Centro Universitário Campos de Andrade (Uniandrade), Curitiba, Brazil
Retaliatory Constitutionalization Theory explains how states may elevate nuclear command and retaliatory commitments from ordinary policy into constitutional or quasi-constitutional form. Drawing on comparative constitutional law, international nuclear law, command-and-control studies, and AI-governance debates, the article argues that this movement becomes more likely where existential threat perception, decapitation anxiety, executive personalization, and weak internal veto points converge. It introduces two exploratory qualitative measures, the Nuclear Constitutionalization Index and the Retaliatory Constitutionalization Score, and applies them to a 45-jurisdiction comparative map. North Korea appears as the paradigmatic case; Russia, Pakistan, Belarus, and anti-nuclear constitutional states operate as contrasting comparators. The article concludes that contemporary nuclear law remains stronger at regulating possession, transfer, testing, and disarmament than decision, delegation, automation, command survival, and constitutional signalling.
International nuclear law has long been organized around a visible object: the weapon. Its core legal imagination is built around possession, transfer, testing, safeguards, prohibition, non-proliferation, and disarmament. [28] [29] That architecture remains indispensable, but it no longer exhausts the legal problem. A nuclear order can be destabilized even when the number of warheads is unchanged, because the decisive institutional question may move from the arsenal to the decision architecture. Who commands, who substitutes, under what conditions authority devolves, how emergency language is codified, and how retaliation is legally pre-committed are not marginal administrative matters. They are constitutional questions with strategic consequences. In a changing world, the legality of nuclear power cannot be separated from the domestic legal forms through which nuclear decision becomes credible, irreversible, or deliberately inflexible.
This article proposes Retaliatory Constitutionalization Theory as a mid-range account of that movement. Retaliatory constitutionalization occurs when constitutional or quasi-constitutional norms are used to codify, signal, or stabilize retaliatory commitments in extreme-security contexts, especially when the survival of the regime, the leadership, the territorial order, or the strategic command structure is perceived as threatened. The claim is not that all nuclear states constitutionalize nuclear use. Most do not. Nor is the claim that constitutions have suddenly replaced doctrine, operational planning, or alliance politics. The narrower claim is that, in a small but analytically consequential set of cases, the constitution or a law with quasi-constitutional status becomes part of the state's deterrent grammar. It speaks outward as well as inward. It instructs domestic elites, but it also warns adversaries, reassures loyal institutions, narrows ambiguity, and raises the political cost of restraint after certain thresholds have been publicly legalized.
The article is written against two forms of reductionism. The first is doctrinal reductionism, which treats constitutional war powers as an internal matter of institutional competence, legislative authorization, and executive discretion. That lens is necessary but incomplete. Constitutional texts can also be technologies of external signalling. The second is strategic reductionism, which treats deterrence language as if it floated above domestic legal form. That is also incomplete. Legal hierarchy matters. A doctrine placed in a speech does not have the same domestic and international texture as a doctrine embedded in a statute, basic law, constitutional amendment, or entrenched quasi-constitutional settlement. The difference is not merely formal. It affects durability, audience, reversibility, and the interpretive horizon available during crisis.
North Korea is used here as a critical case because it brings the mechanism into unusually sharp relief. The 2022 Law on the DPRK's Policy on Nuclear Forces codified leader-centered nuclear command, incorporated an automatic retaliation formula tied to danger to the nuclear command-and-control system, and gave legal form to a deterrent problem that command-and-control scholarship had already identified: a highly personalized regime must make retaliation credible without undermining the leader's monopoly over ultimate authority. [2] [12] Reporting on the 2026 constitutional revision further suggests that at least part of this nuclear command architecture may have been moved into the constitutional register, although the public record must still be handled cautiously because a consolidated official text remains difficult to audit in full. [11] [15] The evidentiary discipline is central. The available sources support the argument that North Korea has constitutionalized, or nearly constitutionalized, nuclear command authority and retaliatory continuity. They do not prove AI-delegated nuclear decision-making, and this article does not claim that they do.
