The Paris Agreement, adopted on 12 December 2015 and entered into force on 4 November 2016, reorganised the global response to climate change around a hybrid design that combines binding procedural duties with nationally determined substantive pledges. This article reviews the first decade of the treaty and asks a single question that matters for students and for policy: has the agreement changed the trajectory of the world, and if so, by how much and through which mechanisms? The study uses a qualitative, document-based approach. It reads the treaty text and the decisions of the Conference of the Parties, and then compares them with peer reviewed assessments, intergovernmental scientific reports, and institutional data published between 2020 and 2026. Three findings emerge. First, the architecture works as a machine for participation and disclosure rather than as a machine for enforcement, and this design choice explains both its near universal membership and its weak ambition. Second, the measurable effect on expected warming is real but modest. Projections have fallen from roughly 3 to 3.5 degrees Celsius at the time of adoption to roughly 2.3 to 2.5 degrees Celsius under full delivery of current pledges, which still misses the treaty's own temperature goal. Third, the centre of gravity in climate governance is shifting from negotiation toward implementation, finance, and courts, a shift confirmed by the 2025 advisory opinion of the International Court of Justice and by the outcome of the thirtieth Conference of the Parties in Belem. The article concludes that the Paris Agreement should be evaluated not as a solution but as an institutional infrastructure whose value depends on the domestic policy, investment, and legal action it makes possible. Keywords: Paris Agreement; climate governance; nationally determined contributions; global stocktake; climate finance; low carbon transition; international environmental law 1. Introduction Few international treaties have carried as much symbolic weight as the #Paris_Agreement. It was negotiated at the twenty first #Conference_of_the_Parties to the United Nations Framework Convention on Climate Change, held in Paris in December 2015, and it was hailed at the time as a diplomatic breakthrough after more than two decades of slow and often bitter bargaining. The agreement did something that no earlier climate treaty had achieved. It secured the participation of almost every state on earth, rich and poor, large emitter and small island, inside a single legal framework with a shared long term goal. The core of that goal is stated in Article 2. Parties commit to holding the increase in global average temperature to well below 2 degrees Celsius above pre-industrial levels, and to pursuing efforts to limit the increase to 1.5 degrees Celsius. The agreement also sets out two further aims that are often forgotten in public discussion: increasing the ability of societies to adapt to #climate_impacts, and making financial flows consistent with a pathway toward low greenhouse gas emissions and climate resilient development. Read together, these three aims describe not only an emissions target but a transformation of the world economy toward a #low_carbon_future. Ten years later, the picture is mixed and it is important to say so plainly. The treaty survived a period of intense geopolitical stress. It attracted new members, new institutions, and vast private capital. Yet greenhouse gas emissions have continued to rise, reaching a record level in 2024, and the world has already experienced a calendar year in which the global average temperature exceeded 1.5 degrees Celsius above the pre-industrial baseline. The United States, the largest historical emitter, completed a second withdrawal from the agreement on 27 January 2026. At the same time, the cost of solar power, wind power, and battery storage has collapsed in ways that few analysts predicted in 2015, and the deployment of clean technology now runs ahead of many official targets. This article is written for students who need a rigorous but readable account of what the Paris Agreement is, how it functions, what it has achieved, and where it has failed. It is structured like a journal article rather than a textbook chapter, because the aim is to model how a scholarly argument is built: a question, a method, a body of evidence, an honest assessment of counter-arguments, and a conclusion that follows from the evidence rather than from hope. Three questions organise the analysis. The first is a question of #institutional_design. Why did states abandon the top down model of the Kyoto Protocol, with its legally binding emission targets for developed countries, and adopt instead a bottom up model in which each country writes its own pledge? What is gained and what is lost by that choice? The second is a question of effectiveness. Effectiveness in international relations is usually measured against a counterfactual: what would the world look like without the treaty? This is difficult to establish, and the article treats claims of causal impact with appropriate caution. Nevertheless, a decade of data allows some careful statements about #emissions trajectories, policy diffusion, and investment patterns. The third is a question of legitimacy and #justice. The agreement rests on the principle of common but differentiated responsibilities and respective capabilities, in the light of different national circumstances. Whether that principle has been honoured in practice, particularly in relation to #climate_finance and to loss and damage, is contested and remains the sharpest political fault line in the regime. The article proceeds as follows. Section 2 reviews the scholarly literature. Section 3 sets out the method. Section 4 provides the historical background from the Rio Convention to Paris. Section 5 analyses the legal architecture in detail. Sections 6 to 11 examine the operating parts of the regime: nationally determined contributions, transparency and the global stocktake, finance, adaptation and loss and damage, cooperative approaches and carbon markets, and the role of non-state actors. Section 12 assesses the empirical record. Section 13 sets out the main criticisms. Section 14 turns to the growing role of courts. Section 15 examines the geopolitics of the past two years, including the outcome of the thirtieth Conference of the Parties. Section 16 offers a research agenda, and Section 17 concludes. 2. Literature Review The scholarship on the Paris Agreement can be sorted into four streams, and a student writing on this topic should be able to name them. 2.1 The regime design stream The first stream is concerned with #regime_design. It asks why the treaty took the shape it did and whether that shape can work. Scholars in this tradition describe the agreement as a hybrid instrument. The obligations of conduct, such as the duty to prepare, communicate, and maintain successive national pledges, and the duty to report emissions, are legally binding. The obligation of result, the actual achievement of the emissions target contained in a pledge, is not. This combination was a deliberate response to the failure of the Copenhagen conference in 2009 and to the political impossibility of securing legally binding targets from the United States Senate and from major emerging economies at the same time. Depledge (2022) has warned against a common caricature in this literature, namely the neat opposition between a rigid top down #Kyoto_Protocol and a flexible bottom up Paris. She shows that the Kyoto regime always contained bottom up features and that the contrast is exaggerated in ways that distort teaching and research. Torstad (2020) frames the design problem as an effectiveness trilemma, in which participation, ambition, and compliance cannot all be maximised at once, and argues that Paris purchased participation by sacrificing enforcement. Victor, Lumkowsky, and Dannenberg (2022) provide experimental evidence that diplomats themselves discount the credibility of pledges that lack domestic policy backing, which suggests that formal architecture matters less than domestic implementation capacity. 2.2 The legal stream The second stream is #doctrinal and is concerned with the interpretation of the treaty. Bodansky (2021) situates the agreement within the longer arc of international environmental law and stresses its dependence on reputational rather than coercive mechanisms. Rajamani and colleagues (2021) analyse how principles of international environmental law, including equity and common but differentiated responsibilities, can be used to assess whether a given national pledge represents a fair share of the global effort. This body of work has become directly operational since the 2025 advisory opinion of the International Court of Justice, which adopted a demanding reading of the treaty's obligations and rejected the argument that climate treaties exhaust the applicable law. 2.3 The effectiveness and modelling stream The third stream is empirical and #quantitative. It asks what the pledges add up to. Meinshausen and colleagues (2022) calculated that full implementation of announced pledges and long term net zero targets could hold warming just below 2 degrees Celsius, a finding that was widely reported as optimistic. Rogelj and colleagues (2023) responded with a credibility analysis showing that most net zero targets lack the legal status, coverage, and near term policy to be treated as reliable. Hohne and colleagues (2021) had earlier described the wave of net zero announcements as an opening window rather than an achievement. The tension between these papers is instructive: the same pledges produce very different projections depending on how much confidence the analyst places in political promises. 2.4 The justice and political economy stream The fourth stream is #critical_political_economy. It examines who pays, who benefits, and who is excluded. Roberts and colleagues (2021) and Timperley (2021) document the failure of developed countries to deliver the promised 100 billion United States dollars per year on schedule and in a form useful to recipients. Boyd and colleagues (2021) argue that loss and damage constitutes a distinct justice agenda that the mitigation centred regime has consistently marginalised. Newell (2021) situates the transition within global political economy and argues that incumbent fossil fuel interests retain structural power that treaty texts cannot dislodge. Van Asselt and Newell (2022) note the striking absence of any explicit reference to fossil fuels in the Paris text itself, a silence that persisted in formal decisions until 2023. 2.5 The gap in the literature Most existing reviews were written before three developments that change the assessment: the completion of the first global stocktake in 2023, the advisory opinion of the International Court of Justice in 2025, and the second withdrawal of the United States, effective January 2026. This article contributes an integrated reading of the decade that incorporates these events. 3. Materials and Methods This study uses a qualitative, interpretive #research_design that combines legal analysis of primary treaty documents with a structured review of secondary literature and quantitative evidence drawn from authoritative institutional sources. 3.1 Research design The design is a single case study of a treaty regime observed over a defined period, from December 2015 to July 2026. A case study is appropriate because the object of analysis is a unique institution whose behaviour cannot be captured by statistical sampling. The analysis is descriptive and evaluative rather than causal in a strict experimental sense, and claims about impact are stated with the uncertainty that the evidence supports. 3.2 Sources of evidence Three categories of source are used. #Primary_legal_materials include the text of the Paris Agreement, the accompanying decision of the twenty first Conference of the Parties, the Katowice rulebook adopted in 2018, the rules on cooperative approaches finalised in 2024, the outcome of the first global stocktake adopted in 2023, and the political package adopted at the thirtieth Conference of the Parties in Belem in 2025. The 2025 advisory opinion of the International Court of Justice is treated as a primary source of legal interpretation. Scientific and institutional assessments include reports of the Intergovernmental Panel on Climate Change, the annual emissions gap and adaptation gap reports of the United Nations Environment Programme, the synthesis reports of the treaty secretariat on national pledges, the Global Carbon Budget dataset, and the analyses of the International Energy Agency. #Peer_reviewed_literature was identified through searches of scholarly databases using terms including Paris Agreement, nationally determined contributions, global stocktake, climate finance, and climate litigation. Preference was given to work published from 2021 onwards, with a small number of earlier items retained where they remain foundational. 3.3 Method of analysis The analysis proceeds in three steps. First, doctrinal reading identifies the obligations created by each provision and classifies them as obligations of conduct, obligations of result, or hortatory language. Second, process tracing links each obligation to the institution created to implement it and then to the observable output of that institution, such as submitted pledges, biennial transparency reports, or disbursed funds. Third, triangulation compares official claims of progress with independent scientific and academic assessments, and reports disagreement where it exists rather than resolving it artificially. 3.4 Limitations Three #limitations should be acknowledged. The counterfactual problem means that attributing emission reductions to the treaty rather than to technology costs or domestic politics is inherently uncertain. Reporting asymmetries mean that data quality varies widely between countries, and self reported figures for finance in particular are contested. Finally, the period under study is short relative to the multi decadal nature of the problem, so any judgement about effectiveness is provisional. 4. Historical Background: From Rio to Paris 4.1 The framework convention The #United_Nations_Framework_Convention_on_Climate_Change was opened for signature at the #Earth_Summit in Rio de Janeiro in 1992 and entered into force in 1994. Its objective, stated in Article 2, is the stabilisation of greenhouse gas concentrations at a level that would prevent dangerous anthropogenic interference with the climate system. The convention established the principle of common but differentiated responsibilities, divided parties into annexes reflecting their level of development, and created the annual Conference of the Parties as the supreme decision making body. It did not, however, set quantified targets. 4.2 The Kyoto Protocol and its limits The 1997 #Kyoto_Protocol supplied the missing targets. It imposed quantified emission limitation and reduction commitments on developed countries listed in Annex I for the period 2008 to 2012, and it created three market based flexibility mechanisms: emissions trading, joint implementation, and the clean development mechanism. Its architecture was legally strong and politically narrow. The United States signed but never ratified. Canada withdrew in 2011. Developing countries, including China, which became the world's largest annual emitter during the protocol's first commitment period, carried no binding targets. By the time of the Doha amendment, the protocol covered a small and shrinking share of global emissions. The lesson drawn by negotiators was uncomfortable. Legal bindingness had been purchased at the price of participation. A treaty that binds a minority of emissions cannot solve a global commons problem. 4.3 Copenhagen and the road to Paris The #Copenhagen conference of 2009 attempted to produce a comprehensive successor and failed. The resulting accord was a political statement negotiated by a small group of leaders and merely noted rather than adopted by the plenary. Yet Copenhagen contained the seed of the later design. It introduced the idea that countries would inscribe their own pledges, and it produced the first political articulation of the 2 degree limit and of the 100 billion dollar finance goal. Between 2011 and 2015 the Durban Platform mandated negotiation of a new instrument applicable to all parties. Two developments made agreement possible. The first was a bilateral understanding between the United States and China announced in 2014, which removed the assumption that the two largest emitters would block each other. The second was the decision to invite all parties to submit intended nationally determined contributions before the conference, which converted the negotiation from a bargaining over shared burdens into a process of registering pre-formed national positions. By the time delegates arrived in Paris, most of the substance was already on the table. 4.4 Adoption and entry into force The agreement was adopted by #consensus on 12 December 2015. It required ratification by at least 55 parties accounting for at least 55 per cent of global emissions, a threshold designed to prevent a repeat of the Kyoto experience. That threshold was crossed with unusual speed, and the treaty entered into force on 4 November 2016, less than a year after adoption. This speed was itself a political act, intended to insulate the agreement from the incoming administration in the United States. 5. The Legal Architecture 5.1 The hybrid model The central innovation of the agreement is its combination of a binding procedural framework with #nationally_determined content. Article 4 requires each party to prepare, communicate, and maintain successive contributions that it intends to achieve, and to pursue domestic mitigation measures with the aim of achieving them. The verbs matter. The duty to communicate and to pursue is binding. The duty to achieve is not expressed in the same mandatory terms. This distinction, described by lawyers as the difference between obligations of conduct and obligations of result, is the hinge on which the entire regime turns. The consequence is that no party can be sanctioned under the treaty for missing its own target. What the treaty provides instead is a system of expectation, disclosure, and comparison. It assumes that reputational pressure, domestic politics, market signals, and peer review will do the work that enforcement cannot. 5.2 The temperature goal Article 2 establishes the #temperature_goal in two limbs: holding warming well below 2 degrees Celsius, and pursuing efforts toward 1.5 degrees Celsius. In 2015 the second limb was widely regarded as aspirational, inserted to satisfy small island states and least developed countries. The special report of the Intergovernmental Panel on Climate Change in 2018 changed that perception by documenting materially greater risks at 2 degrees than at 1.5, and the 1.5 figure progressively became the operative benchmark of the regime. The International Court of Justice in 2025 went further and read the goal dynamically, treating 1.5 degrees Celsius as the applicable standard against which national conduct is measured. 5.3 The ratchet mechanism Article 4 requires that each successive contribution represent a progression beyond the party's current contribution and reflect its highest possible ambition. This is the #ratchet_mechanism. Pledges are submitted on a five year cycle. Combined with the five yearly global stocktake under Article 14, the intended effect is a spiral of increasing ambition: countries pledge, the world assesses collective progress, countries pledge again with better information and under greater scrutiny. The mechanism is elegant on paper. In practice it depends entirely on political will, because the terms progression and highest possible ambition are not defined and no institution has been given the power to reject a pledge as inadequate. 5.4 Differentiation #Differentiation replaces the rigid annex based division of the earlier regime with a more fluid form of differentiation. All parties must submit contributions, but developed countries should continue to take the lead by undertaking economy wide absolute reduction targets, while developing countries are encouraged to move over time toward such targets in the light of different national circumstances. Differentiation is also expressed through flexibility in reporting requirements for countries with limited capacity, and through the finance obligations examined below. This shift was politically necessary and remains politically contested. Emerging economies insist that historical responsibility cannot be erased by a change of drafting. Developed countries insist that a treaty that ignores the current distribution of emissions cannot succeed. The International Court of Justice touched this nerve in 2025 when it observed that the status of a state as developing should be assessed on the basis of its actual and current circumstances rather than assumed from a category fixed decades ago. 5.5 Compliance Article 15 establishes a #compliance_committee to facilitate implementation and promote compliance. Its mandate is explicitly described as expert based and facilitative, and it is required to function in a manner that is transparent, non-adversarial, and non-punitive. It cannot impose penalties. Its principal power is to engage a party in dialogue and, in limited circumstances, to make findings and recommendations. Critics regard this as the clearest evidence that the treaty lacks teeth. Defenders reply that a punitive committee would have produced a treaty with far fewer members and therefore far less coverage of global emissions. 5.6 Withdrawal Article 28 permits a party to #withdraw three years after the agreement has entered into force for that party, with the withdrawal taking effect one year after notification. This provision was designed to slow down impulsive exits. It was invoked by the United States in 2019, taking effect in November 2020, reversed in February 2021, and invoked again in January 2025, taking effect on 27 January 2026. In January 2026 the United States also announced its intention to withdraw from the framework convention itself, an unprecedented step whose legal basis is disputed by constitutional scholars. 6. Nationally Determined Contributions and the Ambition Cycle 6.1 What a contribution contains A #nationally_determined_contribution is a document submitted to a public registry in which a party sets out what it will do. There is no prescribed template, although guidance adopted in 2018 requires information necessary for clarity, transparency, and understanding. In practice, contributions vary enormously. Some contain an economy wide absolute target measured against a base year. Others express intensity targets, deviation from a business as usual baseline, or targets conditional on the receipt of international support. Many include adaptation components and sectoral policies. This heterogeneity is the price of #universality. It is also a serious analytical problem, because comparing a conditional intensity target from one country with an absolute economy wide target from another requires modelling assumptions that are themselves contestable. 