Full Article Text
1. Introduction
Transitional justice has long privileged backward-looking mechanisms such truth-telling, prosecutions, and reparations over the forward-looking task of ensuring that mass violence does not recur. Guarantees of non-repetition (GNR) occupy that comparatively neglected space. Unlike measures that redress past harm, GNR target the institutional architecture that made harm possible; and among the instruments available, constitutional reform is the most ambitious, because by restructuring power, entrenching rights and redefining the political community it purports to write the principle of ‘never again’ into the permanent law of the state.[1]
Bangladesh after the mass uprising of July 2024 presents a rare live instance of the attempt. The fall of Sheikh Hasina’s government, after the killing of more than a thousand demonstrators, produced not merely a change of administration but an explicit constitutional project.[2] The interim government under Professor Muhammad Yunus, sworn in on 8 August 2024, established a Constitution Reform Commission chaired by Professor Ali Riaz, which reported on 15 January 2025.[3] A National Consensus Commission (NCC) was then formed on 12 February 2025 to translate the reform proposals into a cross-party settlement; its work produced the July National Charter 2025, a consolidation of eighty-four proposals, roughly half of them constitutional, adopted on 17 October 2025.[4] A Presidential Implementation Order of 13 November 2025 authorised a national referendum, held on 12 February 2026 concurrently with the thirteenth parliamentary election, at which the Bangladesh Nationalist Party secured a two-thirds supermajority and sixty-one percent of voters endorsed the reform package.[5]
The framing of the exercise was, from the outset, that of non-repetition. The Chief Adviser cast constitutional and electoral reform as essential to halting the recurring cycles of political violence and authoritarianism that had marked Bangladesh’s history, and the Charter’s own preamble invokes the sacrifices of July 2024 as the ground of a new settlement designed to prevent the return of ‘fascist’ rule.[6] The question this article addresses is whether the doctrinal means chosen to secure that guarantee are adequate to it.
A body of able commentary has already formed around the reform, and this article defines itself partly against it. Ridwanul Hoque has assessed the Constitution Reform Commission’s proposals as pragmatically doubtful, questioning whether mega-structural changes such as bicameralism are implementable in a unitary state.[7] Arafat Hosen Khan has argued, most fully, that the process squandered its ‘constitutional moment’ through structural exclusion—an all-male consensus body, the exclusion of the Awami League and other formations and the entrenchment of that exclusion through the later statutory ban of the League, drawing on Ackerman, on Elkins, Ginsburg and Melton and on Amal Sethi’s account of equitable elite bargaining.[8]
These accounts share an analytical centre: the inclusiveness of the constituent process, assessed against theories of legitimacy and endurance. The argument of this article is different, and deliberately narrower. It relocates the critique from the exclusivity of the process to the doctrine of entrenchment itself. Its claim is that the Charter’s substantive design is a credible guarantee of non-repetition, but that the mechanism on which any such guarantee must ultimately rely—entrenchment beyond majoritarian revision—has in Bangladesh’s own constitutional history been the instrument through which abusive and partisan settlements were locked in rather than dismantled. The reform therefore confronts an entrenchment paradox: to make ‘never again’ durable, it must entrench; yet entrenchment adopted through a narrow constituent process, and immunised against later revision, is precisely how each of Bangladesh’s previous ‘never agains’ became the next generation’s grievance. The paradox is doctrinal, not merely political, and it is visible in three specific features of the Charter and in the enforceability of its rights guarantees.
The argument proceeds as follows. Part II sets out the minimal theoretical apparatus, situating constitutional reform within GNR and identifying entrenchment as its characteristic device. Part III reconstructs the Charter’s entrenchment architecture and reads it, sympathetically, as coherent non-repetition. Part IV develops the paradox through Bangladesh’s amendment jurisprudence—the eternity clause of 2011 and the reversals of 1975. Part V isolates three doctrinal defects specific to the Charter. Part VI turns to enforceability. Part VII states the strongest realist counter-argument, and Part VIII replies. Part IX concludes.
2. Guarantees of Non-Repetition and the Constitutional Form
The category of guarantees of non-repetition entered the transitional-justice lexicon as the fourth of the state’s obligations toward victims of gross violations, alongside truth, justice and reparation. Its defining orientation is prospective. Where prosecutions answer for what was done and reparations repair it, GNR ask what institutional conditions permitted the wrong and how those conditions may be altered so that it does not recur. As Lutz observes of the comparative record, the preventive goals of transitional justice are the most numerous and the most various: preventing former perpetrators from reasserting power, reconstructing a justice system that offers a legitimate alternative to violence and involving all sections of society in the design of the new order so that no community’s exclusion becomes the seedbed of the next conflict.[9]
Constitutional reform is the most structurally ambitious GNR instrument because it operates on the architecture of power rather than on individual conduct. A constitution can disperse authority that was previously concentrated, entrench rights that were previously vulnerable and constitute a political community on terms designed to include those whom the prior order excluded. The United Nations’ own formulations of GNR list, among the guarantees, civilian control of security forces, the independence of the judiciary and the reform of laws contributing to violations, each of which is, in the first instance, a constitutional question.[10]
That mapping is not merely rhetorical. As Juan Méndez has argued, the four transitional-justice obligations: truth, prosecution, reparation and institutional reform each translate into constitutional norms and processes, so that a constitution is one of the principal sites at which a society discharges or defaults on its non-repetition duties.[11] Comparative scholarship has begun to taxonomise the forms this takes. Cheng-Yi Huang distinguishes three models of ‘constitutionalizing transitional justice’: written-in constitutions, in which non-repetition values are codified in the text itself, as in South Africa’s 1994 interim epilogue, Poland’s ‘never again’ preamble, or Article 66 of Colombia’s Constitution; court-enforced constitutionalism, in which courts infer transitional-justice mandates from an otherwise silent text; and unilateral executive decisions, in which the chief executive dictates the mechanism without legislative participation.[12] Huang’s own conclusion is pointed: the third, executive-decree model carries ‘the weakest legitimacy of all three,’ because its guarantee rests on a unilateral act that the executive may as handily amend as issue. That ranking becomes directly relevant to the Bangladeshi case in Part V.[13]
Yet constitutional GNR carries a distinctive difficulty absent from truth commissions or trials. A guarantee written into ordinary constitutional text is only as secure as the next amendment. The very majority whose capture of the state produced the original violations may, if it returns to power, unwind the guarantee by the same amendment power that any government commands. The classic response is entrenchment: raising the threshold for amendment, subjecting core provisions to referendum or placing certain commitments beyond amendment altogether through an eternity clause. Entrenchment is what converts a reform from a policy that a later majority may reverse into a guarantee that binds it. It is, in other words, the doctrinal device on which the durability of constitutional non-repetition depends.[14]
Whether entrenchment can actually secure a transitional settlement is, however, contested precisely in the transitional context. Gábor Halmai has asked directly whether unamendable provisions in transitional constitutions can protect the integrity of a post-transition order, and answers with heavy qualification: where a society lacks an established constitutional culture, elite-driven entrenchment may fail to prevent—and can even furnish the legal vocabulary for—subsequent backsliding, as the Hungarian experience shows.[15] This is the theoretical seam the present article works: entrenchment is indispensable to durable constitutional GNR and yet, in polities of thin constitutional culture, is exactly the device through which a later dominant actor captures or unwinds the guarantee. The quality of a reform’s non-repetition guarantee therefore turns on the quality of its entrenchment on what is entrenched, by whom, and through what process.