The contribution is fourfold. First, the article gives a name and analytical form to a phenomenon that sits between comparative constitutional law, nuclear command and control, and international legal theory. Second, it introduces two exploratory tools, the Nuclear Constitutionalization Index, NCI, and the Retaliatory Constitutionalization Score, RCS, not as measurements of military strength, but as qualitative indicators of legal inscription and signalling density. Third, it distinguishes positive, negative, generic, extra-constitutional, and retaliatory modes of nuclear constitutionalization. This distinction is indispensable because Brazil, Palau, Japan, New Zealand, and the Philippines show that nuclear salience in constitutional law can restrain rather than escalate. Fourth, the article argues that artificial intelligence matters here without sensationalism. AI does not need to control launch decisions to destabilize the nuclear order. Its relevance lies in decision-time compression, early-warning uncertainty, target-identification pressure, and the increased value of pre-delegation and automaticity in moments of compressed crisis. [25] [26]
Retaliatory constitutionalization should be distinguished from the older categories with which it overlaps. It is not identical to authoritarian constitutionalism, although it may flourish inside authoritarian or personalist regimes. Authoritarian constitutions may distribute offices, ritualize political supremacy, and manage elite expectations even where liberal limits are weak. [6] Retaliatory constitutionalization refers to a more specific use of constitutional form: the codification of retaliation, command survival, and extreme-security continuity as legal messages to adversaries and internal actors. It is also not reducible to abusive constitutionalism. Landau's account of abusive constitutionalism explains how constitutional tools can be used to weaken democratic competition while preserving a formal constitutional surface. [14] Retaliatory constitutionalization may be abusive, but its distinctive object is different. It does not primarily manipulate democracy. It hardens escalation.
Nor should the concept be confused with emergency constitutionalism. Emergency powers regulate exceptional authority under crisis, sometimes by concentrating power temporarily, sometimes by attempting to domesticate exception through legal form. Retaliatory constitutionalization is narrower and sharper. It appears when the legal order codifies the expectation that certain attacks against leadership, command-and-control, territorial integrity, or regime survival may trigger retaliation. Emergency clauses say who governs during crisis. Retaliatory clauses say, or strongly imply, how the state binds itself to respond when the political center is attacked or strategically incapacitated. The two may reinforce each other, but they are not the same phenomenon.
The theory also differs from nuclear doctrine. Doctrine may remain executive, military, classified, informal, or politically declaratory. A state may have elaborate nuclear doctrine without constitutionalizing it. Retaliatory constitutionalization begins when doctrine migrates upward into a constitutional or quasi-constitutional register. The phrase quasi-constitutional is necessary because some legal systems do not rely on a single codified constitution, and because some national security statutes perform a higher-order function within the political system. The United Kingdom and New Zealand, for different reasons, show why constitutional analysis cannot be limited to a single document. North Korea shows the opposite problem: a formal constitution may interact with statutes and official nuclear laws that already function as part of the regime's highest security settlement.
The causal mechanism can be stated plainly. The independent variable is perceived existential threat, intensified by fear of leadership decapitation, disruption of command-and-control, intervention, regime collapse, or territorial rupture. The key mediating variable is personalization of power. When the leader, the party, the state, the military command, and regime survival are fused, the leader becomes both the symbol of continuity and the single point of political failure. That creates a strategic contradiction. If all authority is concentrated in the leader, adversaries may infer that neutralizing the leader neutralizes retaliation. If authority is broadly delegated, the regime weakens the leader's sacral monopoly over the most consequential decision. Retaliatory constitutionalization tries to solve the contradiction by converting conditional retaliation into an authoritative legal commitment while maintaining leader-centered command as the visible principle of the system.
Several propositions follow. P1: the higher the perceived existential threat to the regime, the stronger the incentive to constitutionalize or quasi-constitutionalize retaliatory commitments. P2: personalist regimes are more likely to adopt retaliatory constitutionalization when the leader functions as a single point of political and military failure. P3: pre-delegation, automaticity, or pre-authorized continuity becomes more attractive when adversaries are perceived as capable of decapitation strikes or command disruption. P4: retaliatory constitutionalization can increase deterrent credibility while decreasing crisis flexibility. P5: AI and military automation do not cause retaliatory constitutionalization by themselves; they increase the legal and strategic significance of delegation, warning, timing, and meaningful human control. P6: international nuclear law remains comparatively stronger at regulating possession than at regulating the legal architecture of decision.
This is a theory of legal signalling under institutional pressure. Its dependent variable is not nuclear use, operational readiness, arsenal size, or battlefield doctrine. Its dependent variable is the constitutional or quasi-constitutional inscription of retaliatory command. The expected strategic effect is deterrence by command survival: adversaries are meant to understand that an attack on leadership or command may fail to eliminate the capacity, obligation, or expectation of response. The effect is double-edged. The same legal hardening that raises the cost of decapitation may reduce space for reinterpretation, delay, bargaining, or de-escalation. Analytically, the theory does not predict a simple positive effect of constitutionalization on security. It predicts a conditional effect shaped by threat perception, regime type, institutional veto points, technological compression, and crisis context.
