmann, 2024b, 1480). Constitutions are no exception. We know from Gödel that no formal system can contain all its basic axioms within itself, and we know from Wittgenstein that not all that is implicit in a practice’s normative background can be made fully explicit. At some point in our attempt to make the implicit explicit, we run out of ‘rule-grounding rules’ (in our case: higher norms that justify constitutional essentials). We reach a customary bedrock that prevents our legal-philosophical, transcendental spade from proceeding further and causes it to turn. Were we to assume the existence of a philosopher’s stone that enables a single observer to tell whether a rule has indeed been followed, we would instantly forfeit the crucial distinction between a rule’s having been followed and being believed to have been followed. In the practice called ‘constitutional democracy’, at that juncture where our reconstructive spade turns, we find appointed interpreters of the essentials of the constitution, called justices, whom we entrust to spell out those fundamentals for us – subject of course to a possible subsequent challenge on the part of us, the addressees of this pronouncement. The judicial interpreters may err, be incompetent, biased, or inadequate in various other respects, but the alternative of making them superfluous by making all the essentials of the constitutional-democratic practice fully explicit is simply impracticable until Wittgenstein’s theses on ‘following a rule’ are convincingly refuted.

Third, the alternative to implicit unamendability is less clear than it seems at first sight. Schupmann’s proposal could be charitably rephrased as follows: instead of relying on the dubious notion of ‘implicit unamendability’, a Rawlsian political-liberal program for constitutional democracy should strive to entrench as many constitutional essentials as possible via explicit ‘eternity clauses’. This strategy would make it much more difficult for populist and anti-democratic forces to disfigure the constitution through legal means, as emblematically Hitler did with the Ermächtigungsgesetz of 23 March 1933. Of the three advantages mentioned by Schupmann – that is, explicit unamendability ‘aligns legality with legitimacy, reduces ambiguities, and limits judicial discretion’ – only the second and the third hold, but his proposal incurs a difficulty that calls for clarification. Should we understand Schupmann’s plaidoyer for explicit unamendability in the sense that only substance covered by eternity clauses – for example, article 89 of the Constitution of France and article 139 of the Constitution of Italy, both barring ‘the republican form’ from amendability – is legally entrenched and all the rest is subject to the will of the voters, provided that it respects the provisions for amending the constitution? That view cannot be simply assumed to be true but needs to be convincingly justified against not only what Rawls has to offer on the implicit unamendability of the First Amendment, but in relation to a blossoming literature on ‘unconstitutional constitutional amendments’. ⁶ And if, for the sake of the argument, we conceded that only eternity clauses ensure effective unamendability, what would Schupmann’s proposal entail? Should a worldwide liberal campaign be called for buttressing all existing democratic orders by introducing as many fundamentals-protecting eternity clauses as locally feasible? Should we understand Schupmann’s proposal as implying that only from now on all newly enacted liberal-democratic constitutions should entrench as many of their constitutional essentials via eternity clauses? All these implications seem to need further clarification.

On containing the unreasonable: How militant can militant democracy (and political liberalism) be?

Finally, in his sympathetic comments and in Schupmann (2024a), aimed at further strengthening the notion of constitutional democracy ‘across generations’, Schupmann suggests revising the idea, articulated by Rawls in the ‘Introduction’ and other passages of Political Liberalism, of containing the unreasonable. I certainly endorse the idea of containing the unreasonable, though I believe that such a notion has to be used as the last resort, when the two tools (outlined in The Democratic Horizon and recalled above in my response to Peter Niesen) of (a) finetuning ‘political justification’ via conjectural arguments designed to make the liberal essentials not weaker, but more palatable for non-liberal constituencies ubiquitously present in liberal-democratic polities, and (b) allowing for prudential, as opposed to principled, endorsement of one or more constitutional essentials on the part of less than fully reasonable actors ⁷ – have run their course without any positive result. And I also fully endorse Rawls’s sobering statement to the effect that

in the long run a strong majority of the electorate can eventually make the constitution conform to its political will. This is simply a fact about political power as such. There is no way around this fact, not even by entrenchment clauses that try to fix permanently the basic democratic guarantees. (Rawls 2005, 233)

An additional specification of ‘containment’, offered by Schupmann, consists of barring blatantly unreasonable actors from eligibility for office. In his words, ‘denying antidemocrats the right to hold public office complements the protective function of unamendability, providing a defense against both antidemocratic laws and actors’ (Schupmann, 2024b, 1483). While I see no objection in principle against such an idea, I wonder if Schupmann isn’t underestimating the prohibitive difficulties raised by its implementation. Who is to grant candidates a certificate of eligibility? If Schupmann has courts in mind, that solution is even more susceptible than judicial review to being perceived as antidemocratic, and even more likely to breed resentment and populist mobilization. Even assuming that the ‘who’-question can be appropriately answered, then how long should ineligibility last? Certainly not forever. Could it be appealed? Could it be reviewed, and by whom? Should it undergo an automatic renewal if the conditions warranting it don’t undergo significant change? Should ineligibility be general or office-specific? Should a candidate ineligible for legislative and executive offices be barred from a profession that offers access to public roles? Should ineligibility entail unappointability to non-elective offices? All these questions suggest that further reflection – not in the least to be understood as a euphemism for belated rejection – is needed, before including such extreme measures of militant democracy within political liberalism.

These few points on which we differ, in the end, do not detract from my gratitude to Ben Schupmann for the chance that his careful remarks have offered me to rethink important aspects of the argument of Sovereignty Across Generations.