Spelled out, that distinction is the pivot of the argument: a guarantee entrenched by a broadly constituted authority, protecting genuinely foundational commitments and remaining open to inclusive revision, advances non-repetition. A guarantee entrenched by a narrow body, protecting a partisan settlement, and immunised against later revision, does the opposite: it locks in the conditions of future grievance and removes the ordinary constitutional means of addressing them. The remainder of this article applies that distinction to the July Charter.
3. The Charter’s Entrenchment Architecture as Non-Repetition
Read against its own diagnosis, the Charter’s substantive design is coherent GNR. The pathology it identifies is the concentration of power in a majoritarian executive able to colonise nominally independent institutions—the judiciary, the Election Commission, the Anti-Corruption Commission—and to distort the electoral and constitutional order to perpetuate single-party dominance.[16] Its remedies track that diagnosis, and several are explicitly counter-majoritarian.
The most important, for present purposes, is the heightened amendment rule. The Charter provides that constitutional amendment shall require a two-thirds majority in the lower house together with a majority in the proposed upper house, and that amendments to a defined core—including Articles 8, 48, 56 and 142, and the reintroduced caretaker-government provisions shall additionally require a national referendum.[17] This is the entrenchment limb proper: it seeks to remove the constitutional core from the reach of any single dominant party, answering the historic complaint that a government commanding a bare two-thirds of a unicameral parliament has enjoyed decisive and unchecked authority over constitutional change.[18]
Reinforcing this are three further counter-majoritarian devices. Bicameralism introduces a one-hundred-member upper house elected by proportional representation on the basis of national vote share, deliberately importing a veto point and minority representation that a single lower-house majority cannot easily override.[19] The dispersal of appointment power replaces prime-ministerial control over constitutional bodies with bipartisan selection committees, drawing in the Speaker, the Leader of the Opposition, and the judiciary for the Election Commission, the Ombudsman, the Public Service Commissions, the Comptroller and Auditor General and a constitutionalised Anti-Corruption Commission.[20] And the revived caretaker-government system, to be entrenched behind the referendum lock, reinstates a non-partisan election-time administration whose abolition in 2011 is widely regarded as having precipitated the electoral-credibility crisis that culminated in 2024.[21]
Taken together, these are not symbolic gestures. They are hard institutional constraints designed to prevent the recurrence of executive capture, and on the substantive axis the Charter is a credible attempt to encode ‘never again’ into enforceable structure rather than aspirational preamble. The Charter also gestures toward the accountability limb that GNR theory pairs with structural reform: proposal 64 provides for an independent commission to investigate officials involved in the killings and repression of the July uprising and in electoral fraud and corruption, situating a truth-and-accountability mechanism within the reform instrument itself.[22]
If the analysis stopped here, the verdict would be favourable: a well-diagnosed reform deploying the standard counter-majoritarian toolkit, entrenched against the very majoritarianism it condemns. The difficulty emerges only when the entrenchment mechanism is examined against Bangladesh’s own constitutional record.
4. The Entrenchment Paradox in Bangladesh’s Amendment Jurisprudence
The claim of this Part is that entrenchment, in Bangladesh, has not functioned as a neutral preservative of foundational commitments. It has functioned as an instrument of capture—a means by which a dominant actor fixes its own settlement in place and forecloses later revision. Two episodes establish the pattern, and both bear directly on the Charter.
4.1 The Eternity Clause of 2011
The basic-structure doctrine was transplanted into Bangladeshi law in 1989, when the Appellate Division in Anwar Hossain Chowdhury invalidated the Eighth Amendment as inconsistent with the Constitution’s basic structure.[23] In 2011 the Fifteenth Amendment went further, converting judicial doctrine into textual entrenchment: Article 7B rendered a large portion of the Constitution—by some estimates more than a third—permanently unamendable. On its face this is the strongest possible non-repetition device, an eternity clause protecting the constitutional core against future majorities.[24]
Yet the manner of its adoption inverts that appearance. The Fifteenth Amendment was passed by a governing party commanding a two-thirds majority; it scrapped the caretaker-government system without a referendum, though a referendum requirement was then in force; the parliamentary select committee that considered it sat without the principal opposition party, which boycotted; and the committee’s recorded position in favour of retaining the caretaker system was reversed, on the record, after a reported meeting with the prime minister. Bangladesh’s leading constitutional scholars have concluded that the entrenchment was itself of doubtful legitimacy—adopted through a narrow, majoritarian process, without the popular participation that alone could justify placing provisions beyond democratic revision, and without any mandate for the eternity clause in the governing party’s election manifesto.[25]
The lesson is precise. Entrenchment did not protect a foundational consensus; it fixed one party’s contested settlement in place and, by doing so, removed from later majorities the ordinary constitutional means of undoing it. Yaniv Roznai’s distinction between facile and demanding amendment powers clarifies why this matters. An amendment procedure is ‘demanding,’ in the participatory sense, only when it draws the people meaningfully into constitutional change; a procedure that raises formal thresholds while excluding popular deliberation remains, in the sense that counts, facile—even when it is difficult to satisfy.[26] The 2011 eternity clause was formally demanding and democratically facile at once, and it is precisely that combination—high threshold, thin participation—that made it an instrument of entrenchment-as-capture rather than of non-repetition.