6.2 The three rounds The first round consisted of the intended contributions submitted in 2015, converted into first contributions upon ratification. The second round, due in 2020 and delayed by the pandemic, produced updated pledges in the run up to the Glasgow conference in 2021. The third round, covering the period to 2035, was due by February 2025. The third round has been disappointing in both timing and content. The large majority of parties missed the February 2025 deadline. By the end of September 2025 only about sixty parties, covering under two thirds of global emissions, had submitted or announced new mitigation targets for 2035. Submissions accelerated in the final weeks before the Belem conference, and by its close more than one hundred and twenty parties had submitted new or updated plans, but the aggregate #ambition_gap remained wide. The United Nations Environment Programme concluded that the new pledges reduced projected warming by roughly 0.1 degrees Celsius, and that once methodological revisions and the effect of the United States withdrawal were taken into account, the new pledges had barely moved the needle. 6.3 The implementation gap Two gaps must be distinguished, and students frequently confuse them. The ambition gap is the distance between what countries have promised and what the temperature goal requires. The implementation gap is the distance between what countries have promised and what their existing policies will actually deliver. Both are large. Current policies point to roughly 2.8 degrees Celsius of warming this century, while full delivery of pledges points to roughly 2.3 to 2.5 degrees Celsius. Closing the implementation gap is therefore a precondition for the ambition gap to have any meaning at all. 6.4 Net zero targets Alongside the five yearly cycle, the agreement invites parties to communicate long term low emission development strategies. This provision, initially treated as marginal, became the vehicle for the extraordinary spread of #net_zero pledges after 2019. Countries representing the large majority of global emissions, gross domestic product, and population now have some form of net zero commitment. The scholarly assessment is cautious. Fankhauser and colleagues (2022) set out the conditions under which a net zero target is meaningful: it must cover all gases and scopes, specify the role of removals, avoid over-reliance on offsets, and be backed by near term action. Rogelj and colleagues (2021, 2023) show that most targets fail several of these tests, and that treating them as reliable inputs to temperature projections produces an unjustified sense of progress. The credibility of a distant target is best judged by the policies adopted today. 6.5 Conditional pledges and the dependence on support A feature of the pledging system that receives too little attention in student essays is #conditionality. A large number of developing country pledges are split into an unconditional component, which the country will deliver from its own resources, and a conditional component, which it will deliver only if it receives international finance, technology transfer, and capacity building. This design has two effects. It makes the aggregate ambition of the system contingent on the delivery of finance, which means that a shortfall in the finance chapter of the treaty translates directly into a shortfall in the mitigation chapter. It also creates an analytical trap: reports that headline the ambition of global pledges often quietly assume the conditional components will be delivered, even though the historical record of support delivery is poor. When conditional pledges are discounted, the projected #emissions_gap widens considerably. For a country such as Kenya, Pakistan, or Vietnam, the conditional component may represent the larger part of the pledge. The architecture therefore builds a mutual dependence into the regime. Developed countries cannot reach the temperature goal without action in emerging and developing economies, and those economies have stated in writing that a substantial part of their action depends on support. This is not a loophole. It is a bargain, and the failure to honour it is a failure of the whole system rather than of one side. 6.6 Sectoral coverage and the quality of pledges Beyond headline numbers, the quality of a contribution depends on what it covers. Many pledges are weak on #methane, on land use and forestry, and on the emissions embedded in international aviation and shipping, which fall outside national inventories and are handled by separate specialised agencies. Nascimento and colleagues (2022) mapped two decades of policy across the Group of Twenty and found persistent gaps in industry, agriculture, and buildings, precisely the sectors that are hardest to decarbonise and least visible to voters. A pledge that promises deep cuts by 2035 while leaving these sectors unregulated is a promise without a mechanism. 7. Transparency, Reporting, and the Global Stocktake 7.1 The enhanced transparency framework If the treaty cannot enforce, it must at least see. This is the logic of #accountability_through_disclosure. Article 13 establishes an enhanced transparency framework applicable to all parties, with built in flexibility for developing countries that need it in the light of their capacities. Under this framework each party must submit a national inventory report of emissions by sources and removals by sinks, and information necessary to track progress in implementing and achieving its contribution. Submissions are subject to technical expert review and to a facilitative, multilateral consideration of progress. The first biennial transparency reports fell due at the end of 2024. Compliance was uneven, and capacity constraints in many developing countries are real rather than rhetorical. Weikmans, van Asselt, and Roberts (2020) argue that transparency is unlikely by itself to drive higher ambition, because information changes behaviour only where a domestic constituency exists to use it. This is an important corrective to the assumption that #transparency is automatically transformative. 7.2 The global stocktake Article 14 requires a periodic collective assessment of progress toward the purpose and long term goals of the agreement, conducted every five years. The first #global_stocktake ran from 2021 to 2023 and concluded at the twenty eighth Conference of the Parties in Dubai. Its technical dialogue produced an unambiguous scientific finding: parties are not collectively on track. Its political outcome, known as the United Arab Emirates Consensus, was more significant than many expected. For the first time in nearly thirty years of negotiation, a consensus decision called on parties to contribute to transitioning away from fossil fuels in energy systems in a just, orderly, and equitable manner, so as to achieve net zero by 2050 in keeping with the science. The same decision called for tripling global renewable energy capacity and doubling the annual rate of improvement in energy efficiency by 2030. The stocktake is best understood as an accountability moment rather than an enforcement moment. It does not assess individual countries. It assesses the world. Its purpose is to inform the next round of contributions, and the weakness of the third round suggests that the informational channel between the stocktake and national planning is not yet working as designed. 8. Climate Finance 8.1 The obligation Article 9 provides that developed country parties shall provide #financial_resources to assist developing country parties with respect to both mitigation and adaptation, in continuation of their existing obligations under the framework convention. The word shall makes this a binding obligation, but the article does not quantify it, and the absence of a figure in the treaty text has produced a decade of argument. The quantified figure comes from political decisions rather than the treaty. Developed countries had pledged in 2009 to mobilise 100 billion United States dollars per year by 2020. The target was missed in 2020, reported as met in 2022 according to figures compiled by the Organisation for Economic Cooperation and Development, and disputed throughout by recipients and by independent analysts. 8.2 Why the figure is disputed Three problems recur in the #literature. The first is definitional. There is no agreed accounting standard for what counts as climate finance, and instruments as different as a concessional grant and a market rate export credit have been reported under the same heading. The second is the loan problem. A large share of reported finance takes the form of debt, which deepens the fiscal stress of countries already spending heavily on debt service. Timperley (2021) and Roberts and colleagues (2021) describe this as a broken promise rather than a delayed one. The third is the adaptation shortfall: mitigation attracts the overwhelming majority of flows because it produces bankable revenue streams, while #adaptation is largely a public good with no obvious return. 8.3 The new collective quantified goal The 2015 decision required parties to set a new collective quantified goal from a floor of 100 billion dollars per year before 2025. That goal was agreed at the twenty ninth Conference of the Parties in Baku in 2024. Developed countries agreed to take the lead in mobilising at least 300 billion United States dollars per year by 2035 for developing countries, within a broader call on all actors to scale up finance from all public and private sources to at least 1.3 trillion dollars per year by 2035. The reception was hostile among many developing country negotiators, who had asked for a core public provision figure an order of magnitude larger and who regarded a mobilisation target that leans heavily on private capital as an evasion of the obligation in Article 9. Pauw and colleagues (2022) had anticipated exactly this dispute, arguing that the post-2025 goal needed to address quality, access, and predictability, not only headline volume. At Belem in 2025, finance remained the most contentious file. The final political decision established a two year work programme covering the whole of Article 9, took note of a roadmap intended to chart how the 1.3 trillion dollar figure might be reached, called for a tripling of adaptation finance by 2035, and launched an accountability framework for climate finance delivery. Developing countries had sought a tripling by 2030 and a dedicated negotiating track on the public provision obligation. They did not get either. 9. Adaptation, Loss and Damage 9.1 The global goal on adaptation Article 7 establishes a #global_goal_on_adaptation: enhancing adaptive capacity, strengthening resilience, and reducing vulnerability. For years this goal existed without indicators, which made progress impossible to measure. A framework was adopted in Dubai in 2023, and at Belem in 2025 parties agreed a set of fifty nine indicators, now referred to as the Belem adaptation indicators, including indicators for means of implementation such as finance, technology transfer, and capacity building. This is genuine institutional progress, but it arrives against a deteriorating financial background. The United Nations Environment Programme reported before the Belem conference that adaptation finance flows to developing countries had fallen rather than risen, and the gap between estimated adaptation needs and available finance remains an order of magnitude wide. Berrang-Ford and colleagues (2021), in a systematic global stocktake of the adaptation evidence, found that documented responses are overwhelmingly fragmented, incremental, and concentrated in individual behaviour change rather than in the deep structural transformation that risk projections imply. 9.2 Loss and damage Article 8 recognises the importance of #averting, minimising, and addressing loss and damage associated with the adverse effects of climate change. It was accompanied by a decision stating that the article does not involve or provide a basis for any liability or compensation. That sentence was the price of the article's inclusion, and it explains why the topic remained frozen for years. The freeze broke at the twenty seventh Conference of the Parties in Sharm el-Sheikh in 2022, which agreed to establish a fund for responding to #loss_and_damage. The fund was operationalised at the following conference in Dubai and began disbursing modest sums thereafter. Its capitalisation, however, is trivial in relation to documented needs, which run into the hundreds of billions of dollars annually. Boyd and colleagues (2021) argue that loss and damage should be understood as a justice agenda concerned with irreversible harm rather than as a residual category left over after mitigation and adaptation. 10. Cooperative Approaches and Carbon Markets Article 6 permits parties to cooperate voluntarily through #cooperative_approaches in implementing their contributions and to use internationally transferred mitigation outcomes. It contains three components: bilateral or plurilateral cooperative approaches under Article 6.2, a centralised crediting mechanism under Article 6.4 supervised by a body accountable to the parties, and a framework for non-market approaches under Article 6.8. The rules proved extraordinarily difficult to negotiate. The central technical problem is #double_counting. If country A sells a mitigation outcome to country B, and both count it toward their own targets, global ambition is reduced rather than increased. The solution adopted is a corresponding adjustment, under which the selling country adds the transferred amount back to its own emissions balance. A second problem is environmental integrity: credits must represent reductions that are real, additional, permanent, and verifiable, and the record of the clean development mechanism on this point is poor. Ahonen and colleagues (2022) draw the direct lesson that governance failures under the earlier mechanism must not be repeated, and Kreibich and Hermwille (2021) show that the credibility of voluntary markets is also at stake. The rulebook was largely completed at Katowice in 2018, extended at Glasgow in 2021, and the remaining standards for the centralised mechanism were adopted at Baku in 2024. Trading has begun under bilateral arrangements, with a small number of countries acting as early buyers and sellers. Whether Article 6 becomes a genuine engine of cost effective mitigation or a laundering mechanism for weak pledges will depend on the strictness of implementation, not on the elegance of the rules. 11. Non-State and Subnational Action The Paris outcome deliberately created space for #non_state_actors. Cities, regions, companies, investors, and civil society organisations were invited into a formal action agenda, and thousands of commitments were registered on a public platform. The theory behind this is that climate governance is polycentric. Authority is distributed across many levels and sites, and international agreement is one node among many. Dubash (2021) documents how national climate institutions have proliferated in the past decade, and Chan and colleagues (2021) argue that the value of the action agenda lies in its catalytic potential rather than in its direct emission reductions. The evidence supports a mixed verdict. When the United States announced its first withdrawal in 2017, a coalition of states, cities, and businesses declared that they remained committed, and subnational action limited the damage. Yet non-state pledges suffer from the same credibility problem as national ones, and often worse, because there is no reporting obligation and no review. #Greenwashing is a genuine risk, particularly in the corporate net zero space, where target setting has outpaced the development of standards for what a target means. 11.1 Cities and regions Urban areas account for the majority of energy related emissions, and #cities control levers that national governments do not: zoning, building codes, public transport, waste management, and municipal procurement. Transnational city networks have become significant policy laboratories, and the diffusion of measures such as low emission zones and building performance standards has often run ahead of national legislation. The limitation is fiscal. Cities in low income countries typically lack the borrowing capacity to finance the infrastructure that their climate plans require, which is why the question of how climate finance reaches the subnational level is more important than its low profile suggests. 11.2 Business, investors, and the disclosure turn The second decade of the regime is likely to be shaped less by pledges and more by disclosure rules. The emergence of mandatory sustainability reporting standards, transition plan requirements for financial institutions, and taxonomies that define what counts as a green investment converts the abstract goal in Article 2, on making financial flows consistent with a low emission pathway, into concrete obligations for firms. Skovgaard (2021) describes this as the economisation of climate change: the migration of the issue from environment ministries into finance ministries, central banks, and international financial institutions. Whether this migration produces real decarbonisation or merely a sophisticated compliance industry is an open empirical question. 11.3 Just transition The phrase #just_transition appears in the preamble of the agreement and has since become a central political demand. It refers to the protection of workers and communities whose livelihoods depend on carbon intensive industries. Coal regions in Poland, South Africa, India, and Indonesia illustrate the difficulty: the closure of a mine is not only an emissions question but a question of pensions, retraining, municipal revenue, and political consent. Newell (2021) argues that transitions fail politically when they ignore this dimension, and the record of populist backlash against climate policy in several democracies supports the point. A treaty that produces technically optimal pathways which are socially unacceptable will not be implemented. 12. The Empirical Record After Ten Years 12.1 Emissions Global #greenhouse_gas_emissions have not fallen. They reached a record of approximately 57.7 gigatonnes of carbon dioxide equivalent in 2024, having grown by about 2.3 per cent on the previous year. Fossil carbon dioxide emissions have set successive records. Lamb and colleagues (2021) show that sectoral drivers remain broadly unchanged, with growth in transport and industry offsetting improvements in electricity generation in many economies. This is the single most important fact in any honest assessment of the agreement. A treaty whose purpose is to reduce emissions has coexisted with a decade of rising emissions. 12.2 Temperature The year 2024 was the #warmest_year in the instrumental record and, according to the World Meteorological Organization, the first calendar year in which the global average temperature exceeded 1.5 degrees Celsius above the pre-industrial baseline. A single year above the threshold does not constitute a breach of the Paris goal, which is defined in terms of a long term average, but the direction of travel is unmistakable. The emissions gap assessment now states plainly that the multi-decadal average will exceed 1.5 degrees Celsius at least temporarily, and the Belem decision acknowledged for the first time in a formal text the likelihood of an #overshoot that must be limited in both extent and duration. 12.3 What has improved Three genuine improvements deserve emphasis, because a purely negative account would be as misleading as a triumphal one. Projected warming has fallen. At the time of adoption, the trajectory implied by policies and pledges was in the range of 3 to 3.5 degrees Celsius. It is now in the range of 2.3 to 2.5 degrees Celsius under full delivery of pledges. This is not sufficient. It is also not nothing: roughly a degree of avoided warming corresponds to an enormous quantity of avoided harm. Policy coverage has expanded. Nearly ninety per cent of global emissions are now covered by some form of national climate pledge, and national climate laws and institutions have proliferated. Eskander and Fankhauser (2020) found evidence that national climate legislation is associated with measurable emission reductions, which supports the theory that the treaty's main effect operates through domestic law rather than through international enforcement. #Technology costs have collapsed. Solar photovoltaic and battery costs have fallen by large margins since 2015, and clean energy investment now substantially exceeds fossil fuel investment globally. The emissions gap assessment for 2025 projected, for the first time, that emissions in China may peak in 2025, driven by renewable generation growth outpacing demand growth. Whether the treaty caused this or merely accompanied it is precisely the counterfactual problem identified in the methods section, but the policy signal created by near universal target setting plausibly shaped investor expectations. 12.4 Regional snapshots Global averages conceal very different national stories, and a good research paper should disaggregate them. The European Union has legislated a binding climate law, operates the world's largest emissions trading system, and has adopted a carbon border adjustment mechanism. Its emissions have fallen substantially since 1990, although part of that reduction reflects the offshoring of industrial production rather than genuine decarbonisation, a point that consumption based accounting makes visible. China is the largest annual emitter and simultaneously the largest manufacturer and installer of #solar_power, wind capacity, and electric vehicles. The 2025 emissions gap assessment projected, for the first time, that Chinese emissions may peak in 2025 as renewable generation growth outpaces demand growth. If confirmed, this would be the most consequential single development of the decade. India frames its position around development and per capita equity. Its per capita emissions remain far below the global average, its energy demand is growing rapidly, and its pledge combines a non-fossil capacity target with an emissions intensity target rather than an absolute cap. The Gulf states, including the United Arab Emirates, occupy an ambiguous position: they are major hydrocarbon exporters and, at the same time, significant investors in renewable capacity and hosts of climate diplomacy. The Dubai conference in 2023 produced the first consensus language on transitioning away from fossil fuels, which illustrates that producer states are participants in the regime rather than simply obstacles to it. Small island developing states and least developed countries contribute a negligible share of emissions and bear a disproportionate share of harm. Their diplomatic influence has been far greater than their material power, and the 1.5 degree limit, the loss and damage fund, and the advisory opinion of the International Court of Justice all originated in their advocacy. This is one of the strongest arguments that multilateralism has value even when it does not deliver enforcement. 