The article uses a mixed legal-comparative design. It combines theory-building, scoping review, comparative constitutional mapping, exploratory index construction, and critical-case analysis. The objective is not statistical inference. The sample is not designed to estimate population-level causal effects. It is designed to test conceptual boundaries, identify contrasts, and make the theory sufficiently falsifiable to be critically tested. The comparative map covers 45 countries: the five NPT-recognized nuclear-weapon states, the four nuclear-armed states outside the NPT framework, post-Soviet inheritance and renunciation cases, South Africa as a dismantlement case, anti-nuclear and pacifist constitutional cases, NATO umbrella or hosting cases, and authoritarian or presidential comparators with strong defense and emergency clauses.
The evidence base privileges constitutional texts, official legal materials, institutional treaty pages, recognized nuclear-security datasets, peer-reviewed scholarship, and specialist analysis. The Constitute Project is used as a constitutional mapping instrument [5], while UNODA, the IAEA, CTBTO, SIPRI, FAS, NTI, and the Bulletin of the Atomic Scientists are used for nuclear status, non-proliferation context, and arsenal background, including the SIPRI Yearbook's comparative nuclear-force data. [27] Media reporting is used only when primary legal texts are unavailable or politically opaque, and it is classified separately from legal evidence. This separation matters most for North Korea, where the 2022 nuclear-forces law is a high-confidence legal source, while reporting on the 2026 constitutional revision is treated as medium-confidence evidence pending wider public access to the consolidated constitutional text.
Two indices structure the comparison. The Nuclear Constitutionalization Index, NCI, measures how far nuclear issues are incorporated into constitutional or quasi-constitutional form. It ranges from 0 to 5: 0 means no meaningful nuclear relevance; 1 means only generic military command; 2 means indirect nuclear relevance through statutory or uncodified arrangements; 3 means the constitution materially shapes nuclear governance or anti-nuclear identity; 4 means strong command, peace, neutrality, or peaceful-use clauses with direct nuclear salience; 5 means direct constitutional or quasi-constitutional regulation of nuclear weapons, nuclear-free status, or nuclear-force command. The Retaliatory Constitutionalization Score, RCS, ranges from 0 to 10 and measures legal-retaliatory signalling.
The RCS uses an equal-weighted additive coding procedure. Each of the five dimensions receives 0, 1, or 2 points: 0 when the dimension is absent; 1 when it appears indirectly, weakly, or through ordinary political practice; and 2 when it appears directly in constitutional, quasi-constitutional, statutory, or officially entrenched strategic form. The five dimensions are: nuclear status in higher or quasi-constitutional law; leader or executive command personalization; automaticity or pre-delegation; formalized retaliatory triggers; and weak legislative or judicial counterweights. The final score is the sum of these five dimensions. Two ceiling rules discipline the coding: cases without a legally cognizable nuclear or strategic-command source normally cannot exceed 4, and cases without a formalized retaliatory trigger normally cannot exceed 6. Scores of 9 or 10 require the combination of leader-centered command, retaliatory triggers, and legal automaticity or pre-delegation.
| Index | Scale | Coding logic |
|---|---|---|
| NCI | 0-5 | Degree to which nuclear issues enter constitutional or quasi-constitutional law: from no nuclear relevance to direct regulation of nuclear weapons, nuclear-free status, or nuclear-force command. |
| RCS | 0-10 | Equal-weighted additive score across five 0-2 dimensions: legal nuclear status, leader/executive command personalization, automaticity or pre-delegation, retaliatory triggers, and weak internal counterweights. Ceiling rules limit cases without legal nuclear sources or formalized retaliatory triggers. |
The indices are intentionally qualitative. They do not measure warhead numbers, technical readiness, probability of use, doctrinal credibility, or actual military effectiveness. They measure legal inscription and signalling density. This limitation is not a defect. It is precisely what allows the comparison to distinguish Brazil from North Korea, Palau from Russia, and Japan from the United States. A high NCI can indicate restraint, prohibition, pacifism, or escalation, depending on legal content. The RCS then separates restraining constitutionalization from retaliatory constitutionalization. In other words, the NCI asks whether the nuclear question enters higher law; the RCS asks whether that entry hardens retaliation.