The instability of the device is confirmed by its fate. In December 2024 a High Court bench declared Articles 7A, 7B and 44(2) of the Fifteenth Amendment void as inconsistent with the basic structure, and revived the Article 142 referendum requirement for amendments; the abolition of the caretaker system was reversed, and the caretaker system itself restored on review in November 2025.[27] That an eternity clause—the strongest entrenchment an order can deploy—was itself struck down within a few years of a change of regime is not a reassurance but a warning. It shows that in Bangladesh entrenchment does not settle a question beyond politics; it merely relocates the political contest to the moment of capture or the moment of judicial undoing. A guarantee of non-repetition that relies on entrenchment in such a system relies on a device that has repeatedly proven neither durable nor neutral.
4.2 The Reversals of 1975
The second episode shows the pattern operating in the domain of accountability rather than structure, and over time rather than in a single enactment. Bangladesh’s first transitional-justice settlement after 1971 was substantial: special tribunals were established, tens of thousands were charged under a Collaborators Order and the International Crimes (Tribunals) Act was enacted in 1973.[28] Within a few years that settlement was dismantled. Following the assassination of Sheikh Mujibur Rahman in August 1975, the tribunals were disbanded, detainees pardoned, the Collaborators Order repealed and an Indemnity Ordinance enacted to shield those responsible for the assassination from prosecution. The International Crimes Tribunal was not revived until 2010, under a returning Awami League government—whereupon it attracted sustained criticism for departures from fair-trial standards and for its use as a partisan instrument.[29]
The 1975 reversals are not an entrenchment episode in the technical sense, but they establish the temporal dynamic that makes the entrenchment paradox acute. Each Bangladeshi settlement designed to secure ‘never again’—the 1972 order, the 1973 tribunals, the 2011 restoration of the 1972 principles—was undone or captured by the succeeding majority, whether by repeal and indemnity or by entrenchment and instrumentalisation. The recurrence the reforms sought to prevent was, repeatedly, a recurrence enacted through the constitutional and legal forms themselves.
The rhyme is not merely historical. The interim government’s own protection order for those who acted against the prior regime between July and August 2024 was converted, by the elected Parliament in April 2026, into the July Mass Uprising (Protection and Liability Determination) Act—an instrument that distinguishes protected ‘political resistance’ from criminal conduct and gives the earlier indemnity measure permanent statutory form.[30] Whatever the merits of protecting those who resisted repression, the form is familiar: a transition once again secures its preferred account of lawful and unlawful conduct through an indemnity-type statute, exactly as the 1975 settlement did in the opposite political direction. The point is not that the two indemnities are morally equivalent—they are not—but that the technique of fixing a transition’s account of justice through entrenchment and indemnity recurs across regimes of opposite character. That recurrence is precisely what should give the designer of a non-repetition guarantee pause.
4.3 The Comparative Pattern: Hungary and Colombia
The paradox is not a Bangladeshi peculiarity; comparative experience supplies both a warning and a qualification. The warning is Hungary. Its post-1989 transition entrenched a rule-of-law order, and successive measures wrote transitional-justice commitments into constitutional form; yet an established constitutional culture was absent, and once Viktor Orbán’s party secured a two-thirds majority it rewrote the Fundamental Law—‘in the name of transitional justice’—deploying the language and forms of the transitional settlement to dismantle the liberal constitutionalism the settlement was meant to secure.[31] Hungary is the entrenchment paradox in its purest comparative form: a settlement entrenched in a polity of thin constitutional culture became not a barrier to capture but the instrument of it, wielded by the very majoritarian actor it was meant to constrain.
The qualification is Colombia. There the 2012 ‘Legal Framework for Peace’ amendment (Article 66) wrote transitional justice into the constitutional text, and the Constitutional Court, drawing on the victims’ right to truth and on inter-American jurisprudence, both legitimated the negotiated settlement and policed its limits—upholding the peace framework while striking down attempts to convert it into a ‘blank cheque’ to remake the political system.[32] What made Colombian entrenchment a guarantee rather than a trap was not the entrenchment itself but the surrounding conditions Huang identifies: a codified, court-reviewable mandate and a ‘deeply rooted legalist culture’ in which the adversarial parties complied with judicial limits on the settlement. The contrast is instructive for Bangladesh. Entrenchment secures non-repetition where a strong, independent court can hold the settlement to its own terms and where the political actors accept that discipline; it becomes capture where the court’s authority is itself contested and where the settlement is immunised against judicial review. The Bangladeshi Charter, as the next Part shows, sits on the wrong side of that line on both counts—precisely the constellation Halmai identifies as the condition under which transitional entrenchment fails.
This is the sense in which the Charter confronts a paradox rather than a mere risk. To make its guarantee durable it must entrench; but entrenchment in Bangladesh has been the medium of capture, and the more effectively the Charter entrenches, the more it forecloses the very revision by which a later, more inclusive majority might correct whatever the founding moment got wrong. The next three defects show the paradox operating in the Charter’s own text.