12.5 What has not improved Finance for adaptation is falling rather than rising. The fund for loss and damage is capitalised at a level far below need. The third round of pledges is late and weak. Fossil fuel production plans remain inconsistent with any Paris compatible pathway, and the formal negotiation process was unable, even at Belem, to adopt a roadmap for transitioning away from fossil fuels, despite support from more than eighty countries. 13. Criticisms and Structural Weaknesses 13.1 The enforcement critique The most common criticism is that the agreement lacks #enforcement. There is no penalty for a missed target, no tariff on a free rider, and no court with compulsory jurisdiction over compliance. Critics argue that a regime built on voluntary pledges and peer pressure cannot solve a problem in which the incentive to defect is structural. The defence is that enforcement is a trade off against participation, and that a treaty binding forty per cent of emissions with sanctions is worth less than a treaty covering ninety per cent without them. This defence is plausible, but it becomes weaker as the decade of delay lengthens, because the value of universal participation depends on that participation eventually producing action. 13.2 The silence on supply The treaty text does not mention #fossil_fuels, coal, oil, or gas. It regulates the accounting of emissions but not the production of the substances that cause them. Van Asselt and Newell (2022) argue that this supply side blind spot allows governments to pledge ambitious reductions while simultaneously licensing new extraction, a contradiction that no provision of the agreement forbids. The Dubai decision partially closed this gap in political terms, but the failure at Belem to agree a transition roadmap shows how contested the terrain remains. 13.3 The equity critique Rajamani and colleagues (2021) demonstrate that when pledges are assessed against principled interpretations of fairness in international law, few if any developed country pledges represent a fair share of the global effort. #Climate_justice scholarship argues further that the regime treats historical responsibility as a rhetorical concession rather than an operational principle, and that finance and loss and damage arrangements consistently under-deliver relative to stated obligations. Gupta and colleagues (2023) extend the argument to earth system justice, insisting that safe boundaries must be defined jointly with just ones. 13.4 The overshoot and removals problem Because near term reductions have been insufficient, most modelled pathways that return warming to 1.5 degrees Celsius now depend on large scale carbon dioxide removal later in the century. Mace and colleagues (2021) identify serious governance gaps in this area: removals at the scale assumed in models would require land, energy, and finance on a vast scale, and there is no international framework governing their quality, permanence, or equity. A regime that quietly transfers the burden of its own delay onto a speculative future technology is storing up both physical and moral risk. 13.5 The credibility problem Victor and colleagues (2022) show that even sophisticated observers discount pledges that lack domestic policy support. If pledges are not believed, they cannot perform the coordinating function that the architecture assigns to them. Credibility, not ambition, may therefore be the binding constraint on the regime. 14. Courts and the Turn to Legal Accountability One of the most significant developments of the past five years has taken place outside the #negotiating_halls. #Climate_litigation has expanded rapidly, with cases filed in dozens of jurisdictions against governments and corporations. Setzer and Higham (2024) document the growth and diversification of this body of cases, and Savaresi and Setzer (2022) show how human rights arguments have become central to it. The Paris Agreement has proved unexpectedly powerful in this setting. Voigt (2023) explains that although the treaty imposes no obligation of result, courts have used it as an interpretive benchmark: as evidence of the standard of care that a reasonable state owes its citizens, as a source of the temperature goal against which national policy can be measured, and as an expression of the best available science. The decisive moment came on 23 July 2025, when the International Court of Justice delivered a unanimous advisory opinion on the obligations of states in respect of climate change, requested by the United Nations General Assembly following a campaign initiated by Pacific island students and led diplomatically by Vanuatu. The Court held that the obligations of states are not confined to the climate treaties. Customary international law, human rights law, and the law of the sea apply simultaneously. It held that the discretion of parties in preparing their pledges is limited, and that parties must exercise stringent due diligence to ensure that their contributions, taken together, are capable of achieving the temperature goal. It treated 1.5 degrees Celsius as the operative standard. And it stated that failure to act, including through the granting of exploration licences or the provision of fossil fuel subsidies, may constitute an internationally wrongful act attributable to the state, with the ordinary consequences of state responsibility, including cessation and reparation. An advisory opinion is not binding in the way a judgment between parties is binding. Its practical importance lies elsewhere. It supplies domestic courts, litigants, and regulators with an authoritative interpretation of what international law requires, and it applies to states that are not parties to the agreement, since customary obligations bind regardless of treaty membership. For a regime long criticised as toothless, this is a structural change: the teeth may come from courts rather than from the treaty's own compliance committee. 15. Geopolitics: Withdrawal, Fragmentation, and Belem 15.1 The United States The second #withdrawal of the United States took effect on 27 January 2026. In January 2026 the administration also announced an intention to withdraw from the framework convention itself and from a large number of other multilateral bodies, including the Intergovernmental Panel on Climate Change. The legal authority of a president to exit a treaty ratified with the advice and consent of the Senate is disputed, and the question may reach the courts. The consequences are material rather than merely symbolic. The emissions gap assessment estimated that the withdrawal would cancel roughly 0.1 degrees Celsius of the improvement otherwise recorded in 2025, and projections for United States emissions in 2030 were revised upward by about one gigatonne of carbon dioxide equivalent because of domestic policy reversals. The loss of United States funding and data also weakens the scientific infrastructure on which the whole regime depends. Yet no other party followed the United States out of the agreement in either withdrawal episode, which is itself a striking fact about the durability of the institution. 15.2 The thirtieth Conference of the Parties The Belem conference in November 2025 was framed by its Brazilian presidency as the implementation conference and, in the event, delivered a package that reflected both determination and deep division. The central decision, named after the Brazilian concept of #mutirao, or collective effort, reaffirmed the agreement on its tenth anniversary, declared that the shift toward low emission and climate resilient development is irreversible, launched a global implementation accelerator and a mission to keep 1.5 degrees Celsius within reach, established a two year work programme on finance, created the first formal space to discuss unilateral trade measures within the climate regime, and for the first time addressed climate disinformation. It also failed to include any reference to fossil fuels or deforestation, despite the support of more than eighty countries for a transition roadmap. The presidency announced that it would take those roadmaps forward outside the formal process. Whether such presidency led initiatives can substitute for negotiated mandates is one of the open questions of the coming years. 15.3 Trade and the new frontier An emerging fault line concerns unilateral measures such as the #carbon_border_adjustment_mechanism. Eicke and colleagues (2021) show that such measures can impose significant costs on exporting developing countries that lack the fiscal or technical capacity to decarbonise quickly. Whether the climate regime can accommodate these instruments without fracturing along a north south line is now, for the first time, formally on the agenda. 16. Implications and a Research Agenda for Students The literature and the record together suggest five directions in which further work is needed, each of which is accessible to a well prepared student researcher. The first is the measurement of #credibility. Existing work distinguishes ambitious from unambitious pledges, but comparatively little work systematically scores the institutional backing of a pledge: the legal status of the target, the existence of a carbon price or sectoral regulation, the alignment of state owned enterprise investment, and the independence of the monitoring body. The second is the political economy of implementation at the sectoral level. Aggregate national targets conceal a heterogeneous reality in which power, steel, cement, agriculture, and transport face very different technical and political constraints. Case studies of sectoral transitions in specific countries remain undersupplied relative to their importance. The third is the interaction between courts and legislatures after the advisory opinion. If a duty of stringent due diligence is now the applicable legal standard, how do national courts translate it into remedies without displacing democratic choice? This is a rich question at the intersection of law and political theory. The fourth is the effectiveness of #climate_finance instruments. There is a large literature on how much finance is delivered and a much smaller one on whether it works. Bhandary, Gallagher, and Zhang (2021) note the thinness of the evidence base on the policy effectiveness of finance instruments, and this remains true. The fifth is the governance of #carbon_dioxide_removal and of overshoot. If the world is going to exceed 1.5 degrees Celsius temporarily, then the rules governing the return path, including who pays for removals, who verifies them, and who bears the risk of their failure, become central rather than peripheral. 17. Conclusion The #Paris_Agreement is neither the triumph its supporters claimed in 2015 nor the failure its critics claimed in the same year. It is an infrastructure. It establishes a shared goal, a common vocabulary, a five yearly rhythm of pledging and review, a system of disclosure, a set of financial obligations, and a legal reference point that courts have begun to use with increasing confidence. It does not, and was never designed to, compel any state to reduce a single tonne of emissions. Judged against its own temperature goal, the first decade must be counted a failure. Emissions have risen, the third round of pledges is late and weak, adaptation finance is falling, and the 1.5 degree limit is now expected to be breached at least temporarily. Judged against the counterfactual of a world with no universal framework, the record is more favourable. Expected warming has fallen by roughly a degree, nearly all emissions are covered by national pledges, national climate law has spread, and clean technology has scaled faster than almost anyone forecast. The decisive question for the second decade is whether the regime can convert breadth into depth. Three developments will determine the answer. The first is domestic: whether the implementation gap between promise and policy is closed in the largest emitting economies. The second is financial: whether the flow of capital to developing countries becomes adequate, concessional, and predictable rather than contested and debt creating. The third is legal: whether the interpretation of state obligations advanced by the International Court of Justice migrates into national courts and reshapes what governments believe they are permitted to do. For students, the deeper lesson is methodological. International agreements should not be evaluated as if they were machines that produce outcomes directly. They should be evaluated as institutions that alter the incentives, the information, the legal arguments, and the expectations available to other actors. On that standard, the Paris Agreement has done a great deal, and still nowhere near enough. References Ahonen, H. M., Kessler, J., Michaelowa, A., Espelage, A., and Hoch, S. (2022). Governance of Article 6 of the Paris Agreement and lessons learned from the Kyoto Protocol. Global Policy, 13(1), 118 to 130. https://doi.org/10.1111/1758-5899.13056 Berrang-Ford, L., Siders, A. R., Lesnikowski, A., et al. (2021). A systematic global stocktake of evidence on human adaptation to climate change. Nature Climate Change, 11(11), 989 to 1000. https://doi.org/10.1038/s41558-021-01170-y Bhandary, R. R., Gallagher, K. S., and Zhang, F. (2021). Climate finance policy in practice: a review of the evidence. Climate Policy, 21(4), 529 to 545. https://doi.org/10.1080/14693062.2020.1871313 Bodansky, D. (2021). Climate change: reversing the past and advancing the future. AJIL Unbound, 115, 80 to 85. https://doi.org/10.1017/aju.2021.4 Boyd, E., Chaffin, B. C., Dorkenoo, K., et al. (2021). Loss and damage from climate change: a new climate justice agenda. One Earth, 4(10), 1365 to 1370. https://doi.org/10.1016/j.oneear.2021.09.015 Chan, S., Boran, I., van Asselt, H., et al. (2021). Climate ambition and sustainable development for a new decade: a catalytic framework. Global Policy, 12(3), 245 to 259. https://doi.org/10.1111/1758-5899.12932 Colenbrander, S., Pettinotti, L., and Cao, Y. (2022). A Fair Share of Climate Finance? An Appraisal of Past Performance, Future Pledges and Prospective Contributors. ODI, London. Depledge, J. (2022). The top down Kyoto Protocol? Exploring caricature and misrepresentation in literature on global climate change governance. International Environmental Agreements: Politics, Law and Economics, 22(4), 673 to 692. https://doi.org/10.1007/s10784-022-09580-9 Dubash, N. K. (2021). Varieties of climate governance: the emergence and functioning of climate institutions. Environmental Politics, 30(S1), 1 to 25. https://doi.org/10.1080/09644016.2021.1979778 Eicke, L., Weko, S., Apergi, M., and Marian, A. (2021). Pulling up the carbon ladder? Decarbonization, dependence, and third country risks from the European carbon border adjustment mechanism. Energy Research and Social Science, 80, 102240. https://doi.org/10.1016/j.erss.2021.102240 Eskander, S. M. S. U., and Fankhauser, S. (2020). Reduction in greenhouse gas emissions from national climate legislation. Nature Climate Change, 10(8), 750 to 756. https://doi.org/10.1038/s41558-020-0831-z Falkner, R., and Buzan, B. (Eds.). (2022). Great Powers, Climate Change, and Global Environmental Responsibilities. Oxford University Press, Oxford. Fankhauser, S., Smith, S. M., Allen, M., et al. (2022). The meaning of net zero and how to get it right. Nature Climate Change, 12(1), 15 to 21. https://doi.org/10.1038/s41558-021-01245-w Forster, P. M., Smith, C., Walsh, T., et al. (2024). Indicators of Global Climate Change 2023: annual update of key indicators of the state of the climate system and human influence. Earth System Science Data, 16(6), 2625 to 2658. https://doi.org/10.5194/essd-16-2625-2024 Friedlingstein, P., O'Sullivan, M., Jones, M. W., et al. (2023). Global Carbon Budget 2023. Earth System Science Data, 15(12), 5301 to 5369. https://doi.org/10.5194/essd-15-5301-2023 Gupta, J., Liverman, D., Prodani, K., et al. (2023). Earth system justice needed to identify and live within Earth system boundaries. Nature Sustainability, 6(6), 630 to 638. https://doi.org/10.1038/s41893-023-01064-1 Hale, T. (2024). Long Problems: Climate Change and the Challenge of Governing across Time. Princeton University Press, Princeton. Hohne, N., Gidden, M. J., den Elzen, M., et al. (2021). Wave of net zero emission targets opens window to meeting the Paris Agreement. Nature Climate Change, 11(10), 820 to 822. https://doi.org/10.1038/s41558-021-01142-2 Intergovernmental Panel on Climate Change. (2022). Climate Change 2022: Mitigation of Climate Change. Contribution of Working Group III to the Sixth Assessment Report. Cambridge University Press, Cambridge. https://doi.org/10.1017/9781009157926 Intergovernmental Panel on Climate Change. (2023). Climate Change 2023: Synthesis Report. Contribution of Working Groups I, II and III to the Sixth Assessment Report. IPCC, Geneva. https://doi.org/10.59327/IPCC/AR6-9789291691647 International Court of Justice. (2025). Obligations of States in Respect of Climate Change. Advisory Opinion of 23 July 2025. International Court of Justice, The Hague. International Energy Agency. (2023). Net Zero Roadmap: A Global Pathway to Keep the 1.5 Degree Celsius Goal in Reach. IEA, Paris. International Energy Agency. (2024). Renewables 2024: Analysis and Forecast to 2030. IEA, Paris. Jewell, J., and Cherp, A. (2023). The feasibility of climate action: bridging the inside and the outside view through feasibility spaces. WIREs Climate Change, 14(5), e838. https://doi.org/10.1002/wcc.838 Kinley, R., Zammit Cutajar, M., de Boer, Y., and Figueres, C. (2021). Beyond good intentions, to urgent action: former UNFCCC leaders take stock of thirty years of international climate change negotiations. Climate Policy, 21(5), 593 to 603. https://doi.org/10.1080/14693062.2020.1860567 Kreibich, N., and Hermwille, L. (2021). Caught in between: credibility and feasibility of the voluntary carbon market post 2020. Climate Policy, 21(7), 939 to 957. https://doi.org/10.1080/14693062.2021.1948384 Lamb, W. F., Wiedmann, T., Pongratz, J., et al. (2021). A review of trends and drivers of greenhouse gas emissions by sector from 1990 to 2018. Environmental Research Letters, 16(7), 073005. https://doi.org/10.1088/1748-9326/abee4e Mace, M. J., Fyson, C. L., Schaeffer, M., and Hare, W. L. (2021). Large scale carbon dioxide removal to meet the 1.5 degree Celsius limit: key governance gaps, challenges and priority responses. Global Policy, 12(S1), 67 to 81. https://doi.org/10.1111/1758-5899.12921 Meinshausen, M., Lewis, J., McGlade, C., et al. (2022). Realization of Paris Agreement pledges may limit warming just below 2 degrees Celsius. Nature, 604(7905), 304 to 309. https://doi.org/10.1038/s41586-022-04553-z Nascimento, L., Kuramochi, T., Iacobuta, G., et al. (2022). Twenty years of climate policy: G20 coverage and gaps. Climate Policy, 22(2), 158 to 174. https://doi.org/10.1080/14693062.2021.1993776 Newell, P. (2021). Power Shift: The Global Political Economy of Energy Transitions. Cambridge University Press, Cambridge. https://doi.org/10.1017/9781108966184 Obergassel, W., Hermwille, L., and Oberthur, S. (2022). From regime building to implementation: harnessing the UN climate conferences to drive climate action. WIREs Climate Change, 13(6), e797. https://doi.org/10.1002/wcc.797 Pauw, W. P., Moslener, U., Zamarioli, L. H., et al. (2022). Post 2025 climate finance target: how much more and how much better? Climate Policy, 22(9 to 10), 1241 to 1251. https://doi.org/10.1080/14693062.2022.2114985 Rajamani, L., Jeffery, L., Hohne, N., et al. (2021). National fair shares in reducing greenhouse gas emissions within the principled framework of international environmental law. Climate Policy, 21(8), 983 to 1004. https://doi.org/10.1080/14693062.2021.1970504 Rajamani, L., and Peel, J. (Eds.). (2021). The Oxford Handbook of International Environmental Law (2nd ed.). Oxford University Press, Oxford. Ripple, W. J., Wolf, C., Gregg, J. W., et al. (2024). The 2024 state of the climate report: perilous times on planet Earth. BioScience, 74(12), 812 to 824. https://doi.org/10.1093/biosci/biae087 Roberts, J. T., Weikmans, R., Robinson, S. A., et al. (2021). Rebooting a failed promise of climate finance. Nature Climate Change, 11(3), 180 to 182. https://doi.org/10.1038/s41558-021-00990-2 Rogelj, J., Fransen, T., den Elzen, M. G. J., et al. (2023). Credibility gap in net zero climate targets leaves world at high risk. Science, 380(6649), 1014 to 1016. https://doi.org/10.1126/science.adg6248 Rogelj, J., Geden, O., Cowie, A., and Reisinger, A. (2021). Three ways to improve net zero emissions targets. Nature, 591(7850), 365 to 368. https://doi.org/10.1038/d41586-021-00662-3 Savaresi, A., and Setzer, J. (2022). Rights based litigation in the climate emergency: mapping the landscape and new knowledge frontiers. Journal of Human Rights and the Environment, 13(1), 7 to 34. https://doi.org/10.4337/jhre.2022.01.01 Schleussner, C. F., Ganti, G., Rogelj, J., and Gidden, M. J. (2022). An emission pathway classification reflecting the Paris Agreement climate objectives. Communications Earth and Environment, 3, 135. https://doi.org/10.1038/s43247-022-00467-w Setzer, J., and Higham, C. (2024). Global Trends in Climate Change Litigation: 2024 Snapshot. Grantham Research Institute on Climate Change and the Environment, London School of Economics and Political Science, London. Skovgaard, J. (2021). The Economisation of Climate Change: How the G20, the OECD and the IMF Contribute to Climate Governance. Cambridge University Press, Cambridge. Timperley, J. (2021). The broken 100 billion dollar promise of climate finance, and how to fix it. Nature, 598(7881), 400 to 402. https://doi.org/10.1038/d41586-021-02846-3 Torstad, V. (2020). Participation, ambition and compliance: can the Paris Agreement solve the effectiveness trilemma? Environmental Politics, 29(5), 761 to 780. https://doi.org/10.1080/09644016.2019.1710322 United Nations Environment Programme. (2023). Adaptation Gap Report 2023: Underfinanced, Underprepared. UNEP, Nairobi. United Nations Environment Programme. (2024). Emissions Gap Report 2024: No More Hot Air, Please. UNEP, Nairobi. United Nations Environment Programme. (2025). Emissions Gap Report 2025: Off Target. UNEP, Nairobi. United Nations Framework Convention on Climate Change. (2023). Technical Dialogue of the First Global Stocktake: Synthesis Report by the Co-Facilitators. UNFCCC, Bonn. United Nations Framework Convention on Climate Change. (2025). Nationally Determined Contributions under the Paris Agreement: Synthesis Report by the Secretariat. UNFCCC, Bonn. van Asselt, H., and Newell, P. (2022). Pathways to an international agreement to leave fossil fuels in the ground. Global Environmental Politics, 22(4), 28 to 47. https://doi.org/10.1162/glep_a_00674 Victor, D. G., Lumkowsky, M., and Dannenberg, A. (2022). Determining the credibility of commitments in international climate policy. Nature Climate Change, 12(9), 793 to 800. https://doi.org/10.1038/s41558-022-01454-x Voigt, C. (2023). The power of the Paris Agreement in international climate litigation. Review of European, Comparative and International Environmental Law, 32(2), 237 to 249. https://doi.org/10.1111/reel.12514 Weikmans, R., van Asselt, H., and Roberts, J. T. (2020). Transparency requirements under the Paris Agreement and their unlikely impact on strengthening the ambition of nationally determined contributions. Climate Policy, 20(4), 511 to 526. https://doi.org/10.1080/14693062.2019.1695571 World Meteorological Organization. (2025). State of the Global Climate 2024. WMO, Geneva. Hashtags #Paris_Agreement #Climate_Change #COP21 #UNFCCC #Global_Warming #Low_Carbon_Future #Climate_Policy #Sustainability #Environmental_Law #Renewable_Energy #Net_Zero_2050 #Climate_Action #Global_Stocktake #Carbon_Markets #Climate_Governance