| Evidence level | Description | Use in article |
|---|---|---|
| High | Primary legal text or official legal publication. | May support direct legal claims. |
| Medium | Specialist institutional analysis or credible news reporting based on documents. | May support cautious claims with caveats. |
| Low | Media paraphrase, advocacy summary, social media, or sensationalized reporting. | Do not use for central claims. |
The comparative map produces a first result that is theoretically important precisely because it is not spectacular: full retaliatory constitutionalization is rare. Most nuclear-armed states do not place nuclear launch authority or retaliatory conditions directly in constitutional text. They constitutionalize command in general terms and leave nuclear doctrine to statutes, executive arrangements, military practice, classified procedures, or political declarations. The United States has a powerful Commander-in-Chief clause and a long practice of presidential primacy in nuclear launch authority, but the Constitution does not itself contain a nuclear clause. [21] Russia's Constitution centralizes supreme command in the presidency and gives the president a decisive role in martial law, but its nuclear posture is largely located in presidential doctrine and decree. China constitutionalizes the Central Military Commission and the responsibility of its chair, yet its nuclear policy remains largely outside constitutional text. France and the United Kingdom likewise show that robust nuclear capacity may coexist with extra-constitutional or semi-constitutional command arrangements.
Russia deserves separate treatment because its nuclear posture sits close to, but still outside, the core of retaliatory constitutionalization. The 2020 Basic Principles of State Policy in the Sphere of Nuclear Deterrence and the 2024 revision are not constitutional provisions, but they are presidentially issued, public, general, and intended to stabilize expectations about nuclear response under extreme threat. [22] [23] The 2024 revision broadens the signalling environment by linking nuclear response more explicitly to aggression supported by a nuclear power and to threats against sovereignty and territorial integrity. For this reason, Russia's RCS is revised from 7 to 8. It remains below North Korea because the Russian materials do not constitutionalize automaticity or substitute command in the same way, but they are closer to retaliatory constitutionalization than ordinary doctrine.
India, Pakistan, and Israel sharpen the comparison among NPT-outside nuclear-armed states. India has a public nuclear doctrine and a Nuclear Command Authority arrangement that keeps authorization within a civilian political council, but its constitutional text does not itself codify nuclear retaliation. [7] Pakistan is more relevant to the quasi-constitutional category because the National Command Authority Act 2010 gives statutory form to strategic command and control; its posture remains legally more explicit than Israel's and more institutionally formalized than mere doctrine. [8] Israel is different again: its nuclear opacity produces strategic signalling through ambiguity and executive-security practice rather than through constitutional or statutory nuclear disclosure. [4] These cases confirm that nuclear capacity, public doctrine, statutory command, and retaliatory constitutionalization must be kept analytically separate.
A second pattern is negative nuclear constitutionalization. Palau and the Philippines are especially useful because they defeat the false inference that constitutional nuclear salience is inherently escalatory. Palau's Constitution constitutionalizes a nuclear-free commitment by requiring an extraordinary referendum threshold for the presence of nuclear weapons or related harmful substances. The Philippine Constitution contains a territorial nuclear-free policy. Brazil's Constitution limits nuclear activity to peaceful purposes and places it under congressional control. Japan's Article 9 is not a nuclear clause in a narrow textual sense, but it creates a peace-constitutional identity that has shaped the legal and political meaning of force, rearmament, and nuclear dependence. New Zealand lacks a single codified constitution, yet its Nuclear Free Zone, Disarmament, and Arms Control Act has a quasi-constitutional political status in the country's security identity. [16] These cases are analytically indispensable because they show that higher-law nuclearization can restrain the state rather than intensify retaliation.
A third pattern is accommodative nuclear re-entry. Belarus is the clearest example in the present sample. The 2022 constitutional referendum removed prior language associated with neutrality and nuclear-free status, creating a legal-political environment more compatible with Russian nuclear deployment. [17] Belarus should not be treated as a full case of retaliatory constitutionalization, because its posture remains derivative of Russian strategic command. Yet it is not a trivial case. It demonstrates that constitutional change can reopen nuclear possibility after a renunciation trajectory. The Belarusian case therefore belongs between negative constitutionalization and retaliatory constitutionalization. It is an enabling case, not a paradigmatic one.