5. Three Doctrinal Defects Specific to the Charter
5.1 The Self-immunising Validity Clause
The Charter’s Pledge of Commitment contains a provision without obvious precedent in a document of this kind. The signatories undertake that they shall not, under any circumstance, raise before any court any question regarding the validity or necessity of the Charter, and shall ensure its full legal and constitutional protection at every stage of implementation.[33] The clause is a form of entrenchment—but of the most self-undermining kind. A guarantee of non-repetition is supposed to restore legality and justiciability after a period in which the executive placed itself above the law. Here the foundational instrument seeks to place itself beyond judicial scrutiny, foreclosing in advance the constitutional adjudication through which any rule-of-law order tests the validity of its own foundations.
The difficulty is doctrinal, not merely symbolic. Bangladesh’s constitutional order rests on the supremacy of the Constitution under Article 7 and on judicial review as the mechanism that gives that supremacy effect.[34] A clause purporting to bar any court from questioning the Charter’s validity is in tension with that structure at its root. It also sits awkwardly beside the Charter’s own commitment to judicial independence and to the restoration of the rule of law: an instrument that constitutes an independent judiciary while simultaneously instructing that judiciary never to examine the instrument’s own validity guarantees non-repetition with one hand and withdraws the means of enforcing it with the other. Whatever else it is, this is not the demanding, participation-anchored entrenchment that legitimate constitutional GNR requires; it is entrenchment by pre-emptive immunisation.
The concern deepens when one attends to the instrument through which the Charter’s constitutional content was to acquire legal force. The substance was routed through a Presidential Order, the July National Charter (Constitutional Reform) Implementation Order 2025 which recited its own authority as flowing from the sovereign will expressed in the uprising, and which attached to the newly elected members a second oath as members of a Constitution Reform Council empowered to complete the amendments within a defined period.[35] Senior constitutional counsel warned that amending the Constitution through a presidential order would set a dangerous precedent, drawing the pointed comparison with General Yahya Khan’s Legal Framework Order of 1970, an executive instrument that purported to prescribe the terms on which a constituent body might act.[36] The objection is doctrinal, and it compounds the validity-clause problem: a guarantee of non-repetition against executive over-reach, entrenched by an executive order whose own constitutional footing is contested and immunised against the courts that would ordinarily test it, borrows the very form, rule by decree over the constituent process that the reform exists to foreclose.
Placed within Huang’s comparative typology, the significance of the instrument sharpens. A reform whose constitutional content is carried by a Presidential Order, ratified by a bundled plebiscite and shielded from judicial review is not a written-in constitutionalisation of the South African or Colombian kind, in which a codified, court-reviewable mandate anchors the settlement; nor is it court-enforced constitutionalism, since the courts are the very institution the validity clause seeks to exclude. It sits closest to the third, executive-decree model—the model Huang identifies as carrying the weakest legitimacy of the three, because its guarantee rests on a unilateral act the executive may amend as handily as it issues.[37] That a non-repetition guarantee against executive over-reach should be delivered through the model of constitutionalising transitional justice with the weakest legitimacy—the unilateral executive act—is not a peripheral irony. It is the entrenchment paradox appearing at the level of institutional form: the reform reaches for the executive-decree route because it is the fastest way to entrench, and in doing so reproduces the very concentration of constituent authority that non-repetition was meant to disperse.
5.2 The Bundled Plebiscite
The referendum of 12 February 2026 was, in form, the popular participation that the 2011 entrenchment lacked. In substance it reproduced the same deficit. The Implementation Order condensed the Charter’s constitutional proposals into four broad questions, which voters were required to answer as a single package rather than severally.[38] A single ‘Yes’ recorded at sixty-one percent was thereby taken to endorse a bundle of distinct and independently contestable constitutional choices, while binding future governments to the reform agenda as adopted.[39]
Assessed through Roznai’s lens, the plebiscite is demanding in threshold but facile in deliberation. It supplies a majoritarian ratification for a project whose stated purpose was to constrain majoritarianism, and it does so through precisely the mechanism: an aggregate, take-it-or-leave-it popular vote that maximises the appearance of participation while minimising its deliberative content. Bundling foreclosed the disaggregated public reasoning that alone would make the entrenchment democratically ‘demanding’ in the sense that legitimacy requires.[40] The 2011 amendment entrenched without a referendum; the 2026 settlement entrenches with a referendum whose design drains the exercise of deliberative force. The form differs; the participatory deficit is continuous.
The deliberative thinness was compounded by a neutrality deficit. The interim government, whose defining claim was to non-partisan stewardship, actively campaigned for a ‘Yes’ vote, prompting the observation from senior constitutional counsel that it was wrong for the government to campaign for one side in a referendum it was administering.[41] A ratifying plebiscite loses much of its legitimating force where the administering authority is not neutral between the options, and the point sharpens the doctrinal difficulty: the very body invoking popular sovereignty to entrench the settlement placed its thumb on the scale of the vote through which that sovereignty was said to speak.
5.3 The Partisan Framing Of Justice
The Charter’s account of the wrong it exists to prevent is, in places, explicitly partisan. Its Pledge commits the signatories to justice for the victims of ‘Awami fascism’ and for killings committed by the ‘fascist Awami League,’ and the settlement was negotiated and declared with the Awami League excluded, the party having been banned under anti-terrorism legislation.[42] Khan has traced the subsequent conversion of the executive ban into statute by the incoming BNP-led legislature, describing the sequence as ‘weaponised legalism’—the encoding of political exclusion into the legal architecture itself.[43] The subsequent record bears the description out with precision. When the ordinance banning the League was placed before the elected Parliament, the relevant committee recommended it for passage in amended form; Parliament nonetheless enacted the Anti-Terrorism (Amendment) Bill 2026 with no change to its substance, retaining the power to ban all activities of an entity. The party that had, during the interim period, opposed banning any political party by executive order thus moved, once in government, to fix that same ban into permanent statute—unchanged, and over its own committee’s recommendation to amend it.[44]
This article accepts Khan’s description of the phenomenon and locates it within the entrenchment paradox rather than within a theory of the constitutional moment. The point, in doctrinal terms, is that a ‘never again’ defined against a named party is not a general guarantee against the recurrence of atrocity but a specific guarantee against the recurrence of one set of perpetrators—and, entrenched, it fixes that partisan definition into the constitutional foundation. Comparative experience with post-authoritarian settlements that exclude the former ruling party suggests two characteristic failure modes: the excluded party’s eventual unconstitutional return, or the new dominant party’s instrumentalisation of the settlement for factional advantage.[45] Either way, a guarantee framed as ‘not them again’ rather than ‘never again’ invites its own future contestation, and entrenchment ensures that the contestation, when it comes, must take an extraordinary rather than an ordinary constitutional form.