The Kyoto Protocol, adopted in the Japanese city of Kyoto on 11 December 1997, was the first treaty in history to place legally binding limits on the greenhouse gas emissions of industrialized states. It did not create a new legal order from nothing. Instead, it turned the general promises of the 1992 United Nations Framework Convention on Climate Change into concrete numbers, deadlines, accounting rules, and consequences. This article examines how that operationalization worked, what it achieved, where it failed, and what it taught the world about designing international environmental agreements. Using an integrative review of primary treaty texts and recent peer reviewed scholarship, the study analyses five dimensions of the Protocol: its legal architecture of quantified targets for Annex B parties, its three market based flexibility mechanisms, its monitoring and compliance system, its equity logic rooted in common but differentiated responsibilities, and its political life cycle from the Berlin Mandate to the Doha Amendment. The article then reviews the empirical evidence on effectiveness, which remains genuinely divided. Some econometric work finds meaningful reductions relative to a no treaty counterfactual, while other work using similar methods finds little detectable effect for the largest emitters. The article argues that this apparent contradiction is best resolved by separating three different questions that are often confused: whether parties complied, whether emissions fell, and whether the Protocol caused the fall. The evidence supports a strong answer to the first, a qualified answer to the second, and an uncertain answer to the third. The article concludes that the Protocol's most durable contribution was infrastructural rather than atmospheric. It built the measurement systems, registries, accounting conventions, market institutions, and review procedures that the Paris Agreement inherited and now depends upon. For students of international law, environmental policy, and global governance, Kyoto remains the essential case study in the difficulty of turning shared scientific knowledge into binding collective obligation. Keywords: Kyoto Protocol, UNFCCC, greenhouse gas mitigation, carbon markets, Clean Development Mechanism, compliance, common but differentiated responsibilities, Paris Agreement, treaty design, climate governance 1. Introduction Every international agreement faces a gap between what it says and what it does. The #UNFCCC, agreed at the Rio Earth Summit in 1992, is a clear example. It set out a powerful objective, namely the stabilization of #greenhouse_gases in the atmosphere at a level that would prevent dangerous human interference with the climate system. It established principles, institutions, and a permanent negotiating forum. What it did not do was tell any country how much to cut, by when, or what would happen if the country failed. The Convention was a framework in the literal sense. It framed a problem and created a room in which to argue about it. The #Kyoto_Protocol was the answer to that gap. Adopted at the third Conference of the Parties in December 1997, it took the Convention's open ended promise and converted it into arithmetic. Thirty eight industrialized countries and the European Community accepted quantified limits on their emissions. The limits were expressed as percentages of a base year, usually 1990. They applied to a defined basket of gases. They were to be met over a defined period, 2008 to 2012. They were legally binding, and a compliance system was created to check them. For the first time, a state could be said, as a matter of treaty law, to have broken or kept a climate promise. This article is written for students who need to understand not only what the Protocol contained but why it was built the way it was, and why so much of the later architecture of #climate_governance still carries its fingerprints. The topic is often taught badly. Textbooks tend to summarize Kyoto in a paragraph, describe it as a top down failure, and move quickly to the Paris Agreement of 2015 as the successor that fixed everything. Recent scholarship has challenged almost every part of that summary. Depledge (2022) shows in detail that the description of Kyoto as a purely top down instrument is a caricature that misrepresents both how the targets were set and how they were implemented. Asadnabizadeh and Moe (2024) show that the market mechanisms Kyoto invented did not disappear, but rather migrated, incompletely and painfully, into the Paris system. Harstad (2023) shows formally that the shift from Kyoto style bargaining to Paris style pledging involved real trade offs rather than a simple improvement. The purpose of this article is therefore threefold. First, it reconstructs the legal and institutional design of the Protocol in enough detail to be useful, without assuming prior legal training. Second, it reviews the empirical evidence on whether the Protocol reduced emissions, and explains honestly why that evidence is contested. Third, it draws out the design lessons that the Protocol offers to anyone thinking about how to make international cooperation on a global commons problem actually work. The central argument is that the Protocol should be judged as an act of institution building rather than as a single policy intervention. Judged as a policy intervention, its record is modest and its causal footprint is hard to isolate. Judged as institution building, it was transformative. It produced the counting rules, the review procedures, the registries, and the market plumbing without which the current climate regime could not function at all. 2. Methodology This study uses an integrative, qualitative review design. It is not an experiment, a survey, or a statistical analysis of original data. It is a structured synthesis of two kinds of material. The first body of material is primary and documentary. It consists of the text of the 1992 Convention, the text of the 1997 Protocol, the decisions of the Conference of the Parties that gave the Protocol operational detail, most importantly the Marrakesh Accords of 2001, and the Doha Amendment of 2012. These documents are the legal facts of the case. Any claim in this article about what the Protocol required is grounded in them. The second body of material is secondary and scholarly. It consists of peer reviewed research published mainly between 2021 and 2025 on the Protocol's design, its mechanisms, its compliance record, and its relationship to the Paris Agreement, together with the assessment findings of the Intergovernmental Panel on Climate Change (IPCC, 2022, 2023). Where the article reports empirical results, such as estimated emission reductions attributable to the treaty, it reports them as findings of specific studies rather than as settled fact, because the literature does not speak with one voice. The selection logic was thematic rather than exhaustive. Sources were included when they addressed at least one of five themes: legal architecture and obligation, market mechanisms and #environmental_integrity, compliance and #transparency, equity and differentiation, or the institutional transition from Kyoto to Paris. Preference was given to recent work, both because older work has been superseded on questions of outcome and because the Protocol's full record only became visible after its second period ended in 2020. Three limitations should be stated plainly at the outset. First, an integrative review inherits the biases of the literature it reviews, and the Kyoto literature is dominated by authors writing in English from European and North American institutions. Second, the article cannot resolve the causal question of the Protocol's effect on emissions, because that question is unresolved in the primary econometric literature itself. It can only map the disagreement and explain its sources. Third, the article does not attempt a country by country legal analysis. It works at the level of regime design. 3. From Rio to Kyoto: Why a Protocol Was Needed The 1992 Convention was drafted at a moment of scientific consensus but political caution. Its drafters knew that a treaty demanding immediate cuts would not be ratified by the countries whose emissions mattered most. So they built a framework convention, a familiar device in international environmental law, in which the parent treaty establishes principles and institutions and leaves the hard commitments to later protocols. The Vienna Convention on ozone, followed by the Montreal Protocol, had used the same design and had worked. The Convention did contain one important obligation for developed countries, an aim to return emissions to 1990 levels by the year 2000. But the language was hedged, the aim was not a binding cap, and by the mid 1990s it was clear that several major economies would miss it. The first Conference of the Parties, held in Berlin in 1995, responded with what became known as the #Berlin_Mandate. The parties agreed to negotiate a protocol with quantified limitation and reduction objectives for developed countries within a defined timeframe, and they agreed explicitly that no new commitments would be introduced for developing countries in this round. That second half of the Berlin Mandate shaped everything that followed. It was a direct application of the Convention's principle of #common_but_differentiated_responsibilities and respective capabilities. The logic was straightforward and, at the time, widely accepted. The rich industrialized world had produced the large majority of the cumulative carbon dioxide already in the atmosphere. It had grown rich by doing so. It had the money and technology to act first. Developing countries had immediate needs of poverty reduction, electrification, and industrialization, and their per person emissions were a small fraction of those in the North. Asking a country with widespread energy poverty to cap its emissions at the same moment as asking the same of a wealthy economy was, on this view, neither fair nor politically possible. The Berlin Mandate also created the political fault line that eventually broke the Protocol's universality. In July 1997, five months before Kyoto, the United States Senate passed the #Byrd_Hagel_Resolution by ninety five votes to zero. The resolution declared that the United States should not sign any climate agreement that imposed binding commitments on developed countries without imposing comparable commitments on developing countries in the same period, or that would seriously harm the American economy. The vote was unanimous across both parties. Any negotiator paying attention knew, in advance, that a Protocol built on the Berlin Mandate could not be ratified in Washington. The Kyoto negotiations therefore took place under a contradiction. The mandate that made agreement possible among the parties was the same mandate that made ratification impossible in the largest emitter of the time. Understanding this is essential. Kyoto's later difficulties were not an accident of implementation. They were written into the terms of reference before the drafting began. 4. The Legal Architecture: Turning a Goal into a Number The core of the Protocol is Article 3, read together with Annex A and Annex B. Article 3 requires the parties listed in #Annex_B to ensure that their aggregate emissions of the listed gases do not exceed their assigned amounts, and it sets the collective objective of reducing overall emissions of those gases by at least five percent below 1990 levels over the commitment period 2008 to 2012. Four design choices in that sentence deserve close attention, because each one was contested and each one had consequences. The choice of a basket. Annex A lists six gases: carbon dioxide, methane, nitrous oxide, hydrofluorocarbons, perfluorocarbons, and sulphur hexafluoride. Rather than setting a separate target for each, the Protocol converts them into a single common unit using global warming potentials, so that a tonne of methane counts as a fixed multiple of a tonne of carbon dioxide. This #basket_of_gases approach gave states flexibility to cut whichever gas was cheapest to cut. It was efficient. It also introduced a technical dependence on global warming potential values that are themselves scientific estimates subject to revision. The choice of a base year. Emissions are measured against a #base_year, normally 1990. This looks neutral. It was not. In 1990 the economies of Eastern Europe and the former Soviet Union were at the end of a period of heavy, inefficient, state directed industrial production. Those economies then collapsed. Their emissions fell dramatically for reasons that had nothing to do with climate policy. Choosing 1990 therefore handed these #economies_in_transition an enormous quantity of surplus permits before the treaty even began, a phenomenon that came to be called #hot_air. Section 9 returns to this problem, which is the single most damaging accounting flaw in the Protocol. The choice of differentiated targets. The targets in Annex B are not uniform. The European Community accepted minus eight percent, the United States minus seven, Japan minus six. Some parties accepted stabilization at 1990 levels, and a few, including Australia and Iceland, were permitted increases. The numbers were the product of hard bargaining, not of any formula. This is one reason Depledge (2022) resists the label of a purely top down treaty. No central authority allocated these figures. Each government negotiated the number it believed it could deliver domestically, which is much closer to a bottom up process than the standard textbook description admits. The choice of a budget rather than an annual cap. The target applies to cumulative emissions across the five years of the commitment period, not to any single year. This turned the obligation into a carbon budget. A party could exceed its notional annual level in one year and compensate in another. Combined with the trading mechanisms, this made the obligation economically manageable while keeping it legally firm. Two further elements complete the architecture. Article 3 allows parties to count net changes in emissions and removals from #LULUCF, that is, land use, land use change, and forestry. This means that a country planting forests could count the resulting #carbon_sinks toward its target. The rules for doing so were among the most technically difficult and politically bitter of the entire negotiation, because sinks are hard to measure, are reversible, and could in principle substitute for real reductions in energy systems. Finally, Article 4 permits parties to fulfil their commitments jointly, a provision used by the European Union to redistribute its collective minus eight percent target internally among its member states, an arrangement known informally as the bubble. The result was a treaty that was legally binding in form but economically flexible in operation. That combination was deliberate. It was also the source of both its intellectual elegance and its practical vulnerability. 5. The Flexibility Mechanisms: Markets Inside a Treaty The most influential and most criticized part of the Protocol is its trio of #flexibility_mechanisms. Their underlying logic comes from environmental economics. A tonne of carbon dioxide has the same effect on the climate wherever it is emitted. If the goal is a fixed global reduction at least cost, then reductions should happen wherever they are cheapest, not wherever they happen to be politically convenient. Allowing countries to buy reductions from one another should therefore lower the total cost of hitting the same environmental target, or equivalently, allow a more ambitious target for the same cost. That claim of #cost_effectiveness is the intellectual foundation of every carbon market that exists today. International Emissions Trading (Article 17). Annex B parties receive #assigned_amount_units equal to their permitted emissions. A party that emits less than its allocation may sell the surplus to a party that has emitted more. This is a classic cap and trade system operating between sovereign states rather than between firms. Joint Implementation (Article 6). An Annex B party may invest in an emission reduction project in another Annex B party and receive emission reduction units generated by that project, which are then subtracted from the host country's assigned amount. Because both countries are capped, the transaction moves permits around without changing the aggregate cap. Joint implementation was the smallest of the three mechanisms in practice. The Clean Development Mechanism (Article 12). This is the mechanism that mattered. Under the #clean_development_mechanism, an entity in an Annex B country may finance an emission reduction project in a developing country that has no cap, and receive #certified_emission_reductions for the emissions avoided. The CDM had two stated purposes, and this dual purpose is essential to understanding both its appeal and its problems. It was meant to help Annex B parties meet their targets at lower cost, and it was meant to assist developing countries in achieving #sustainable_development. The CDM was, in effect, a large scale experiment in linking climate finance to development. It offered developing countries a channel for #technology_transfer and investment without accepting an emissions cap. It offered developed countries cheaper compliance. And it created, almost as a by product, a global #carbon_market with prices, brokers, verifiers, auditors, and financial products. Because CDM projects take place outside any cap, the mechanism raises a problem that capped trading does not. If a project would have happened anyway, and credits are issued for it, then the buyer has legally offset real emissions against reductions that were never additional. Total emissions rise. This is the problem of #additionality, and it is the permanent, unsolved difficulty of every offset system. The Protocol's answer was procedural. A project must demonstrate, against an approved #baseline methodology, that its reductions are additional to what would have occurred in its absence, and this demonstration must be validated by an accredited independent auditor and registered by an Executive Board. The detailed rules for all three mechanisms, together with the compliance system and accounting procedures, were not settled in 1997. They were negotiated over four more years and finalized as the #Marrakesh_Accords in 2001. This gap matters. The Protocol was adopted in 1997, made operational on paper in 2001, and entered into force in 2005. Eight years passed between the political agreement and the first day of legal effect. 6. Compliance, Measurement, and the Machinery of Trust An emissions target is meaningless unless someone can verify it. This is where the Protocol made contributions that outlived every other part of it. The starting point is measurement. Annex I parties must submit annual #national_inventories of emissions and removals, prepared according to agreed methodologies, covering all sources and sinks. These inventories are then subjected to review by international #expert_review_teams, which examine the data, question the assumptions, and publish their findings. This creates an obligation not merely to reduce emissions but to be able to prove, in public and to a technical standard, what one's emissions are. Before Kyoto, most countries simply did not know their national emissions with any precision. The Protocol forced them to find out, and to keep finding out, every year. The second element is the tracking of units. Because the mechanisms create tradable instruments, there must be a way to ensure that a single tonne is not sold twice. The Protocol's answer was a system of national registries linked to an #international_transaction_log maintained by the secretariat, which checks every proposed transaction for consistency with the rules before it is allowed to complete. This is the direct institutional ancestor of the accounting infrastructure that today underpins Article 6 of the Paris Agreement, and it is why the problem of #double_counting is now technically manageable even where it remains politically contested. The third element is the #compliance_committee, established under the Marrakesh Accords. Its structure is the most legally interesting feature of the entire regime. It has two branches. The #facilitative_branch is designed to help parties that appear likely to fall short, through early warning, advice, and the facilitation of finance and technical assistance. The #enforcement_branch determines whether a party has actually failed to meet its target and applies consequences. Those consequences are unusual. A party found in non compliance must make up the shortfall in the next commitment period with a penalty rate of thirty percent, must submit a compliance action plan, and loses eligibility to sell units under emissions trading until it is reinstated. This is not enforcement in the sense that a domestic court enforces a contract. There is no fine, no seizure of assets, no coercive power. The consequences are what international lawyers call reputational and participatory. A non compliant party is named, is required to explain itself, and is excluded from the market it wishes to use. The design assumes that states care about their standing, about their access to institutions, and about being seen to keep their word. Whether that assumption holds is exactly what the Canadian withdrawal tested, and Section 8 examines the result. 7. Equity, Differentiation, and the Fault Line of Fairness No aspect of the Protocol has generated more argument than its treatment of #equity. The Annex based structure divides the world in two. Annex B parties have quantified obligations. Everyone else does not. There is no intermediate category and no mechanism for moving between them. The case for this structure rests on #historical_responsibility and capability. Cumulative emissions, not annual emissions, drive warming, because carbon dioxide persists in the atmosphere for a very long time. On a cumulative basis, the industrialized world was overwhelmingly responsible for the problem as it stood in 1997. It also possessed the wealth, the institutions, and the technology to respond. To require identical action from a country with a tenth of the income and a twentieth of the per person emissions would have been to freeze existing inequality in place under an environmental justification. This is the argument from #climate_justice, and it remains, in its core logic, powerful. The case against the structure is not that differentiation is wrong but that this particular form of differentiation was static. The Annexes were fixed in 1992 and were never revised. The world was not. Over the Protocol's lifetime, the emissions of several large developing economies grew enormously, and China became the largest annual emitter on the planet. A treaty that permanently exempted the fastest growing sources of the problem from any quantitative obligation could not, by construction, solve the problem, however fair its intentions. The binary division also produced a perverse dynamic. It gave capped countries a strong reason to worry about #carbon_leakage, the migration of emissions intensive industry to uncapped jurisdictions, and it gave domestic opponents of climate policy a permanent and effective argument, namely that unilateral action would export jobs without reducing global emissions. Recent scholarship has grown more careful about this debate. Aklin and Mildenberger (2020) argue that the standard framing of climate change as a global collective action problem, in which the main obstacle is #free_riding by other states, is substantially wrong. In their account, the decisive obstacles are domestic and distributive. Climate policy creates winners and losers inside each country, and it is the losers, usually well organized and well financed incumbents, who block action. On this reading, the Byrd Hagel Resolution was not really about what China would or would not do. It was about what American coal, oil, and manufacturing interests would tolerate, dressed in the language of international fairness. This matters for how students should interpret the Protocol's failure to secure universal participation. It was not simply that the equity formula was too generous to developing countries. It was that in the country whose participation mattered most, the domestic political economy would have rejected almost any binding cap, whatever formula had been chosen. 