A fourth pattern is military-generic constitutionalization. Many constitutions regulate armed forces, territorial integrity, war declaration, emergency powers, and civilian command without giving nuclear questions any direct textual status. Germany, Italy, the Netherlands, Belgium, South Korea, Turkey, Canada, Australia, Spain, Portugal, Mexico, Argentina, Colombia, Indonesia, Egypt, Venezuela, Cuba, and Vietnam fall into different subtypes of this broad family. Their constitutional texts matter because they structure who may declare war, command forces, approve emergency measures, authorize deployments, or discipline military power. But they do not generally constitutionalize nuclear retaliation. They are necessary comparators because they show the baseline from which retaliatory constitutionalization deviates.
The conceptual payoff is discriminant validity. Retaliatory constitutionalization is not any mention of the military. It is not any strong presidency. It is not any nuclear arsenal. It is not any emergency clause. It is a tighter configuration: perceived existential threat, leader-centered or executive-centered command, legal elevation of nuclear or strategic retaliation, conditional continuity under attack, and weak internal constraints on escalatory interpretation. On this definition, North Korea is not merely one case among many. It is the case that exposes the full mechanism.
| Country or group | Type | NCI | RCS | Interpretation |
|---|---|---|---|---|
| North Korea | Retaliatory constitutionalization | 5 | 10 | Paradigmatic case of leader-centered nuclear command and conditional retaliatory continuity. |
| Russia | Presidential nuclear-doctrine entrenchment | 4 | 8 | Public presidential doctrine and 2024 revision create strong legal-retaliatory signalling without constitutional automaticity. |
| Pakistan | Quasi-constitutional strategic command | 4 | 7 | National Command Authority Act 2010 gives statutory form to strategic command and control. |
| United States | Extra-constitutional nuclear command on strong executive base | 4 | 7 | Powerful commander-in-chief architecture, with nuclear specificity largely outside constitutional text. |
| India | Public doctrine and civilian nuclear command | 4 | 6 | Nuclear Command Authority and public doctrine; no direct constitutionalization of nuclear triggers. |
| China | Party-state military centralization | 4 | 6 | Central Military Commission is constitutionally central, but nuclear triggers are not constitutionalized. |
| Israel | Strategic opacity | 2 | 5 | Nuclear ambiguity operates through executive-security practice rather than disclosed constitutional command. |
| Belarus | Accommodative nuclear re-entry | 3 | 5 | Constitutional change removed barriers associated with neutrality and nuclear-free status. |
| Palau | Negative constitutionalization | 5 | 0 | Direct constitutional nuclear restrictions with extraordinary popular threshold. |
| Philippines | Negative constitutionalization | 5 | 0 | Territorial anti-nuclear constitutional commitment. |
| Brazil | Peaceful-use constitutionalization | 4 | 1 | Nuclear activity constitutionally limited to peaceful purposes and congressional control. |
| Japan | Pacifist constitutionalization | 4 | 0 | Article 9 structures force and nuclear debate through peace-constitutional identity. |
| NATO umbrella cases | Generic/extra-constitutional nuclear dependence | 2-4 | 3-4 | Constitutional command exists, but nuclear dependence is mainly alliance-based. |
North Korea's importance lies in the interaction between legal text, personalist command, and deterrence under perceived decapitation risk. The 2022 Law on the DPRK's Policy on Nuclear Forces is the central document. It identifies nuclear forces as integral to state defense and sovereignty, places those forces under the monolithic command of the president of the State Affairs Commission, and gives that leader decisive authority over nuclear weapons. It also provides that if the command-and-control system over the state nuclear forces is placed in danger by hostile attack, a nuclear strike shall be launched automatically and immediately according to an operation plan decided in advance. [12] This language is legally and strategically unusual. It translates the problem of command survival into a public legal commitment and should be read against recent scholarship on the DPRK's changed nuclear doctrine and nuclear-force development. [3] [13]
The provision should be read with care. It does not prove a fully autonomous technical system. It does not prove AI launch authority. It does not necessarily describe a mechanical dead-hand apparatus. Bernstein and Smith's analysis for 38 North is persuasive precisely because it rejects that shortcut. They read the law as pointing toward an automatic or pre-planned command model that attempts to preserve Kim Jong Un's primacy while reducing the strategic vulnerability created by one-person rule. [2] That is the core constitutional puzzle. A personalist nuclear regime needs credible retaliation after decapitation, but it cannot comfortably empower alternative centers of authority. Legal automaticity offers a way to communicate continuity without openly pluralizing sovereignty.