6. The Enforceability Limb: Entrenching the Unenforceable
Entrenchment secures a guarantee only if the guarantee is itself enforceable. A right that cannot be vindicated in court gains nothing from being made difficult to amend; entrenchment merely fixes its unenforceability in place. On this axis the Charter reproduces a long-standing structural weakness of the Bangladeshi Constitution that is directly relevant to non-repetition.
The Constitution divides its rights provisions between the Fundamental Principles of State Policy in Part II, which Article 8(2) renders judicially unenforceable, and the enforceable fundamental rights of Part III. Economic, social, cultural and environmental rights sit on the unenforceable side of that line, functioning as directives to the state rather than as claims a citizen may press.[46] The Constitution Reform Commission had proposed to collapse this distinction, unifying Parts II and III into a single enforceable charter of ‘Fundamental Rights and Freedoms’ subject to a general limitation clause and a proportionality test. The July Charter did not adopt that proposal. In its place it retained only a broadly worded commitment that the state shall take measures to expand, protect and implement citizens’ fundamental rights—language whose vagueness leaves the enforceability question unresolved and, on the better reading, leaves socio-economic rights where they were: unenforceable.[47]
The significance for GNR is direct. Several of the structural conditions implicated in the 2024 crisis—inequality, the denial of economic and social entitlements, the vulnerability of marginalised communities—are precisely the conditions that enforceable socio-economic rights would address, and precisely the conditions that non-repetition, properly understood, must reach. A reform that entrenches its counter-majoritarian architecture behind a referendum lock while leaving its rights guarantees in the non-justiciable domain entrenches the enforceable and leaves the unenforceable unenforceable. It is a selective entrenchment, and the selection tracks the distinction between the interests of political elites, whose competition the structural provisions regulate, and the interests of citizens, whose social entitlements the rights provisions would secure. That asymmetry is itself a feature of the paradox: entrenchment is deployed where it protects the settlement among parties, and withheld where it would protect the population against the parties.
7. The Counter-Argument: Process Realism in Transition
The strongest objection to this analysis is realist, and it deserves statement at full strength. Transitions are not constitutional seminars. They occur under acute time pressure, incomplete information and the standing threat of counter-mobilisation or renewed violence. On this view three replies blunt the paradox.
First, entrenchment is the price of any guarantee at all. A reform that declined to entrench, for fear of locking in the founding moment’s imperfections, would leave the concentrated executive powers that produced the 2024 crisis fully available to the next majority. Between an imperfectly entrenched guarantee and no guarantee, the former prevents more future harm. The Charter’s architects themselves describe it as one step in a longer journey rather than a completed settlement, which implies that later, more inclusive constituent stages can revisit what the founding moment fixed.[48]
A fourth reply, deeper than the other three, comes from the theory of transitions itself. Colleen Murphy locates the distinctiveness of transitional contexts in two forms of uncertainty that are pervasive in transitions and rare in consolidated democracies: existential uncertainty and uncertainty about authority\u2014about who is entitled to speak and decide for the political community.[49] On this account, the drive to entrench, to immunise and to lock in is not mere overreach but a rational response to radical uncertainty about authority: an interim body with contested legitimacy reaches for the strongest available devices precisely because its authority to decide is in doubt. Judith Heimbach’s study of Libya adds the practical corollary that, where a transition is fragile and divided, pressing constitution-making and transitional-justice measures too hard can deepen division rather than resolve it, and that holding back may sometimes be the wiser course.[50] Together these suggest that the Charter’s hard entrenchment may be the understandable product of a body acting under uncertainty about its own authority, and that the alternative—declining to entrench at all—carries its own risks of drift and reversal.
Second, some exclusions are themselves guarantees of non-repetition. Excluding a party credibly associated with the prior repression is, on this account, not a betrayal of ‘never again’ but an instance of it—closer to lustration or militant democracy than to erasure. A constitutional order may legitimately defend itself against those who would dismantle it, and the bundled referendum can likewise be defended as a means of preventing well-resourced spoilers from unpicking an interdependent package provision by provision.[51]
Third, entrenchment self-corrects over time. Because the elected parliament sits, under the Implementation Order, as a Constitution Reform Council charged with completing the amendments within a defined period, the initial narrowness of the constituent body is arguably transitional: future majorities inherit a structure that constrains them regardless of who was in the room at the founding, and the referendum lock protects the reform from casual reversal while ordinary politics resumes.[52]
8. Reply: Why the Defects Are Not Incidental
These replies are serious, and each is partly right. But each establishes only that the Charter’s entrenchment defects were understandable; none establishes that they were costless for non-repetition specifically. The distance between those two propositions is where the paradox lives.
The ‘price of any guarantee’ argument proves too much. If any degree of entrenchment-through-narrow-process is licensed by the risk that a fuller process would fail, then GNR imposes no constraint on constituent method at all, and ‘never again’ reduces to whatever the strongest transitional bloc can entrench. The distinctive claim of constitutional GNR is that how a society entrenches its guarantee is part of the guarantee, because an entrenched settlement adopted over the exclusion of a major community writes that community’s grievance into the foundation and then removes the ordinary means of addressing it. The self-correction reply concedes the premise it needs to deny: if the settlement genuinely required later inclusive revision, the referendum lock and the validity clause are the two features most calculated to obstruct it.