8. Ratification, Entry into Force, and the Doha Amendment The Protocol's #ratification story is one of the strangest in modern treaty practice. Adoption in 1997 required only agreement among negotiators. #entry_into_force required ratification by at least fifty five parties, including Annex I parties accounting for at least fifty five percent of Annex I carbon dioxide emissions in 1990. The United States signed the Protocol under President Clinton but never submitted it to the Senate, where the Byrd Hagel vote had already made its fate obvious. In 2001, the incoming administration announced that it would not implement the treaty. Because the United States accounted for roughly a third of Annex I emissions in the 1990 baseline, its absence made the fifty five percent threshold arithmetically dependent on the participation of nearly every other significant Annex I party, and above all on Russia. Russia thus acquired decisive leverage over the treaty's existence, and used it. Ratification came in late 2004, and the Protocol entered into force on 16 February 2005, more than seven years after adoption. The delay had a compounding effect. The commitment period began in 2008, only three years later. Countries had known their targets since 1997, but the legal obligation to meet them arrived very late in the day, which weakened the incentive to begin the slow work of transforming energy systems early. The second shock was #withdrawal. Canada, facing a large shortfall against its target and the prospect of the compliance consequences described in Section 6, formally withdrew from the Protocol in 2011, with effect at the end of 2012. This is the crucial test case for the theory that reputational and participatory consequences are sufficient to secure compliance. The answer the case gives is uncomfortable. When the cost of compliance exceeded what a government was willing to pay, exit was available, lawful, and chosen. A compliance system that a party can escape simply by leaving is not, in the end, a constraint on that party. The third episode is the #Doha_Amendment of 2012, which established a #second_commitment_period running from 2013 to 2020, with a collective target of at least eighteen percent below 1990 levels and the addition of nitrogen trifluoride to the basket of gases. But the second period was a shadow of the first. Japan, Russia, New Zealand, and Canada did not take on targets. What remained was essentially the European Union, Australia, and a small group of others. The share of global emissions covered by binding targets fell to a small fraction, widely estimated at well under fifteen percent. The Doha Amendment did not receive enough ratifications to enter into force until 31 December 2020, which is to say, the day the period it governed came to an end. That fact is worth pausing on. The second commitment period of the Kyoto Protocol became legally effective on the last day of its own existence. It is difficult to imagine a clearer illustration of the gap between international legal process and physical atmospheric time. 9. Did It Work? The Contested Empirics of Effectiveness Students often ask a simple question about Kyoto: did it reduce emissions? The honest answer requires separating three questions that are routinely blurred together. Question one: did parties comply? On the legal record, yes, and overwhelmingly so. Parties with targets in the first commitment period, as a group, ended the period with emissions below their assigned amounts. The compliance system functioned. Inventories were submitted, reviews were conducted, and the enforcement branch did not need to impose penalties on a broad scale. Measured as a test of whether states can be made to keep a quantified environmental promise, the first commitment period was a success. Question two: did emissions fall? In the Annex B group, yes. But the reasons are heavily contaminated. Three factors dominate. The first is the collapse of the post Soviet economies, which delivered enormous reductions relative to 1990 for reasons that predate the treaty entirely. The second is the global financial crisis of 2008 and 2009, which happened to coincide exactly with the start of the commitment period and depressed industrial output across the developed world. The third is the shift from coal to gas in several major economies, driven mainly by relative fuel prices. None of these is a policy effect of the Protocol. Question three: did the Protocol cause the reductions? Here the literature genuinely disagrees, and it is important to represent the disagreement fairly rather than to pick a convenient side. One strand of econometric work, using #counterfactual methods to construct a synthetic version of each treaty country from a weighted combination of untreated countries, finds a measurable treaty effect, with ratifying countries emitting several percent less than the constructed no treaty scenario. Another strand, using closely related #synthetic_control techniques, finds very little evidence of any reduction effect among the major Annex B emitters and warns that country level evaluation of international environmental policy is beset by methodological traps that can generate spurious findings. Both strands are published in serious journals. Both are methodologically careful. They reach opposite conclusions. Why? Three reasons. First, the treatment is not clean. Countries did not receive Kyoto in the way a patient receives a drug. Many were already adopting climate policies, and the European Union in particular was moving toward its own emissions trading system regardless. Second, the control group is not clean either. Non ratifying and non Annex B countries were affected by the same global energy prices, the same technological changes, and the same financial crisis. Third, and most fundamentally, the Protocol's effect may operate through channels that country level emissions data cannot capture, such as the diffusion of policy instruments, the creation of measurement capacity, and the normalization of carbon pricing. This third channel is where more recent work is most illuminating. Stechemesser and colleagues (2024), in a systematic ex post evaluation of around fifteen hundred climate policies implemented across forty one countries between 1998 and 2022, identify a modest number of cases of large emission reductions and find that they were typically produced by combinations of instruments rather than by single measures, with pricing instruments frequently playing a role within those combinations. The IPCC's Sixth Assessment on mitigation similarly attributes meaningful avoided emissions to the accumulated body of national climate policies adopted over this period, a body of policy that Kyoto helped to catalyse even where the treaty itself cannot be credited with the reductions directly. Set against this, work on the effect of #carbon_price signals is sobering. Lilliestam, Patt, and Bersalli (2021, 2022) review the empirical ex post evidence and find that while carbon pricing schemes have sometimes been associated with emission reductions, there is no persuasive evidence that they have triggered the deep, zero carbon investment and technological change that decarbonization requires. If correct, this cuts at the heart of Kyoto's economic theory. The mechanisms may have delivered cheap short term abatement while doing little to shift the long term technological trajectory, and cheap short term abatement can even entrench existing carbon intensive infrastructure by making it marginally cleaner rather than replacing it. Overlaying all of this is the #hot_air problem. Because the 1990 baseline handed the economies in transition a large surplus of assigned amount units that corresponded to no real abatement effort, and because those units could be sold, the effective stringency of the aggregate cap was lower than the headline numbers suggested. Every hot air unit sold and used for compliance permitted a real tonne to be emitted elsewhere against a reduction that had already happened for unrelated reasons. Some purchasing countries and the European Union restricted the use of such units, which limited the damage, but the design flaw was real and it is the clearest single lesson in baseline setting that the Protocol offers. The defensible summary is this. The Protocol produced full legal compliance, coincided with real reductions in the covered group, and probably contributed to those reductions through the policies it induced, but its independent causal contribution cannot be cleanly measured and was almost certainly smaller than its supporters claimed and larger than its harshest critics allow. 10. The Clean Development Mechanism in Practice The CDM deserves separate treatment because it was, by volume, the largest environmental market institution ever created, and because its afterlife is now central to the Paris regime. At its peak the mechanism registered thousands of projects. The literature reports on the order of eight thousand registered activities and close to two billion certified emission reductions issued over its lifetime. Renewable energy, industrial gas destruction, methane capture, and energy efficiency dominated the portfolio. Measured as a piece of institutional engineering, this is remarkable. A United Nations body, working through accredited private auditors and national designated authorities, verified emission reductions in dozens of countries and issued tradable instruments against them, and those instruments were bought, sold, and surrendered for legal compliance. Measured against its two stated purposes, the record is far more mixed, and four criticisms recur in the literature. Geographic concentration. Projects clustered heavily in a small number of large host economies, with China, India, Brazil, and Mexico accounting for the great majority of registered activities, while sub Saharan Africa hosted only a very small share. The reason is structural rather than malicious. The CDM was a market, and markets flow toward low transaction costs, large project pipelines, established financial institutions, and reliable regulatory counterparties. The countries with the greatest need for #capacity_building and investment were precisely the countries least able to compete for it. Sustainable development as a residual. The mechanism made emission reductions measurable, verifiable, and tradable. It made development benefits none of these things. Host countries were left to certify for themselves that a project contributed to sustainable development, with no common standard and no international review. Predictably, the dimension that was measured, priced, and audited crowded out the dimension that was not. Reviews of the community level co benefits of CDM projects find that outcomes were highly variable and frequently modest, and that projects with the largest credit yields were often not those with the largest local development impact. Additionality and environmental integrity. Because credits are issued against a hypothetical baseline, the integrity of the entire system rests on a counterfactual judgement that can never be directly observed. Where baselines were set generously, or where projects would have proceeded anyway, credits were issued for reductions that did not exist in any real sense, and their use for compliance increased net emissions. This is not a marginal technical concern. It is the central risk of every offsetting arrangement, and it is why Ahonen and colleagues (2022) treat the governance of environmental integrity across fragmented compliance and voluntary markets as the defining problem of carbon markets under the Paris Agreement. Price collapse. The market depended on demand from capped parties. When the second commitment period shrank to a small group and the European market restricted the use of international credits, demand evaporated. Credit prices collapsed to a level at which the mechanism could not finance anything. Thousands of projects were stranded. This taught an unglamorous but vital lesson: an offset market is entirely a creature of the compliance demand created by regulation. Without a binding cap somewhere, a carbon credit is worth nothing. Alongside the criticisms, one genuine achievement should be recorded. The CDM levied a share of proceeds on issued credits, which financed the #Adaptation_Fund. This established, for the first time, an automatic and non discretionary channel of #climate_finance from mitigation activity in one part of the world to adaptation needs in another. The mechanism was small in scale but conceptually important, and its logic survives in the Paris architecture. 11. The Top Down Caricature and What It Obscures The standard narrative says that Kyoto was a top down treaty in which targets were imposed from above, that this design failed, and that Paris succeeded by replacing it with a bottom up system of voluntary national pledges. Depledge (2022) subjects this narrative to close examination and finds it seriously misleading in at least three respects. First, as noted in Section 4, the Annex B targets were not imposed by any central authority. They were negotiated by governments on the basis of what each believed it could deliver at home. That is a bottom up process expressed through a binding legal form. Second, the Protocol left the choice of domestic policy instruments entirely to each party. It said what the outcome had to be, not how to achieve it. A party could use a carbon tax, an emissions trading system, regulation, subsidy, or any combination. This is precisely the kind of national discretion that the Paris Agreement is praised for. Third, the Protocol's own architecture already contained the review and adjustment features that are usually presented as Paris innovations. Commitment periods were designed to be successive, so that ambition could be revisited. Inventories, expert review, and public reporting created a continuous cycle of scrutiny. Why does this matter for students? Because the caricature encourages a false inference: that binding obligation is the thing that failed, and that voluntarism is therefore the solution. The evidence does not support that inference. What failed at Kyoto was not bindingness. Bindingness worked. Parties that stayed in complied. What failed was participation, coverage, and baseline design. The United States never joined, key parties left, developing country emissions were not covered, and the 1990 baseline leaked hot air into the system. The Paris Agreement fixed participation and coverage. It did so by giving up on bindingness of outcome. That is not an unambiguous improvement. It is a trade. Harstad (2023) models this trade formally, comparing a bargaining regime of the Kyoto type against a #pledge_and_review regime of the Paris type, and derives predictions consistent with the observed transition, including the changes in participation levels, in the depth of commitments, and in incentives to invest in low carbon technology. The point of such analysis is not to declare a winner but to make the trade explicit. Broader participation was purchased with weaker commitment, and weaker commitment has consequences for the incentive to invest in the technologies that actually decarbonize an economy. Asadnabizadeh and Moe (2024) reach a compatible conclusion from a different direction. Reviewing the literature on global carbon markets from Kyoto through Paris, they document a clear historical drift away from top down and toward bottom up approaches, and toward national rather than international level solutions, while noting that the international market project itself has repeatedly failed to reach implementation. The mechanisms Kyoto built were not replaced by something better. They were left in an unfinished state for two decades while negotiators argued. 12. From Kyoto to Paris: Inheritance, Not Replacement The Paris Agreement of 2015 is usually described as a new beginning. It is more accurate to describe it as a new political settlement built on old technical foundations. What Paris changed is the nature of the obligation. Every party, developed and developing, submits a #nationally_determined_contribution. The content of that contribution is determined nationally, not negotiated internationally. The obligation is to prepare, communicate, and maintain a contribution and to pursue domestic measures aimed at achieving it. The obligation is not to achieve it. Universal participation was bought at the price of enforceable outcome. Differentiation survives, but in a softer and more flexible form. A #ratchet_mechanism requires successive contributions to represent a progression in ambition, and a periodic #global_stocktake assesses collective progress and is intended to inform the next round of pledges. What Paris inherited is almost everything else. The #Enhanced_Transparency_Framework, under which parties report inventories and progress and submit to technical expert review, is a direct descendant of the Kyoto reporting and review system, extended to all parties and adapted for flexible capacity. Without three decades of accumulated inventory methodology, national statistical capacity, and expert review practice, the framework could not exist. That capacity was built under Kyoto. The same is true of markets. #Article_6 of the Paris Agreement provides for cooperative approaches involving internationally transferred mitigation outcomes, and for a centralized crediting mechanism now known as the #Paris_Agreement_Crediting_Mechanism. Its supervisory body has spent years developing standards for additionality, baselines, leakage, and permanence, and has established procedures for transitioning eligible CDM activities and methodologies into the new mechanism. The intellectual and administrative lineage is unbroken. The problems are also unbroken. Additionality remains contested, baseline setting remains the crucial integrity risk, and the requirement for #corresponding_adjustments to prevent double counting between buyer and seller countries is the direct technical descendant of the Kyoto registry and transaction log system. There is one further inheritance, and it is not encouraging. The Kyoto compliance committee's work has now essentially concluded, and its counterpart under the Paris Agreement is a facilitative and non punitive body. The enforcement branch, with its thirty percent penalty and suspension of trading eligibility, has no equivalent in the Paris system. The most legally innovative instrument the climate regime ever built has not been carried forward. The political inheritance is likewise sobering. In 2025 the United States again began the process of withdrawing from the central climate treaty of its era, as it declined to ratify the central climate treaty of the previous era. The structural problem that defeated Kyoto, namely the domestic political economy of the world's largest economies, has not been solved by any change in treaty design. 13. Lessons for Treaty Design and Global Environmental Governance The Protocol's greatest value to students is as a source of transferable design lessons. Eight stand out. Lesson one: legal bindingness and political feasibility trade against each other. Kyoto chose depth of commitment among a coalition of the willing. Paris chose breadth of participation with shallow commitment. Neither has yet produced adequate mitigation. The lesson is not that one is right, but that #treaty_design cannot escape this trade off by clever drafting. It can only choose where to sit on it, and the choice must be made with open eyes. Lesson two: baselines are where integrity is won or lost. The hot air problem was not a failure of enforcement. It was a failure of arithmetic, decided in 1997 and impossible to correct afterwards. In any crediting or capping system, whether Kyoto, Article 6, or a voluntary market, the reference point against which performance is measured is the point of maximum vulnerability. Lesson three: an exit option undermines an enforcement mechanism. The Canadian withdrawal demonstrates that consequences a party can avoid by leaving are not consequences at all. Any future regime that seeks real accountability must make continued membership more valuable than escape, whether through market access, finance, trade linkage, or reciprocal benefit. Lesson four: what gets measured is what gets delivered. The CDM measured tonnes precisely and development loosely, and it delivered tonnes and neglected development. If a treaty has two objectives, both must be given measurable form, or the measurable one will consume the other. Lesson five: markets require regulatory demand. The collapse of the credit price after 2012 shows that an offset market has no independent existence. It lives entirely on the compliance demand that regulation creates. Building the market first and hoping demand will follow inverts the logic. Lesson six: static categories fail in a dynamic world. The Annex structure of 1992 was a reasonable reflection of the world of 1992 and an increasingly poor reflection of the world of 2012. Differentiation was and remains right in principle. Fixed and unrevisable differentiation was the error. Lesson seven: infrastructure outlives instruments. Kyoto's targets expired. Its inventories, review teams, registries, accounting rules, and market institutions did not. When judging an international agreement, ask what capacity it leaves behind, not only what it achieved during its own life. Lesson eight: international agreements do not by themselves defeat domestic opposition. The Aklin and Mildenberger argument deserves to be taken seriously. If the binding constraint is distributive conflict inside states rather than free riding between them, then #policy_effectiveness depends on building domestic coalitions that benefit from #decarbonization, and international treaties are useful mainly insofar as they help those coalitions win at home. 14. Limitations and Directions for Future Research Three limitations of the present analysis, and of the field it reviews, should be acknowledged. First, the causal question remains open. Until the econometric literature converges on a defensible identification strategy for international environmental agreements, claims that Kyoto did or did not reduce emissions should be treated as contested rather than established. A useful contribution from a student researcher would be a careful replication and comparison of the leading counterfactual studies, examining precisely which modelling choices drive the divergence in their conclusions. Second, the literature is unbalanced. Research on the Protocol is dominated by scholars in Europe and North America. The experience of host countries under the CDM, the reasons for the near absence of African participation, and the institutional legacies left in developing country administrations are all under researched relative to their importance. Third, the transition of Kyoto era mechanisms into the Paris Agreement Crediting Mechanism is happening now, and its outcome will determine whether the integrity failures of the CDM are repeated. This is the most valuable open research area in the field, and it is one where new work is genuinely needed rather than merely welcome. Two further questions deserve attention. What explains variation in domestic implementation among Annex B parties that faced comparable targets? And what would a compliance mechanism with real consequences look like in a regime built on nationally determined pledges, given that the enforcement branch model has been abandoned rather than reformed? 