The 2026 constitutional revision intensifies the case, but it must not be overstated. Specialist reporting following the March 2026 Supreme People's Assembly session indicates that the constitution was revised, that Kim Jong Un's State Affairs Commission chairmanship was further formalized as the headship position of the state, and that the amended text appears to have removed reunification language while introducing a provision on delegated nuclear command authority. [11] [15] If confirmed in a consolidated official constitutional text, this would mark a clear movement from statutory nuclear policy toward constitutional nuclear command. Yet the article's argument does not depend on a maximalist reading. It is enough to say that the evidence strongly supports legal codification of leader-centered nuclear command and conditional automatic retaliation in 2022, and credibly suggests constitutional elevation of nuclear-force command in 2026.
A robustness check is therefore necessary. If later access to the consolidated 2026 constitutional text shows that the revision is more limited than current reporting suggests, three conclusions would remain intact: the 2022 law would still support the claim of quasi-constitutional nuclear command; North Korea would remain the strongest case of legal automaticity tied to command survival; and the theory would still explain why personalist regimes under decapitation anxiety may legally harden retaliation. What would change is narrower: the claim of formal constitutional elevation in 2026 would be softened, and North Korea's RCS could be read as 9 rather than 10 if delegation language is not confirmed. The theory therefore does not depend on the strongest version of the 2026 claim.
North Korea therefore scores high on both indices. It receives NCI 5 because nuclear command and nuclear-force status enter constitutional or quasi-constitutional legal form. It receives RCS 10 because the legal arrangement combines leader-centered command, explicit retaliatory continuity, conditional automaticity, weak countervailing institutions, and external signalling toward adversaries. The score is not a claim that North Korea possesses the largest arsenal or the most sophisticated command system. It is a claim about legal architecture. Indeed, uncertainty about operational capability makes the legal signal more important rather than less. Law becomes one of the instruments through which the regime tries to compensate for uncertainty, vulnerability, and asymmetry.
The case also reveals the theory's darker side. Retaliatory constitutionalization may improve deterrence by warning adversaries that leadership decapitation will not necessarily neutralize the state. At the same time, it can harden crisis behavior. A legal commitment to rapid retaliation under command threat may reduce the interpretive space needed to pause, verify, communicate, or de-escalate. Command survival becomes legally fused with national survival. The state's constitutional imagination begins to treat restraint as possible betrayal of the founding security commitment. That is why the constitutionalization of retaliation deserves legal attention even when no nuclear weapon is used. It alters the grammar of decision before the decisive moment arrives.
Artificial intelligence enters this argument as a moderator, not as an independent proof of autonomous nuclear decision-making. The distinction is essential. There is no reliable legal evidence that North Korea has constitutionalized AI-based nuclear launch authority. Normative debate has largely moved in the opposite direction, toward preserving meaningful human control over nuclear-use decisions. SIPRI's recent work is more useful than sensational claims because it shows how AI may increase escalation risk even when humans formally remain in the loop. AI can affect warning, surveillance, data fusion, target recognition, decision support, cyber vulnerability, and the tempo of crisis interpretation. [25] [26]
This matters for retaliatory constitutionalization because the legal value of pre-delegation and automaticity rises as decision time shrinks. If leaders believe that adversaries can disable command, blind sensors, paralyze communications, or exploit automated vulnerabilities, the temptation to pre-authorize retaliation grows. Law then becomes a stabilizer and a destabilizer at once. It stabilizes expectations by specifying who commands and what follows from attack on command. It destabilizes crisis management by reducing flexibility exactly when information is degraded. The legal form becomes part of the escalation environment.
International law is underdeveloped on this dimension. The NPT [28] remains the central non-proliferation treaty, structured around non-acquisition, safeguards, peaceful use, and disarmament commitments. The TPNW [29] prohibits a wide range of nuclear-weapon activities for its parties. The CTBT's legal project centers on explosive testing, although entry into force remains incomplete. IHL debates emphasize humanitarian consequences, distinction, proportionality, unnecessary suffering, and the legality of use. These frameworks are indispensable, but their center of gravity is still possession, testing, transfer, use, or prohibition. They do not adequately regulate the domestic legal architecture of nuclear decision, delegation, automaticity, or constitutional signalling. The second-generation crisis of nuclear law is therefore a crisis of decision architecture.