The uncertainty-of-authority reply explains the impulse to entrench without excusing its form. That an interim body reaches for strong devices under uncertain authority is understandable; but the response to uncertain authority that a non-repetition settlement requires is to broaden the constituent base so that the settlement can be attributed to the community rather than to a faction, not to narrow it and then seal the result against challenge. Heimbach’s Libyan study cuts the other way from the reply it is enlisted to support: her lesson is that selective transitional justice tends to spill into selective constitution-making, each narrowing the other in a spiral that deepens division.[53] A guarantee framed against a named party, negotiated over the exclusion of that party and of the women’s commission, and then entrenched and court-proofed, is the selective-into-selective spiral in constitutional form. The correct inference from transitional uncertainty is not that entrenchment excuses exclusion, but that exclusion plus entrenchment is the most dangerous of the available combinations, because it fixes a contested authority-claim as if it were settled.
The lustration reply is sound for the banned party and unsound as a general defence. There is a coherent militant-democracy rationale for excluding, or conditioning the return of, a party credibly responsible for atrocity; that rationale is doctrinally distinct from erasure and may properly inform a non-repetition settlement. But it does not extend to the exclusion of the women’s commission from the constituent body, for which no self-defence rationale exists, nor to the bundling of unrelated constitutional questions, nor to the immunisation of the Charter against judicial review.[54] Those defects are not instances of a constitutional order defending itself; they are instances of the founding settlement insulating itself from scrutiny and from inclusive revision. The militant-democracy frame, invoked to justify the one exclusion for which it fits, cannot be stretched across the others.
The self-correction reply is the most powerful, and it is double-edged in exactly the way Bangladesh’s history predicts. Entrenchment locks in whatever was decided at the constituent moment, including its exclusions; and the higher the amendment threshold, the harder it becomes for a later, more inclusive majority to revisit a settlement struck by a narrow founding body. This is the through-line from Part IV. The 2011 eternity clause was defended, too, as a protection of foundational commitments; it functioned as the entrenchment of a contested, party-specific settlement adopted without popular deliberation, and it removed from later majorities the constitutional means of correction. A bundled, court-proofed ‘Yes’ that binds future governments to the reform agenda as adopted does not cure that pattern; it constitutionalises it. Entrenchment is a virtue for the substance the Charter gets right and a liability for the exclusions it carries, and the same referendum lock serves both functions indiscriminately.
The self-correction reply also assumes a smoothly functioning Council; the early record is less reassuring. When the Constitution Reform Council was to be convened, several formations—including Jamaat-e-Islami and others—declined the Council oath, and senior counsel argued that the mechanism attaching a second, Council oath to elected members by presidential order was, in substance, without valid legal foundation, describing the instrument as neither ordinance nor law.[55] A self-correcting mechanism whose very convocation is contested at the first opportunity is a thin reed on which to rest the claim that entrenchment’s exclusions will be cured downstream. The more the entrenchment holds, the less the correction; and where the correction depends on a body whose own legal footing is in doubt, the reassurance the reply offers is largely notional.
It is worth marking precisely where this argument parts company with the existing commentary. Khan’s critique is framed in the register of legitimacy and endurance: the process was insufficiently inclusive to constitute ‘the People,’ and constitutions built through exclusion endure less well.[56] That is persuasive on its own terms, but it is an argument about the legitimacy of the process. The argument here is about the doctrine of entrenchment: even granting, for the sake of argument, that the process was as legitimate as a transition allows, the entrenchment mechanism on which the guarantee depends has in Bangladesh reliably converted such settlements into instruments of capture, and the Charter’s specific entrenchment devices—the validity clause, the bundled lock, the partisan framing, the selective enforceability—reproduce that conversion in its own text. The two critiques are complementary, but the doctrinal one survives even where the legitimacy one is contested, because it does not depend on a theory of who ‘the People’ are; it depends only on the observable behaviour of entrenchment in this constitutional order.
The paradox therefore survives the counter-argument in modified form. The realist is right that some entrenchment was unavoidable and that some exclusion may have been defensible; the GNR analyst is right that these particular entrenchment choices—an unreviewable validity clause, a bundled plebiscite, a partisan definition of the wrong and an enforceability line drawn to protect elites rather than citizens—are not the neutral costs of transition but latent grievances now fixed in the foundational order, behind the highest amendment threshold Bangladesh has ever adopted. ‘Never again’ is weakened, not fulfilled, when it is entrenched by the doctrinal means through which Bangladesh’s previous guarantees were captured.
9. Conclusion
Post-July 2024 Bangladesh shows both the promise and the specific limit of constitutional reform as a guarantee of non-repetition. On substance, the July Charter is a genuine attempt to encode ‘never again’ into durable structure: it disperses appointment power, introduces counter-majoritarian friction through bicameralism and proportional representation, revives the caretaker system and raises the amendment threshold to protect the reformed core. These are the right instruments for the pathology the reform diagnoses.
But the durability those instruments require is supplied by entrenchment, and entrenchment is the point at which Bangladesh’s constitutional history turns against the reform. The eternity clause of 2011 and the reversals that began in 1975 show that entrenchment here has been the medium through which contested and partisan settlements were fixed in place and later inherited, rather than dismantled, by their successors. The Charter reproduces the pattern in its own text: a validity clause that seeks to place the foundational instrument beyond judicial review; a bundled plebiscite that entrenches without deliberation; a definition of the wrong framed against a named party; and an enforceability line that entrenches the structural settlement among elites while leaving the population’s socio-economic guarantees unenforceable. Each is a way of making the guarantee durable; each is also a way of making its defects durable.