15. Conclusion The Kyoto Protocol converted a framework into an obligation. That is its historical function and its lasting significance. Before 1997, no state had ever accepted a quantified, legally binding, internationally verified limit on its greenhouse gas emissions. After 1997, thirty eight of them had. The treaty's shortcomings are real and should not be minimized. Its largest intended participant never joined. One party left rather than comply. Its baseline leaked surplus permits into the market. Its offset mechanism concentrated investment where it was easiest rather than where it was most needed, and let development benefits fall away. Its second period entered into force on the day it expired. Global emissions rose throughout. And yet the case against Kyoto is frequently overstated, and overstated in a way that leads to the wrong lesson. The parties that remained under the treaty complied with it. The compliance and review machinery worked. The measurement systems it created are the reason the world can now discuss national emissions as facts rather than as assertions. The market institutions it built are being rebuilt, under new names, inside the Paris Agreement, because no one has designed anything better. The #ambition_gap that remains between current pledges and the trajectory required for #net_zero is not evidence that Kyoto's approach was misconceived. It is evidence that the underlying political problem, which is distributive, domestic, and deeply entrenched, has never been solved by any treaty of any design. For students, the correct conclusion is neither celebration nor dismissal. It is that the Kyoto Protocol was a serious, sophisticated, and partially successful attempt to solve the hardest coordination problem humanity has faced, and that its failures were mostly failures of politics rather than of legal imagination. The next generation of climate agreements will be built out of Kyoto's parts. It is worth understanding how those parts were made. #KyotoProtocol1997 #ClimateTreatyDesign #GHG_Mitigation #InternationalClimateLaw #CarbonMarketsResearch #UNFCCC_Implementation #FromKyotoToParis #ClimatePolicyStudies #GlobalEnvironmentalGovernance #EmissionsAccountability #AnnexICountries #ClimateDiplomacy #CleanDevelopmentMechanism_Review #ClimateEquityDebate #StudentResearchArticle References Ahonen, H. M., Kessler, J., Michaelowa, A., Espelage, A., and Hoch, S. (2022). Governance of fragmented compliance and voluntary carbon markets under the Paris Agreement. Politics and Governance, 10(1). doi:10.17645/pag.v10i1.4759 Aklin, M., and Mildenberger, M. (2020). Prisoners of the wrong dilemma: why distributive conflict, not collective action, characterizes the politics of climate change. Global Environmental Politics, 20(4), 4 to 27. doi:10.1162/glep_a_00578 Asadnabizadeh, M., and Moe, E. (2024). A review of global carbon markets from Kyoto to Paris and beyond: the persistent failure of implementation. Frontiers in Environmental Science, 12, 1368105. doi:10.3389/fenvs.2024.1368105 Depledge, J. (2022). The top down Kyoto Protocol? Exploring caricature and misrepresentation in literature on global climate change governance. International Environmental Agreements: Politics, Law and Economics, 22(4), 673 to 692. doi:10.1007/s10784-022-09580-9 Harstad, B. (2023). Pledge and review bargaining: from Kyoto to Paris. The Economic Journal, 133(651). Intergovernmental Panel on Climate Change (2022). Climate Change 2022: Mitigation of Climate Change. Contribution of Working Group III to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge: Cambridge University Press. Intergovernmental Panel on Climate Change (2023). Climate Change 2023: Synthesis Report. Contribution of Working Groups I, II and III to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. Geneva: IPCC. Lilliestam, J., Patt, A., and Bersalli, G. (2021). The effect of carbon pricing on technological change for full energy decarbonization: a review of empirical ex post evidence. WIREs Climate Change, 12(1), e681. doi:10.1002/wcc.681 Lilliestam, J., Patt, A., and Bersalli, G. (2022). On the quality of emission reductions: observed effects of carbon pricing on investments, innovation, and operational shifts. Environmental and Resource Economics, 83, 733 to 758. doi:10.1007/s10640-022-00708-8 Mor, S., Aneja, R., Madan, S., and Ghimire, M. (2024). Kyoto Protocol and Paris Agreement: transition from bindings to pledges, a review. Millennial Asia. doi:10.1177/09763996221141546 Rajamani, L., and Peel, J. (Eds.) (2021). The Oxford Handbook of International Environmental Law. Second edition. Oxford: Oxford University Press. Stechemesser, A., Koch, N., Mark, E., Dilger, E., Kloesel, P., Menicacci, L., Nachtigall, D., Pretis, F., Ritter, N., Schwarz, M., Vossen, H., and Wenzel, A. (2024). Climate policies that achieved major emission reductions: global evidence from two decades. Science, 385(6711), 884 to 892. doi:10.1126/science.adl6547 United Nations Environment Programme (2024). Emissions Gap Report 2024. Nairobi: United Nations Environment Programme.

This article examines the Treaty for the Prohibition of Nuclear Weapons in Latin America and the Caribbean, signed in Mexico City on 14 February 1967 and known as the #Treaty_of_Tlatelolco. It was the first agreement to establish a nuclear-weapon-free zone across a large, populated region, and it created a permanent regional body, the Agency for the Prohibition of Nuclear Weapons in Latin America and the Caribbean (OPANAL), to supervise compliance. The article traces the origins of the treaty in the aftermath of the #Cuban_Missile_Crisis, reconstructs the negotiating process inside the Preparatory Commission for the Denuclearization of Latin America (#COPREDAL), analyses the legal architecture of the instrument and its two Additional Protocols, and evaluates the slow but eventual movement from a partial zone to a fully universal one across all thirty-three states of the region. It then assesses the treaty as a model that shaped later zones in the South Pacific, Southeast Asia, Africa and Central Asia, and as a precedent that informed both the Treaty on the Non-Proliferation of Nuclear Weapons (NPT) and the Treaty on the Prohibition of Nuclear Weapons (TPNW). The final sections consider contemporary pressures on the regime, including naval nuclear propulsion programmes, unresolved interpretive declarations attached by nuclear-armed states, renewed #great_power_competition, and the United Nations decision in December 2024 to commission the first comprehensive study of #nuclear_weapon_free_zones since 1975. The article argues that the durability of Tlatelolco rests less on its prohibitions than on its institutional design: it converted a political aspiration into a standing bureaucracy with a legal mandate, and that bureaucracy has outlived the Cold War conditions that produced it. Keywords: Treaty of Tlatelolco; nuclear-weapon-free zone; OPANAL; nuclear non-proliferation; Latin America and the Caribbean; regional security; disarmament law; IAEA safeguards 1. Introduction On 14 February 1967, delegates from fourteen states of Latin America signed a document in the Tlatelolco district of #Mexico_City that committed their region to remain permanently free of nuclear weapons. The agreement was not the first attempt to keep nuclear arms out of a defined territory. The Antarctic Treaty of 1959 had already excluded them from a continent, and the Outer Space Treaty of the same year as Tlatelolco extended a similar logic beyond the atmosphere. What made the Latin American instrument different was that it applied to a region where people actually lived, farmed, traded and governed themselves. The #zone_of_application covered millions of square kilometres and, eventually, hundreds of millions of citizens. This was #denuclearisation not of an empty continent or of the vacuum of space, but of a populated political space with rival states, ongoing border disputes, military governments and, in at least two cases, serious ambitions to master the complete nuclear fuel cycle. The literature on nuclear order has often treated the treaty as a footnote to the #NPT, a regional curiosity that happened to precede the global bargain struck in 1968. That reading is difficult to sustain in light of recent historical work. Archival research on Mexican diplomacy shows that the negotiators in Mexico City and the negotiators in Geneva were, in several cases, the same people, and that they deliberately used the regional process to shape the global one (Rodriguez, 2025a). Latin American diplomats did not simply accept a nuclear order designed elsewhere. They built an instrument of their own and then insisted that the global treaty accommodate it. Article VII of the NPT, which preserves the right of groups of states to conclude regional treaties assuring the total absence of nuclear weapons in their territories, exists in part because the Tlatelolco coalition insisted on it. This article makes three arguments. The first is historical: the treaty was not a spontaneous reaction to the missile crisis of October 1962 but the product of a sustained, four-session negotiating process in which the participants worked through genuinely difficult questions about sovereignty, transit, verification and the peaceful uses of the atom. The second is legal and institutional: the durability of the regime rests on a design choice that later zones copied only partially, namely the creation of a standing regional agency with its own General Conference, Council and Secretariat rather than reliance on periodic diplomatic conferences. The third is contemporary: the regime now faces a set of pressures, above all around #naval_nuclear_propulsion and the persistence of interpretive declarations by nuclear-armed states, that test whether a treaty written in the 1960s can govern the technological and geopolitical realities of the 2020s. The article is written for students and early-career researchers approaching #arms_control and #disarmament_law for the first time. It therefore explains technical terms as it goes and keeps the prose plain, while retaining the structure and evidentiary standards of a peer-reviewed article. 1.1 Research questions The study is organised around four questions. What political and diplomatic conditions made a regional nuclear-weapon-free zone possible in Latin America in the 1960s, when comparable proposals failed elsewhere? How does the legal architecture of the treaty, including its two Additional Protocols and its safeguards provisions, distribute obligations between zone members and external powers? How did the regime move from a partial zone, in which several key states withheld full adherence for decades, to a universal one? What are the principal contemporary challenges to the regime, and what do they suggest about the future of the #nuclear_weapon_free_zone as an instrument of #regional_security? 1.2 Structure Section 2 reviews the relevant scholarship and sets out the conceptual framework. Section 3 explains the method and the sources. Section 4 reconstructs the historical background. Section 5 analyses the legal architecture. Section 6 examines the institutional machinery of #OPANAL and the treaty Control System. Section 7 traces the long path to universality. Section 8 assesses the effectiveness of the regime against several benchmarks. Section 9 considers diffusion and the influence of the treaty on later zones and on the #TPNW. Section 10 addresses contemporary challenges. Section 11 states the limitations of the analysis. Section 12 concludes. 2. Literature review and conceptual framework 2.1 What a nuclear-weapon-free zone is A #nuclear_weapon_free_zone is a legally binding arrangement, recognised as such by the United Nations General Assembly, under which a group of states agrees to exclude nuclear weapons entirely from a defined geographical area, establishes a system to verify that commitment, and seeks corresponding assurances from the states that possess nuclear weapons. The definition matters because it separates zones from weaker constructs. A political declaration that a region is a #zone_of_peace has moral weight but no verification. A unilateral renunciation binds one state only. A zone, properly constituted, combines three elements: a prohibition, a verification mechanism, and a set of external guarantees usually delivered through protocols. The prohibition normally covers testing, use, manufacture, production, acquisition, receipt, storage, installation, deployment and possession of nuclear weapons, whether directly or indirectly, and whether by the state itself or by anyone acting on its behalf. The verification mechanism normally rests on comprehensive #safeguards_agreements with the International Atomic Energy Agency, sometimes supplemented by regional bodies. The external guarantees, usually called #negative_security_assurances, are undertakings by nuclear-armed states not to use or threaten to use nuclear weapons against members of the zone. 2.2 Existing scholarship Three broad strands of scholarship bear on the treaty. The first is diplomatic and international history. Recent archival work has substantially revised the older picture of Latin American passivity. Hunt (2022) situates the region within a wider account of how the United States and its partners policed the atom in the two decades after Hiroshima, showing that the emergence of a nuclear club was contested from below as well as managed from above. Rodriguez (2025a) demonstrates that Mexican representatives consciously linked the regional and global lawmaking processes, using the Tlatelolco precedent and the coalition behind it to amplify their voice in the NPT debates and to ensure that the global treaty did not interfere with consolidation of the zone. Gibbons (2022) analyses how the hegemon assembled the tools, incentives and pressures that persuaded most states to join the non-proliferation regime, which by contrast highlights how unusual it was for a group of #Global_South states to arrive first with an instrument of their own. The second strand is critical and postcolonial. Gabriel and Sanchez Cobaleda (2023) read the treaty as a form of contestation: Latin American governments accepted the substance of non-proliferation but demanded concessions from the great powers in return, and in doing so refined instruments that had been drafted with the priorities of nuclear-armed states in mind. Egeland (2021) offers a broader critique of what he calls the ideology of nuclear order, arguing that the discourse of pragmatic nuclear management tends to entrench existing hierarchies while formally endorsing the goal of abolition. Egel and Ward (2022) analyse the #TPNW as an act of subversion by subordinate actors within a hierarchical order, a reading that places Tlatelolco in a longer lineage of institutional challenges from states without nuclear weapons. The third strand is policy analysis of the zones themselves. Hamel-Green (2021) assesses the scope conditions for a zone in Northeast Asia and draws explicitly on the Latin American precedent. Kassenova (2022) provides a detailed account of how Kazakhstan renounced the arsenal it inherited and helped to create the Central Asian zone. Onderco (2021) explains how the coalition that made the NPT permanent in 1995 was assembled, an episode in which the Latin American caucus mattered. Rodriguez (2024, 2025b) argues that Latin American practice offers transferable lessons in #verification and irreversibility, while warning that the regional institutions should not be taken for granted. Musto (2026) identifies a new and under-examined pressure on all zones, namely the spread of nuclear-powered submarines to states that do not possess nuclear weapons. 2.3 Conceptual framework The article draws on three concepts. Regime theory. A regime is a set of principles, norms, rules and decision-making procedures around which actor expectations converge in a given issue area. The Tlatelolco regime consists of the treaty text, the two protocols, the associated safeguards agreements, the resolutions of the OPANAL General Conference, and the settled practice of the member states. Treating it as a regime rather than a single document explains why the meaning of the treaty has changed over time without any amendment to its central prohibitions. Norm entrepreneurship. Individuals and small states can promote new normative standards and persuade others to adopt them. #Alfonso_Garcia_Robles, the Mexican diplomat who chaired the preparatory commission and later received the #Nobel_Peace_Prize in 1982, is a textbook case. So is the wider Mexican delegation, which combined legal craftsmanship with a willingness to accept imperfect compromises in order to secure a functioning instrument. Institutionalisation. The decisive variable in the durability of the regime is not the strength of its prohibitions on paper but the existence of a permanent organisation with a budget, a secretariat, a legal personality and an annual reporting cycle. Institutions create constituencies, generate records, socialise officials and outlast the governments that founded them. 3. Method and sources This is a qualitative, document-based study. It combines doctrinal analysis of the treaty text and its protocols with process tracing of the negotiation and of subsequent adherence decisions, and with an interpretive reading of the secondary literature. Three categories of source are used. The first is primary legal material: the treaty and its Additional Protocols, the amendments adopted between 1990 and 1992, and the safeguards agreements concluded under Article 13. The second is the institutional record of #OPANAL, in particular the resolutions of its General Conference and the reports of its Secretariat. The third is peer-reviewed scholarship and policy analysis published since 2021, which forms the evidentiary backbone of the interpretation offered here. The study does not use interviews or unpublished archives. Where it makes historical claims about the negotiating process, it relies on the published work of historians who have consulted those archives, above all Rodriguez (2025a) and Hunt (2022). This is a limitation, and it is stated openly in Section 11. The analytical strategy is comparative in a limited sense. The Latin American zone is compared with the four other regional zones in force, and with proposals that have failed, in order to isolate the features that account for its success. The comparison is illustrative rather than statistically controlled, which is unavoidable given that only five regional zones exist. 4. Historical background 4.1 Before the crisis Interest in keeping nuclear weapons out of the region predates the missile crisis. Costa Rica raised the idea of a regional arrangement against nuclear arms at a meeting of the Organization of American States council in 1958. The proposal went nowhere, but it established that the idea was in circulation before the events that are usually credited with producing it. Several Latin American states were also active in the wider debates of the late 1950s about testing and fallout, and the region had begun to build its own scientific capacity in radiobiology, dosimetry and radiation protection, often in cooperation with the International Atomic Energy Agency (Barahona, 2022). The point deserves emphasis because it complicates the assumption that the region approached nuclear questions as a passive recipient of external expertise. Latin American scientists and diplomats had their own institutions, their own priorities and their own reasons to care. 4.2 The Cuban Missile Crisis as catalyst In October 1962, the discovery of Soviet ballistic missiles in Cuba brought the superpowers to the edge of war. The crisis was resolved by negotiation between Washington and Moscow. Latin American governments, including the government on whose territory the missiles stood, were largely spectators to a confrontation that could have destroyed them. That experience of exclusion, more than the fear of the weapons themselves, supplied the political energy for the zone. The immediate diplomatic response came in April 1963, when the presidents of #Brazil, #Chile, Bolivia, Ecuador and #Mexico issued a joint declaration announcing their intention to negotiate a multilateral agreement that would keep nuclear weapons out of Latin America. The declaration was brief and did not commit anyone to specific obligations, but it converted a national idea into a regional project and provided the basis for taking the matter to the United Nations. In November 1963, the #UN_General_Assembly welcomed the initiative. The following year, a preliminary meeting in Mexico City created the Preparatory Commission for the Denuclearization of Latin America, universally known by its Spanish acronym #COPREDAL. 