Recent scholarship has already begun to connect artificial intelligence to fundamental rights, public security, and national-security governance across legal systems. [1] [10] [24] The present argument extends that discussion into a domain where law, security, technology, and constitutional form converge. The novelty is not that AI creates every risk. The novelty is that AI makes the legal architecture of decision more consequential. When decision time is compressed, the difference between flexible doctrine and constitutionalized retaliation becomes legally and strategically sharper.
The theory developed here shifts attention from nuclear possession to nuclear decision. That shift does not diminish non-proliferation law. It reveals its limits. A state may comply with certain visible formalities while constructing a domestic command architecture that heightens crisis instability. Conversely, a state may constitutionalize nuclear restraint, as Brazil, Palau, the Philippines, Japan, and New Zealand demonstrate. The legal question is no longer whether nuclear matters enter constitutional law. They already do, in different forms. The question is what kind of constitutionalization occurs and what strategic signal it sends.
Retaliatory constitutionalization also complicates the standard liberal assumption that constitutionalization is inherently restraining. Constitutional law can limit executive violence, but it can also solemnize it. It can democratize decisions over war, but it can also concentrate them. It can humanize security through rights, procedure, and parliamentary control, but it can also transform retaliation into an identity claim. This ambivalence is familiar in abusive constitutionalism, but nuclear command gives it a different scale. When escalation is constitutionalized, the constitutional text becomes part of the state's coercive posture. It is no longer only a forum of internal legality. It becomes a signal in the adversary's risk calculation.
The proposed indices are modest instruments for capturing that ambivalence. Their purpose is not to rank dangerous states. Their purpose is to separate legal salience from retaliatory salience. Without that separation, one might place Palau and North Korea in the same category merely because both constitutionalize nuclear questions. That would be analytically absurd. Palau constitutionalizes a barrier; North Korea constitutionalizes command and retaliation. Brazil constitutionalizes peaceful-use restriction; the United States constitutionalizes executive command in general while leaving nuclear specificity largely outside the text. Belarus removes a constitutional obstacle to nuclear dependence without becoming the source of independent retaliatory command. The conceptual map matters because it prevents nuclear constitutional analysis from collapsing into a single scale of threat.
There are limits. First, the dataset is exploratory and should be expanded. Second, authoritarian legal systems are difficult to interpret because text, practice, and propaganda overlap. Third, uncodified systems require careful treatment of statutes, conventions, and political entrenchment. Fourth, the North Korean constitutional revision of 2026 remains partly dependent on credible reporting rather than full public access to a consolidated official text. Fifth, the indices cannot measure actual operational reliability. A legal signal may be intentionally ambiguous, aspirational, deceptive, or performative. These limits delimit the article's claims and indicate where subsequent research should concentrate: larger coding exercises, deeper country studies, and closer integration between comparative constitutional law and nuclear-command scholarship.
Retaliatory Constitutionalization Theory identifies a narrow but important legal phenomenon: the elevation of retaliatory command, strategic continuity, and nuclear decision architecture into constitutional or quasi-constitutional form. Its central insight is that constitutions can become external signalling devices in extreme-security contexts. They may speak to citizens, courts, legislatures, and bureaucracies, but they may also speak to adversaries when the message concerns leadership survival, command continuity, and retaliation after attack.
The comparative evidence most directly supports P1 and P2: perceived existential threat and personalization of power explain why North Korea, and to a lesser degree Russia and Pakistan, occupy the high-retaliatory end of the map. It also supports P3 and P4, because pre-delegation, automaticity, and legal hardening increase deterrent credibility while narrowing crisis flexibility. P5 is supported more indirectly: AI does not cause retaliatory constitutionalization, but it increases the legal significance of pre-authorization and compressed decision time. P6 follows from the gap between nuclear treaty law and domestic decision architecture. International nuclear law remains stronger at regulating possession than at regulating command, delegation, and constitutional signalling.
The forward-looking agenda is therefore precise. Future research should refine the NCI and RCS, expand the coding matrix, distinguish constitutional rhetoric from operational reality, and ask whether emerging norms of meaningful human control can reach domestic nuclear-command architectures. Possession, transfer, testing, safeguards, prohibition, and disarmament remain central. Yet the next crisis may be decided not only by who possesses nuclear weapons, but by how domestic law structures the decision to use, delegate, delay, automate, or retaliate. A changing world requires a legal vocabulary for that architecture. Retaliatory constitutionalization is one proposed entry into that vocabulary.