The lesson is not that constitutional GNR is futile, nor that the Charter should not have entrenched. It is that the designers of a non-repetition settlement must treat the mode of entrenchment as part of the guarantee itself: what is locked, by whom, and whether the lock can later be opened by inclusive constitutional means. A guarantee against repetition entrenched through the doctrinal form most associated, in that polity, with capture carries the seed of the next rupture within its own foundation. For Bangladesh, whether the settlement of February 2026 becomes a durable ‘never again’ or a court-proofed ‘not them again’ will turn less on the elegance of its counter-majoritarian architecture than on whether the excluded the banned party where its return can be safely conditioned, the silenced women’s commission, and above all the citizens whose enforceable rights the Charter declined to entrench can be brought back into the constitutional settlement before the referendum lock sets. The paradox is that the very device meant to guarantee the future may be the one that forecloses its correction.
Footnotes
[1] Ellen Lutz, ‘Transitional Justice: Lessons Learned and the Road Ahead’ in Naomi Roht-Arriaza and Javier Mariezcurrena (eds), Transitional Justice in the Twenty-First Century: Beyond Truth versus Justice (CUP 2006) 325.
[2] Office of the United Nations High Commissioner for Human Rights, Human Rights Violations and Abuses Related to the Protests of July and August 2024 in Bangladesh (OHCHR, 12 February 2025).
[3] Constitution Reform Commission (Bangladesh), ‘Report’ (2025); Muhammad Ekramul Haque and Sumit Bisarya, ‘July Charter and Constitutional Reforms in Bangladesh’ (ConstitutionNet, 1 December 2025) <https://constitutionnet.org/news/voices/july-charter-and-constitutional-reforms-bangladesh> accessed 20 June 2026.
[4] July National Charter 2025 (ConstitutionNet) <https://constitutionnet.org/sites/default/files/2025-11/Bangladesh%20July%20National%20Charter%202025%20%28English%20translation%29.pdf> accessed 9 March 2026.
[5] July National Charter (Constitutional Amendment) Implementation Order 2025 (Presidential Order, 13 November 2025) (hereafter ‘Implementation Order’).
[6] July Charter (n 4), Introduction and ‘Historical Context’.
[7] Ridwanul Hoque, ‘Bangladesh at a Constitutional Crossroads: Reform or Overhaul?’ (NLSIR Online, 20 May 2025) <https://www.nlsir.com/post/bangladesh-at-a-constitutional-crossroads-reform-or-overhaul> accessed 2 July 2026.
[8] Arafat Hosen Khan, ‘Bangladesh’s Constitutional Crossroads: Reforms, Exclusion, and the Quest for Democratic Legitimacy’ (IACL-AIDC Blog, 27 March 2025) <https://blog-iacl-aidc.org/2025-posts/2025/3/27/bangladeshs-constitutional-crossroads-reforms-exclusion-and-the-quest-for-democratic-legitimacy> accessed 2 July 2026; Arafat Hosen Khan, ‘Bangladesh’s Broken Constitutional Moment: Part I’ (iconnectblog, 8 May 2026) <www.iconnectblog.com/bangladeshs-broken-constitutional-moment-part-i-the-architecture-of-failure-the-national-consensus-commission-and-the-problem-of-selective-constitutionalism/> accessed 2 July 2026; Arafat Hosen Khan, ‘Bangladesh’s Broken Constitutional Moment: Part II’ (iconnectblog, 9 May 2026) <www.iconnectblog.com/bangladeshs-broken-constitutional-moment-part-ii-the-bnps-inheritance-and-the-foreclosed-constitutional-moment-can-reform-be-salvaged/> accessed 2 July 2026.
[9] Lutz (n 1) 325–27.
[10] ‘Updated Set of Principles for the Protection and Promotion of Human Rights through Action to Combat Impunity’ (8 February 2005) UN Doc E/CN.4/2005/102/Add.1, principles 35–38.
[11] Juan E Méndez, ‘Constitutionalism and Transitional Justice’ in Michel Rosenfeld and András Sajó (eds), The Oxford Handbook of Comparative Constitutional Law (OUP 2012) 1270.
[12] Cheng-Yi Huang, ‘Constitutionalizing Transitional Justice’ in Cheng-Yi Huang (ed), Constitutionalizing Transitional Justice (Routledge 2023) 1.
[13] ibid 5.
[14] Yaniv Roznai, Unconstitutional Constitutional Amendments: The Limits of Amendment Powers (OUP 2017); Silvia Suteu, Eternity Clauses in Democratic Constitutionalism (OUP 2021).
[15] Gábor Halmai, ‘Transitional Constitutional Unamendability?’ (2019) 21 European Journal of Law Reform 259.
[16] July Charter (n 4), ‘Historical Context’.
[17] July Charter, proposal 3 (‘Constitutional Amendment Procedure’).
[18] M Rafiqul Islam and Muhammad Ekramul Haque (eds), The Constitutional Law of Bangladesh: Progression and Transformation at Its 50th Anniversary (Springer 2023).
[19] July Charter, proposals 17–19 (‘formation, structure, and functions of the upper house’).
[20] ibid, proposals 38–42.
[21] ibid proposal 16. Fatima Zahra Ahsan Raisa and Suriya Tarannum Susan, ‘Bangladesh Through the Prism of Doctrine’ (Verfassungsblog, 11 September 2024) <https://verfassungsblog.de/bangladesh-through-the-prism-of-doctrine/> accessed 2 July 2026; Khan, ‘Bangladesh’s Constitutional Crossroads: Reforms, Exclusion, and the Quest for Democratic Legitimacy’ (n 8).
[22] July Charter (n 4) proposal 64; Mahia Rahman, ‘Can the Revised ICT-BD Ordinance Facilitate Transitional Justice for the Bangladesh Genocide?’ Global Human Rights Defence (4 April 2025).
[23] Anwar Hossain Chowdhury v Bangladesh (1989) 41 DLR (AD) 165. Islam and Haque, The Constitutional Law of Bangladesh (n 18); Ridwanul Hoque, ‘Eternal Provisions in the Constitution of Bangladesh: A Constitution Once and for All?’ in Richard Albert and Bertil Emrah Oder (eds), An Unamendable Constitution? Unamendability in Constitutional Democracies (Springer 2019) 195.