4.3 COPREDAL and the negotiation COPREDAL met in four sets of sessions between 1964 and 1967, all of them in Mexico City. At its largest the process involved twenty-one Latin American states, together with observers from outside the region, among them Austria, Canada, Denmark, France, India, Japan, Sweden, the United Kingdom and the United States, and from the International Atomic Energy Agency. The negotiation was not smooth. Four questions proved especially difficult. The definition of the zone. How far out to sea should the prohibition extend? Some delegations wanted a broad maritime zone that would constrain the movements of the superpower navies. Others recognised that an expansive definition would guarantee that no nuclear-armed state would ever accept the protocols. The compromise, embodied in Article 4, defined the zone in two stages: an initial area comprising the territories of the parties, and a much larger area, described by precise coordinates in the western hemisphere, that would apply only once the conditions of Article 28 were fulfilled. Transit. May a warship or aircraft carrying nuclear weapons pass through the territorial sea or airspace of a zone member? The treaty does not answer the question explicitly. The prevailing interpretation, encouraged by the negotiating history, is that each state party decides for itself whether to grant transit rights, consistent with its rights under international law. This deliberate silence, sometimes called #transit_ambiguity, was the price of agreement. It has been criticised ever since as a loophole, and it was one of the features the Indian observer had in mind when he described the treaty as containing gaps while still calling it a real achievement. Peaceful nuclear explosions. In the 1960s, the idea that nuclear devices could be used for civil engineering, harbour excavation or canal digging was taken seriously by serious people. Brazil in particular was unwilling to renounce a technology it believed might have economic value. Article 18 therefore permits parties to carry out explosions of nuclear devices for peaceful purposes, provided this is consistent with the prohibition on nuclear weapons and subject to notification and observation. Since the physics of a peaceful device and a weapon are identical, the provision was always uncomfortable. It has never been used, and the entry into force of the Comprehensive Nuclear-Test-Ban Treaty regime has made it a dead letter in practice, but it remains in the text as a monument to a debate that once mattered. Entry into force. Several delegations feared that a treaty requiring universal ratification before it bound anyone would remain a dead letter if one state held out. Others feared that a treaty binding a subset of states while others stayed outside would be dangerous. Article 28 solved the problem with an elegant device discussed in Section 5.4: it set demanding conditions for general entry into force, but allowed any state to waive those conditions and bring the treaty into force for itself immediately. The treaty was opened for signature on 14 February 1967. On 5 December of that year the United Nations General Assembly endorsed it, with the United States voting in favour and a substantial number of abstentions reflecting unease among states that had not been consulted. 5. The legal architecture 5.1 The core obligation Article 1 contains the central undertaking. Parties agree to use the nuclear material and facilities under their jurisdiction exclusively for peaceful purposes, and to prohibit and prevent, in their respective territories, the testing, use, manufacture, production or acquisition by any means of any nuclear weapon, and the receipt, storage, installation, deployment and any form of possession of any nuclear weapon, directly or indirectly, by the parties themselves, by anyone on their behalf, or in any other way. They also undertake not to engage in, encourage or authorise, directly or indirectly, or in any way participate in the testing, use, manufacture, production, possession or control of any nuclear weapon. Two features of this drafting deserve attention. First, it prohibits both national programmes and the stationing of foreign weapons. A state party cannot build a bomb and cannot host someone else's. Second, it prohibits participation and encouragement, not merely possession. The obligation is thus broader than a simple ban on ownership. Article 5 defines a nuclear weapon as any device capable of releasing nuclear energy in an uncontrolled manner and which has a group of characteristics appropriate for use for warlike purposes. It adds that an instrument that may be used for transport or propulsion of the device is not included in the definition if it is separable and not an indivisible part of it. This definitional carve-out, drafted with delivery vehicles in mind, has acquired unexpected relevance in the debate over #naval_nuclear_propulsion discussed in Section 10. 5.2 Safeguards and the peaceful use of the atom Article 13 requires each party to negotiate multilateral or bilateral agreements with the International Atomic Energy Agency for the application of #IAEA_safeguards to its nuclear activities. This is the verification backbone of the regime. It is also the provision that ties the regional instrument into the global system: the safeguards applied under Tlatelolco are the same safeguards applied under the NPT, administered by the same inspectorate, using the same technical standards. Article 17 preserves the right of parties to use nuclear energy for peaceful purposes, in particular for economic development and social progress. This provision was not decorative. For Brazil, Argentina and Mexico, all of which had ambitions in #peaceful_nuclear_energy, the treaty had to be a guarantee of access as well as a renunciation of weapons. The bargain was the same one later struck globally in the NPT: give up the weapon, keep the technology, and accept inspection as the price of both. 5.3 The two Additional Protocols The treaty comes with two protocols that bind states outside the region. They are the mechanism by which a regional agreement acquires extra-regional force. #Additional_Protocol_I applies to states outside the region that are internationally responsible, in law or in fact, for territories inside the zone. Four states fell into this category: France, the Netherlands, the United Kingdom and the United States. Each undertook to apply the denuclearisation provisions of the treaty to those territories. For the United States this meant Puerto Rico, the United States Virgin Islands and the naval base at Guantanamo Bay. All four eventually signed, and all four ultimately became bound. #Additional_Protocol_II applies to the five states recognised as possessing nuclear weapons under the NPT: China, France, Russia (as successor to the Soviet Union), the United Kingdom and the United States. Each undertakes to respect the statute of denuclearisation of the zone and not to use or threaten to use nuclear weapons against the parties. This is the source of the #negative_security_assurances that give the zone its external dimension. All five have ratified. The significance of full protocol adherence should not be understated. In no other zone have all five recognised nuclear-weapon states ratified the relevant protocols without qualification of substance. The qualification, and it is a serious one, lies in the #interpretive_declarations attached at the time of ratification, which are discussed in Section 8.3. 5.4 Entry into force and the waiver in Article 28 Article 28 sets out the conditions for general entry into force: all states in the region must ratify, all extra-regional states with territories in the zone must ratify Protocol I, all nuclear-weapon states must ratify Protocol II, and the necessary safeguards agreements must be concluded. These conditions were, in 1967, plainly unattainable in the short term. The second paragraph of Article 28 therefore allows any signatory to waive, wholly or in part, the requirements of the first paragraph, and thereby bring the treaty into force for itself at once. Mexico waived immediately. Others followed. The result was that the treaty entered into force for individual states from 1968 onward, and the zone grew state by state, rather than waiting for a moment of collective perfection that might never arrive. This is one of the most instructive design features of the whole instrument. It converted a collective action problem into an individual decision. A state did not have to wait for its rivals. It could simply act. The choice of the waiver had a signalling effect too: states that waived demonstrated confidence, and states that did not, notably Argentina until 1994, revealed their hesitation to the region and to the world. 5.5 The amendments of 1990 to 1992 Three amendments modernised the text. In 1990 the General Conference amended the title to add the words "and the Caribbean", acknowledging that the zone had ceased to be a purely Latin American project and now included the Anglophone, Francophone and Dutch-speaking Caribbean. In 1991 Article 25 was amended. The original paragraph had excluded political entities whose territory was, before the treaty was opened for signature, the subject of a dispute or claim between an extra-continental state and one or more Latin American states. In practice this excluded Belize and Guyana. The amendment removed the barrier and opened the zone to both. In 1992 the #Control_System of the treaty was amended, principally by modifying the provisions on special inspections. The changes were driven by Argentina and Brazil, whose accession depended on assurances that the special inspection procedure could not be used abusively by a rival. The amendments are a good illustration of a general principle: verification arrangements are not merely technical, they are political settlements about how much intrusion states will tolerate from one another. 6. Institutional design: OPANAL and the Control System 6.1 The Agency Article 7 establishes the Agency for the Prohibition of Nuclear Weapons in Latin America and the Caribbean, #OPANAL, with its seat in Mexico City. This is the feature that most sharply distinguishes Tlatelolco from the zones that followed. The South Pacific, Southeast Asian, African and Central Asian treaties either rely on existing regional organisations or on consultative committees that meet intermittently. Latin America built a dedicated international organisation. OPANAL has three organs. The #General_Conference is the supreme organ. All member states are represented, each with one vote. It meets in ordinary session every two years and can be convened in special session. It adopts the budget, elects the Secretary-General, admits new members, and takes decisions on compliance. Its resolutions constitute an important source of subsequent practice for interpreting the treaty. The Council consists of five member states elected by the General Conference, serving four-year terms with staggered renewal. It functions continuously and is responsible for ensuring the proper operation of the Control System. The Secretariat, headed by a Secretary-General, carries out the day-to-day work: preparing reports, maintaining relations with the IAEA and with other zone organisations, representing the treaty in United Nations disarmament forums, and running the disarmament education programme that the Agency has developed in recent decades. 6.2 The Control System Articles 12 to 16 establish the #Control_System, whose declared purpose is to verify that devices, services and facilities intended for peaceful uses of nuclear energy are not used for the testing or manufacture of nuclear weapons, that none of the activities prohibited by Article 1 are carried out in the zone, and that no nuclear weapons are received, stored, installed, deployed or possessed there. The system has three pillars. The first is the network of #safeguards_agreements with the IAEA required by Article 13. Every member state of the zone has concluded a comprehensive safeguards agreement, a fact that the Agency and its members cite regularly as evidence of the maturity of the regime. The second is a reporting obligation. Parties submit semi-annual reports to the Agency and to the IAEA stating that no activity prohibited by the treaty has occurred in their territories. The third is the special inspection procedure. Any party that suspects another of conducting a prohibited activity may request a special inspection. The procedure was amended in 1992, as noted above, to address concerns about politicisation. 6.3 The relationship with the IAEA and with ABACC The regime is best understood as a layered system. The IAEA supplies the technical verification. OPANAL supplies the political and legal supervision. In the specific case of #Argentina and #Brazil, a third layer exists. In 1991 the two states created the Brazilian-Argentine Agency for Accounting and Control of Nuclear Materials, #ABACC, a bilateral body that inspects nuclear facilities in both countries. In the same year they concluded, with ABACC and the IAEA, the #Quadripartite_Agreement, under which all nuclear material in both states is subject to a double layer of verification: comprehensive IAEA safeguards and, in parallel, safeguards applied by ABACC. ABACC can conduct inspections without prior notification. The arrangement solved a specific regional problem. Neither Argentina nor Brazil was willing to submit to inspection unless the other did so on identical terms, and neither trusted a purely international inspectorate to be sensitive to their industrial and commercial interests. The bilateral agency allowed each to watch the other while both were watched by the IAEA. It was the confidence generated by this mutual scrutiny that made ratification of Tlatelolco possible for both states in 1994 (Rodriguez, 2024). The lesson has been drawn explicitly by analysts working on other regions: a dual approach, combining a global inspectorate with a regionally owned verification body, may be more robust than either alone, because it satisfies both the international demand for assurance and the regional demand for ownership (Rodriguez, 2025b). 7. From partial zone to universal zone 7.1 The holdouts For most of its history the treaty was a partial instrument. The states that mattered most for #nonproliferation in the region were precisely the ones that had not fully adhered. Argentina signed in 1967 but did not ratify until 1994. Its nuclear programme was advanced, its military governments were suspicious of external constraint, and its rivalry with Brazil made unilateral restraint appear costly. Argentina had mastered uranium enrichment and possessed the technical basis for a weapons programme had it chosen to pursue one. Brazil signed in 1967 and ratified in 1968, but invoked the Article 28 requirements rather than waiving them, so the treaty did not enter into force for Brazil. It embarked on a secret military nuclear programme in the 1970s. Full adherence came only in 1994, after democratisation and after the bilateral settlement with Argentina. Chile signed in 1967 and ratified in 1974, but likewise did not waive the entry-into-force conditions, so the treaty did not bind it until 1994. #Cuba refused to participate in the negotiations at all. Its conditions were political: it demanded the inclusion of United States military installations in the region, of Puerto Rico and of the Panama Canal within the zone, and the return of the base at Guantanamo Bay. None of these demands was met. Cuba signed in 1995 and ratified only in 2002. 7.2 What changed Four developments explain the transformation between 1990 and 2002. The first is democratisation. Argentina, Brazil and Chile all returned to civilian rule in the 1980s. Civilian governments faced electorates that had little enthusiasm for expensive weapons programmes and considerable enthusiasm for reintegration into the world economy. The second is the resolution of the Argentine-Brazilian rivalry. The two states moved from suspicion to cooperation, culminating in the 1991 bilateral agreement on the exclusively peaceful use of nuclear energy, the creation of #ABACC and the #Quadripartite_Agreement. The third is the end of the Cold War, which removed the strategic rationale for hedging and reduced the perceived cost of accepting external inspection. The fourth is institutional patience. The Agency continued to exist, to convene, to report and to lobby throughout the decades in which the zone was incomplete. When the political conditions changed, the machinery was already there. This is the strongest single argument for the #institutionalisation thesis advanced in this article. With the Cuban ratification deposited on 23 October 2002, the treaty came into force throughout the region. All thirty-three states of #Latin_America and the #Caribbean are now parties. #treaty_universality, in the regional sense, has been achieved. 7.3 The zone of peace declaration In January 2014, at the second summit of the Community of Latin American and Caribbean States, held in Havana, the region declared itself a #zone_of_peace, expressly invoking the Treaty of Tlatelolco and international law. The declaration is political rather than legally binding, and it should not be confused with the treaty. Its significance is discursive: it shows that the zone has become part of how the region describes itself, and it links #CELAC, a body with no nuclear mandate, to the legal architecture built half a century earlier. 8. Assessing effectiveness 8.1 Benchmarks Assessing a treaty requires an explicit standard. Four are used here. Non-acquisition. Have any parties acquired nuclear weapons? Non-deployment. Have nuclear weapons been stationed in the zone? External recognition. Have the nuclear-armed states accepted binding obligations towards the zone? Diffusion. Has the model been reproduced elsewhere? 8.2 Performance against the benchmarks On the first benchmark the record is unambiguous. No party to the treaty possesses nuclear weapons. Two states, Argentina and Brazil, had the technical capacity and, for a period, the intention to move towards a weapons option, and both abandoned that path. Attributing this outcome solely to the treaty would be an error, since democratisation, economic crisis, bilateral reconciliation and the wider #nonproliferation regime all contributed. But the treaty supplied the legal frame within which reconciliation could be locked in and made difficult to reverse. This is the point emphasised in the recent literature on irreversibility: the value of the regional instruments is that they raise the political cost of reversing a decision that a future government might otherwise reconsider (Rodriguez, 2025b). On the second benchmark the record is good but not perfect, because of #transit_ambiguity. No nuclear weapons are stationed in the zone. Whether nuclear-armed vessels transit the territorial waters of member states is, by design, not something the treaty resolves. On the third benchmark the record is unusually strong. All five recognised nuclear-weapon states have ratified Protocol II, and all four relevant extra-regional states have accepted Protocol I. No other zone has achieved this. On the fourth benchmark, discussed in Section 9, the record is strong. Every subsequent regional zone treaty borrows from Tlatelolco. 8.3 The problem of interpretive declarations The principal qualification to the third benchmark concerns #interpretive_declarations. When ratifying Protocol II, the nuclear-weapon states attached statements setting out how they understood their obligations. Some of these statements reserve positions on transit. Others reserve the right to reconsider the negative security assurance in the event of an armed attack carried out with the support of a nuclear-armed state. The member states of the zone have objected consistently. Successive sessions of the #General_Conference have called on the parties to the Protocols to review or withdraw their interpretive declarations, and the point is raised at every #NPT review cycle. The nuclear-weapon states have not complied. The dispute matters for two reasons. Legally, it raises the question of how far a unilateral statement can qualify a treaty obligation without amounting to a reservation, which the Protocols do not permit. Politically, it means that the #negative_security_assurances at the heart of the bargain are less than absolute. Critics argue that this asymmetry illustrates a wider pattern in which the nuclear-armed states extract compliance from others while retaining flexibility for themselves (Egeland, 2021; Gabriel and Sanchez Cobaleda, 2023). 8.4 Was the treaty necessary? A sceptical reading holds that Latin America was never going to acquire nuclear weapons in any case, that the region faced no existential threats requiring them, and that the treaty therefore prohibited something that would not have happened. This argument is weaker than it appears. Argentina and Brazil both invested heavily in sensitive fuel cycle technology under military governments, both concealed elements of their programmes, and both were assessed by outside observers in the 1970s and 1980s as plausible future possessors. The counterfactual in which the region simply drifted into a bilateral nuclear rivalry is not fanciful. What the treaty and its associated institutions did was to provide a ready-made framework into which the reconciliation of the early 1990s could be poured. Institutions do not create political will, but they determine what political will can be converted into once it appears. 9. Diffusion: the treaty as a template 9.1 The four later zones Four further regional zones have entered into force, and each bears the imprint of the Latin American original. The South Pacific Nuclear Free Zone Treaty, signed at #Rarotonga in 1985, adopts the same basic structure of a core prohibition plus protocols for external powers, and adds provisions on the dumping of radioactive waste at sea, reflecting regional concerns about testing and disposal in the Pacific. The Southeast Asian zone, established by the #Bangkok_Treaty of 1995, extends the prohibition to exclusive economic zones and continental shelves, a more ambitious geographical scope that has contributed to the reluctance of the nuclear-weapon states to sign its protocol. The African zone, created by the #Pelindaba treaty of 1996 and in force since 2009, follows a South African decision to dismantle the only nuclear arsenal ever built and voluntarily surrendered, and it establishes an African Commission on Nuclear Energy with functions loosely analogous to those of OPANAL. The Central Asian zone, created by the #Semipalatinsk treaty of 2006, is the first zone in the northern hemisphere and the first to require its members to conclude an Additional Protocol with the IAEA, thereby setting a higher verification standard than Tlatelolco itself (Kassenova, 2022). To these should be added the status of #Mongolia as a single-state nuclear-weapon-free territory, recognised by the United Nations, an arrangement that stretches the concept but demonstrates its flexibility. 