[24] Constitution (Fifteenth Amendment) Act 2011, art 7B. Islam and Haque, The Constitutional Law of Bangladesh (n 18), and Hoque, ‘Eternal Provisions in the Constitution of Bangladesh’ (n 23).
[25] Islam and Haque, The Constitutional Law of Bangladesh (n 19), chapter on constitutional change.
[26] Islam and Haque (n 17); Roznai (n 13) ch 6.
[27] On the December 2024 High Court decision striking arts 7A, 7B and 44(2) and reviving the art 142 referendum requirement, and the subsequent restoration of the caretaker system on review by the Appellate Division in November 2025, see the contemporaneous reporting collected in The Daily Star and Dhaka Tribune (December 2024–March 2026). The full-annulment challenge to the remainder of the Fifteenth Amendment remained pending at the time of writing.
[28] Asia Justice and Rights and Liberation War Museum, Transitional Justice: Bangladesh (2022), which sets out the 1972–74 prosecutions, the tribunals, and the subsequent reversals.
[29] Rahman, ‘Can the Revised ICT-BD Ordinance Facilitate Transitional Justice for the Bangladesh Genocide?’ (n 21); Asia Justice and Rights and Liberation War Museum (n 27).
[30] July Mass Uprising (Protection and Liability Determination) Act 2026 (passed 8 April 2026), giving statutory form to the interim government’s October 2024 protection measure. The structural resemblance to the Indemnity Ordinance of 1975 lies in the use of a legal instrument to fix the protected status of one side of a transition; the political valence is, of course, opposite.
[31] Halmai (n 15) 259–77; Huang (n 12) 4–5; Gábor Halmai, ‘Transitional Justice, Transitional Constitutionalism and Constitutional Culture’ in Cheng-Yi Huang (ed), Constitutionalizing Transitional Justice (Routledge 2023).
[32] Huang (n 12) 11–12.
[33] July Charter, ‘Pledge of Commitment to the Implementation of the July National Charter 2025’, cl 3.
[34] Nafiz Ahmed, ‘Rhetorical Invocation of Constitutional Guardianship as a Justificatory Tool: The Case of Bangladesh’ (2025) 58(2) Verfassung und Recht in Übersee / World Comparative Law 239.
[35] Implementation Order (n 5); ConstitutionNet translation; Haque and Bisarya, ‘July Charter and Constitutional Reforms in Bangladesh’ (n 3).
[36] Shakhawat Liton, ‘Amending Constitution thru Presidential Order Will Set Dangerous Precedence: Dr Shahdeen Malik’ (The Business Standard, 27 September 2025) <https://www.tbsnews.net/features/panorama/amending-constitution-thru-presidential-order-will-set-dangerous-precedence-dr> accessed 22 April 2026.
[37] Huang (n 12) 4–5.
[38] Implementation Order (n 5), s 4; Hossain (n 5).
[39] Hossain (n 5); ‘Bangladesh’s Election and Referendum: Contesting Reform and Political Futures’ (NUS Institute of South Asian Studies, 9 February 2026).
[40] Roznai (n 14) ch 6.
[41] Shahdeen Malik, ‘It’s Wrong for the Government to Campaign for One Side in the Referendum’, (Prothom Alo, 18 January 2026).
[42] July Charter, ‘Pledge of Commitment’, cl 5; Hossain (n 5).
[43] Khan, ‘Bangladesh’s Broken Constitutional Moment’, Part II (n 8).
[44] On the passage of the Anti-Terrorism (Amendment) Bill 2026 (8 April 2026) without the amendments the committee had recommended, and on the earlier ordinance (11 May 2025) and gazette ban (12 May 2025), see The Business Standard and The Daily Star (May 2025–April 2026).
[45] For the comparative pattern, see the discussion of Adem Kassie Abebe’s analysis of exclusionary post-authoritarian settlements in Khan, ‘Bangladesh’s Broken Constitutional Moment’, Part II (n 8); and, on endurance and inclusiveness generally, Tom Ginsburg, James Melton and Zachary Elkins, The Endurance of National Constitutions (John M Olin Program in Law and Economics Working Paper No 511, University of Chicago Law School 2010).
[46] Constitution of the People’s Republic of Bangladesh, arts 8(2), Parts II–III. On the consequences of relegating socio-economic and environmental rights to non-justiciable state policy, and the judiciary’s partial workaround through an expansive reading of the right to life, see the sustained treatment in the post-July literature on environmental constitutionalism, which argues that a right located in the Fundamental Principles is “resilient” in name only.
[47] July Charter proposal 9; Haque and Bisarya, ‘July Charter and Constitutional Reforms in Bangladesh’ (n 3).
[48] Haque and Bisarya, ‘July Charter and Constitutional Reforms in Bangladesh’ (n 3).
[49] Colleen Murphy, The Conceptual Foundations of Transitional Justice (Cambridge University Press 2017).
[50] Judith Heimbach, ‘The Interaction Between Transitional Justice and Constitution-Making in a Conflict Setting: Reconciliation in Libya’ (2023) 3 Journal of Comparative Law in the Middle East and North Africa 69.
[51] Roznai (n 14); Lutz (n 1) 327–28.
[52] Implementation Order (n 5); Riadul Karim, ‘Yes Vote to Open Path to Reforms’, (Prothom Alo, 30 January 2026) <https://en.prothomalo.com/bangladesh/government/kcx1fa4cjn> accessed 22 April 2026
[53] Heimbach (n 50).
[54] Hossain (n 5).
[55] On the refusal of the Council oath by several formations and the argument that the Council-oath mechanism was legally unfounded, see the reporting of the parliamentary sittings of March 2026 in Prothom Alo, Dhaka Tribune, and bdnews24. The point here is not to resolve that dispute but to note that the self-correcting mechanism on which the counter-argument relies was itself contested at the first opportunity.
[56] Khan, ‘Bangladesh’s Broken Constitutional Moment’, Parts I–II (n 8).