9.2 The relationship with the NPT The relationship between the regional and global instruments is often presented as hierarchical, with the NPT as the parent and the zones as its offspring. The chronology alone should give pause. Tlatelolco was signed in February 1967 and the NPT was opened for signature in July 1968. The regional instrument came first. Recent scholarship goes further and argues that the causal arrow ran in both directions. Mexican negotiators used the existence of the Tlatelolco coalition to strengthen their bargaining position in Geneva, and worked to ensure that the emerging global treaty would not undercut the regional one (Rodriguez, 2025a). Article VII of the NPT, which explicitly preserves the right of groups of states to conclude regional treaties, is the trace of that effort in the text of the global bargain. Hunt (2022) situates this within a broader account in which the global nuclear order was assembled through negotiation with, and not merely imposition upon, states outside the nuclear club. 9.3 The relationship with the TPNW The Treaty on the Prohibition of Nuclear Weapons, adopted in 2017 and in force since 2021, prohibits nuclear weapons outright for its parties (Kmentt, 2021). Latin American and Caribbean states were among its strongest supporters, and Mexico described the #TPNW during the negotiations as a global extension of the treaties establishing zones free of nuclear weapons. The conceptual link is real. Both instruments rest on the proposition that possession of nuclear weapons is not a right that some states legitimately enjoy but a condition to be eliminated. Both create obligations for states without nuclear weapons while seeking to change the behaviour of those that have them. Both have been read as acts of institutional subversion by states outside the nuclear club, working within the forms of international law to challenge a hierarchy they did not design (Egel and Ward, 2022). There are also differences. A zone is territorial: it removes weapons from a place. The TPNW is categorical: it removes the legitimacy of the weapons as such. A zone can coexist comfortably with #nuclear_deterrence practised elsewhere; the prohibition treaty cannot. This is why some states that are enthusiastic members of the Latin American zone have been cautious about the newer treaty, and why the relationship between the two instruments will remain a live question in the coming review cycles. 10. Contemporary challenges 10.1 Naval nuclear propulsion The most technically demanding challenge now facing the regime concerns nuclear reactors used to propel submarines. Brazil is developing a nuclear-powered, conventionally armed submarine, together with a land-based prototype reactor. The programme is indigenous. Brazil maintains that #naval_nuclear_propulsion is a peaceful use of nuclear energy and therefore permitted, a position it grounds in the text of its safeguards arrangements and in the 1991 bilateral agreement with Argentina, which characterises propulsion as a peaceful application. Brazil has been in consultation with the IAEA since December 2021 about the special procedures that would apply, under the relevant article of the #Quadripartite_Agreement, to nuclear material used in propulsion. Those consultations have involved ABACC and have addressed design information for the prototype reactor, the fuel to be used, and the verification techniques that could be applied. The difficulty is structural rather than a matter of bad faith. Comprehensive safeguards agreements, both the standard model used under the NPT and the quadripartite arrangement, contain provisions allowing a state to withdraw nuclear material from safeguards for use in a non-proscribed military activity such as propulsion. When material leaves safeguards, the inspectorate loses continuity of knowledge about it. If that mechanism became routine, it could create a lawful route out of verification. The issue is not confined to Latin America. The #AUKUS arrangement, under which Australia is to acquire nuclear-powered submarines with support from the United Kingdom and the United States, has generated an intense and unresolved debate within the IAEA Board of Governors about whether and how safeguards can be maintained. Australia is a member of the South Pacific zone. Any precedent established in one zone will be cited in the others (Das, 2025). Musto (2026) frames the problem in terms of the difference between the letter and the spirit of the zone treaties. Nothing in the text of Tlatelolco prohibits a reactor that moves a ship. Article 5, as noted above, excludes separable transport and propulsion instruments from the definition of a nuclear weapon. But the zones were created to insulate their regions from the dynamics of nuclear competition, and the arrival of nuclear-powered fleets brings those dynamics closer. The risk identified in the recent literature is fragmentation: zones that respond to the question in different ways will weaken the coherence of the model as a whole. The counter-argument, advanced by Brazil and by analysts sympathetic to the region, is that the Latin American case is the best available demonstration that the problem is manageable. The Brazilian programme is indigenous and therefore does not involve the transfer of weapons-usable material from a nuclear-armed state to one without nuclear weapons. It is subject to a double layer of verification through the IAEA and #ABACC. It uses low-enriched fuel rather than the highly enriched fuel used in some other navies. And it has been the subject of consultation with the inspectorate from the outset rather than presented as a fait accompli. On this reading, the flexible but institutionalised #Control_System of Tlatelolco is not the weak point of the regime but the model for how other regions might respond. 10.2 The unresolved interpretive declarations As set out in Section 8.3, the #interpretive_declarations attached by the nuclear-weapon states to Protocol II remain in place despite repeated demands from the #General_Conference for their review or withdrawal. This is a slow-burning grievance rather than an acute crisis, but it shapes the political temperature of the regime and provides continuing evidence for critics who argue that the assurances offered to states without nuclear weapons are conditional and revocable. 10.3 Great power competition The regional environment has changed. Nuclear-armed states are modernising their arsenals. Arms control agreements between the major powers have lapsed or are under strain. Competition between the United States, China and Russia now reaches into Latin American nuclear energy, space technology and critical mineral supply chains, and cooperation between Chinese entities and Latin American governments in these areas has begun to be framed in Washington as a security question rather than a commercial one (Rodriguez, 2025b; Alexander Thew, 2023). The danger for the regime is not that a member state will seek a bomb. It is that the region will be drawn into the strategic competition of others, that its nuclear energy choices will be securitised, and that the political space for the kind of autonomous, regionally owned diplomacy that produced the treaty in the first place will narrow. 10.4 The renewed international attention to zones In December 2024 the United Nations General Assembly adopted a resolution requesting the Secretary-General to convene a qualified group of experts to prepare a comprehensive study of the question of nuclear-weapon-free zones in all their aspects. This is the first study of its kind since 1975. The group began meeting in December 2025 and has continued through 2026. The timing is significant. It reflects an emerging recognition that the zones, long treated as settled achievements, now face questions their drafters did not anticipate: propulsion reactors, small modular reactors, new enrichment technologies, the militarisation of space and the erosion of the arms control architecture between the major powers. It also creates an opportunity. If the zones are to remain relevant, they will need a common position on these questions, and OPANAL, as the oldest and most institutionalised of the zone bodies, is well placed to help shape it. 10.5 Nuclear energy expansion in the region A final, quieter challenge concerns the growth of civil nuclear power. Interest in small modular reactors is growing across the region, and several states are examining nuclear options as part of decarbonisation strategies. Expanded civil programmes are entirely lawful under Article 17 and were always part of the bargain. They nevertheless increase the amount of material and the number of facilities that must be safeguarded, and they place additional demands on an inspectorate whose resources have not grown in proportion. #verification capacity, not political will, may prove to be the binding constraint of the next two decades. 11. Limitations of the study Three limitations should be acknowledged. First, the study relies on published sources and does not draw on original archival research or interviews. Historical claims about the negotiating process are therefore secondhand, mediated through the work of scholars who have consulted the Mexican, Brazilian, Argentine and United States archives. Second, the comparison with other zones is illustrative rather than systematic. With only five regional zones in force, generalisation about the conditions for success rests on a very small number of cases, and the danger of selecting on the dependent variable is real. Failed proposals, in the Middle East, in South Asia and in Northeast Asia, would need to be analysed in comparable depth before firm causal claims could be made. Third, the assessment of contemporary challenges, in particular the naval propulsion question, concerns a situation that is still developing. The consultations between Brazil, ABACC and the IAEA are ongoing, and the conclusions offered here are provisional. 12. Implications and conclusion 12.1 Lessons for other regions Proposals for zones in the Middle East and in Northeast Asia have been on the international agenda for decades without result. The Latin American experience suggests several conditions that appear to matter. A regional catalyst helps but does not suffice. The #Cuban_Missile_Crisis created urgency, but Costa Rica had raised the idea four years earlier, and the negotiation still took more than four years to complete. Rivalry can be managed rather than resolved first. Argentina and Brazil did not settle their competition before joining the zone; the zone and the associated bilateral machinery were part of how they settled it. This is directly relevant to regions where the standard objection is that adversaries will never accept restraint while distrust persists (Hamel-Green, 2021). Partial entry into force is better than none. The waiver device in Article 28 allowed the zone to grow incrementally rather than requiring simultaneous universal agreement. Ownership matters. The verification architecture that worked in Latin America was regionally designed and regionally staffed as well as internationally supervised. Externally imposed inspection regimes generate resistance; regionally owned ones generate constituencies. An institution is worth more than a text. The single most consequential decision of the 1967 negotiators may have been Article 7, which created OPANAL. A treaty with no organisation would have had no advocate during the decades when the zone was incomplete. 12.2 Conclusion The #Treaty_of_Tlatelolco was signed by states that were not powerful, in a region that was not central to the strategic calculations of the superpowers, at a moment when the global rules on nuclear weapons had not yet been written. It nevertheless produced the first #nuclear_weapon_free_zone in an inhabited region, secured binding commitments from all five recognised nuclear-weapon states, and supplied the template for every regional zone that followed. Its architect received the #Nobel_Peace_Prize, and its central prohibition has never been violated. Its achievement was not, however, a triumph of moral clarity over power. It was a negotiated compromise, and the compromises are visible in the text: the ambiguity over transit, the provision for #peaceful_nuclear_explosions, the long delay before the most nuclear-capable states in the region accepted the full discipline of the regime. The treaty worked not because it was perfect but because it was institutionalised, because it created an agency that survived the years of partial adherence, and because it left a door open through which states could pass when their politics changed. That is the lesson worth carrying forward. The regime is now confronting technologies and rivalries its drafters could not have foreseen. Whether it adapts will depend on the same quality that carried it through the 1970s and 1980s: the existence of a permanent regional institution, with a mandate, a memory and a seat at the table, capable of speaking for the region when the region is otherwise spoken about rather than heard. Hashtags #Treaty_of_Tlatelolco #nuclear_weapon_free_zone #OPANAL #nuclear_disarmament #nonproliferation_regime #Latin_America_and_the_Caribbean #IAEA_safeguards #Alfonso_Garcia_Robles #Additional_Protocol_II #ABACC_verification #TPNW_and_NWFZ #naval_nuclear_propulsion_debate #disarmament_law #regional_security_governance #arms_control_history References Alexander Thew, J. (2023). Negating non-proliferation? The potential Sino-South challenge to nuclear market regulation. The International Spectator, 58(4), 36-53. https://doi.org/10.1080/03932729.2023.2256103 Barahona, A. (2022). Radiation risk in Cold War Mexico: Local and global networks. NTM Zeitschrift fur Geschichte der Wissenschaften, Technik und Medizin, 30. https://doi.org/10.1007/s00048-022-00331-0 Das, D. (2025). Cascades of competition: Southern Asia, the Indo-Pacific, and AUKUS. Texas National Security Review, Roundtable: Navigating the New Nuclear Map. Egel, N., & Ward, S. (2022). Hierarchy, revisionism, and subordinate actors: The TPNW and the subversion of the nuclear order. European Journal of International Relations, 28(4), 751-776. https://doi.org/10.1177/13540661221108384 Egeland, K. (2021). The ideology of nuclear order. New Political Science, 43(2), 208-230. https://doi.org/10.1080/07393148.2021.1886772 Gabriel, J. P. N., & Sanchez Cobaleda, A. (2023). Tlatelolco Treaty: The Global South's postcolonial contestation of the nuclear non-proliferation regime. Peace Review: A Journal of Social Justice, 35(4), 634-646. https://doi.org/10.1080/10402659.2023.2249415 Gibbons, R. D. (2022). The hegemon's tool kit: US leadership and the politics of the nuclear nonproliferation regime. Cornell University Press. Hamel-Green, M. (2021). Nuclear deadlock, stalled diplomacy: The Northeast Asia nuclear weapon free zone alternative. Proposals, pathways, prospects. Journal for Peace and Nuclear Disarmament, 4(sup1), 201-224. https://doi.org/10.1080/25751654.2021.1917028 Hunt, J. R. (2022). The nuclear club: How America and the world policed the atom from Hiroshima to Vietnam. Stanford University Press. Kassenova, T. (2022). Atomic steppe: How Kazakhstan gave up the bomb. Stanford University Press. Kmentt, A. (2021). The treaty prohibiting nuclear weapons: How it was achieved and why it matters. Routledge. Musto, R. A. (2026). A new era of nuclear-powered submarines is making waves in nuclear-weapon-free zones. Carnegie Endowment for International Peace. Onderco, M. (2021). Networked nonproliferation: Making the NPT permanent. Stanford University Press. Rendall, M. (2022). Nuclear war as a predictable surprise. Global Policy, 13(5), 782-791. https://doi.org/10.1111/1758-5899.13142 Rodriguez, J. L. (2024). Arms control lessons from Latin America. The Washington Quarterly, 47(4), 177-193. https://doi.org/10.1080/0163660X.2024.2435723 Rodriguez, J. L. (2025a). Mexico and the balancing of nuclear perils and promises in the 1960s. Cold War History, 1-22. https://doi.org/10.1080/14682745.2024.2404835 Rodriguez, J. L. (2025b). Latin America and contemporary nuclear challenges. Texas National Security Review, Roundtable: Navigating the New Nuclear Map. Sukin, L., & Mukherjee, R. (2025). The evolving global landscape of nuclear security. Texas National Security Review, Roundtable: Navigating the New Nuclear Map.

International conventions are the main legal tools that states use to manage shared resources that no single country owns. This article studies four related families of agreements: the environmental conventions that grew out of the 1972 Stockholm Conference, the climate conventions built around the United Nations Framework Convention on Climate Change, the nature and biodiversity conventions led by the Convention on Biological Diversity, and the space conventions that begin with the 1967 Outer Space Treaty. The study uses a doctrinal and comparative method. It reads the treaty texts, the decisions of their governing bodies, and recent peer reviewed literature published between 2021 and 2026. The article asks three questions. First, what legal design features do these conventions share? Second, why do some regimes change state behaviour while others do not? Third, what does the recent record, including the 2025 advisory opinion of the International Court of Justice, the entry into force of the high seas agreement in January 2026, the outcome of the thirtieth climate conference in Belem, and the failure of the plastics negotiations in Geneva, tell us about the future of treaty based cooperation? The analysis finds that success depends less on the strength of the words in a treaty and more on four practical factors: a narrow and measurable problem definition, a credible financial mechanism, a reliable science to policy channel, and a review system that makes non compliance visible. Regimes that manage a small number of substances, such as the ozone regime, perform far better than regimes that try to steer whole economies. The article also finds that space governance is now repeating an early mistake of environmental law by relying on voluntary guidelines while commercial activity grows faster than the rules. The article closes with recommendations for treaty design, for national implementation, and for student researchers who wish to enter this field. Keywords: international environmental law; climate governance; biodiversity conventions; space law; global commons; treaty compliance; sustainability 1. Introduction The air, the climate system, the deep ocean, the migratory species that cross borders, and the orbits around the Earth belong to everyone and to no one. Because they belong to no one, no single government can protect them alone. A country that cuts its emissions still breathes the pollution of its neighbours. A country that protects a bird in summer cannot protect it in winter if the wintering ground lies far away. A satellite operator that follows every safety rule can still lose a spacecraft to a fragment left behind by someone else. This is the basic problem of #Global_Commons management, and international conventions are the main answer that the world has produced so far. An international convention is a written agreement between states that creates rights and duties under #International_Law. Conventions differ from political declarations because they are meant to bind. Yet in practice the line between binding and non binding is soft. Many conventions contain firm procedural duties, such as the duty to report, and vague substantive duties, such as the duty to protect. This gap between form and effect is the central puzzle of the field, and it is the reason that students often find #Environmental_Governance confusing when they first meet it. The modern system began at the United Nations Conference on the Human Environment, held in Stockholm in 1972. Twenty years later the Rio Earth Summit produced the two great framework conventions on climate and on biological diversity. Thirty years after Rio, the world adopted the Kunming Montreal Global Biodiversity Framework. In 2023 states agreed a new treaty for marine life beyond national borders, and that treaty entered into force on 17 January 2026 after sixty states ratified it. Meanwhile, the space regime, which was written in the 1960s and 1970s for a world of two superpowers and a handful of satellites, now faces tens of thousands of objects in orbit and a private launch industry that the drafters never imagined. This article brings these separate stories together. Most textbooks treat climate law, biodiversity law, ocean law, chemicals law, and space law as separate subjects. Treating them as separate subjects hides the family resemblance between them. They use the same legal building blocks: a framework #Treaty that sets goals, protocols or annexes that set details, a conference of the parties that meets regularly, a secretariat, a scientific body, a financial mechanism, and a reporting system. They also share the same weaknesses: consensus decision making that lets a small minority block progress, weak #Enforcement, and a chronic gap between promises and money. The article has three aims. The first aim is descriptive. It sets out, in plain language, what the main conventions actually say and what their governing bodies have recently decided. The second aim is analytical. It compares the regimes and asks why some work better than others. The third aim is practical. It offers recommendations for treaty design and a short guide for students who want to study or work in this field. The article proceeds as follows. Section 2 reviews the literature and the main concepts. Section 3 explains the method. Section 4 traces the historical development of the system. Sections 5 to 9 examine the climate regime, the nature regime, the ocean and atmosphere regimes, the chemicals and waste regime, and the space regime in turn. Section 10 offers a cross cutting comparison. Section 11 discusses the findings. Section 12 states the limits of the study. Section 13 concludes with recommendations. 2. Literature Review and Conceptual Framework 2.1 The growth of the field International environmental law is young. Rajamani and Peel (2021), reviewing a decade of change in the field, describe a body of law that has grown quickly in volume but unevenly in strength. They note that the field has moved away from strict rules of prohibition towards flexible, procedural, and cooperative techniques. This move made agreement easier to reach, but it also made obligations harder to measure. The Paris Agreement is the clearest example. Each party decides its own contribution, and no international body can reject it as too weak. Scholars generally group the explanations for treaty effectiveness into three schools. The first school looks at power. Treaties work when powerful states want them to work. The second school looks at interests. Treaties work when the cost of compliance is low and the benefit of cooperation is visible. The third school looks at norms and knowledge. Treaties work when a strong scientific community and an active civil society make non compliance embarrassing. The evidence in this article supports a mixed explanation, with the interest based account carrying the most weight in the short term and the norm based account carrying more weight over decades. 2.2 Core principles A small set of principles runs through every regime studied here. The principle of prevention requires a state to stop activities inside its territory from causing serious harm to other states or to areas beyond national control. This principle is the oldest and the firmest. In July 2025 the International Court of Justice confirmed that it applies to the climate system as well, which extends a rule written for smoke and rivers to the atmosphere as a whole (Foster, 2025). The #Precautionary_Principle says that a lack of full scientific certainty should not be used as a reason to postpone action when the threat of serious or irreversible damage exists. It appears in the Rio Declaration, in the climate convention, and in the biodiversity regime. The principle of #Common_But_Differentiated_Responsibilities recognises that all states share the duty to protect the environment but that they do not share it equally. Historic emitters and wealthier states are expected to act first and to help others. This principle is the source of most of the political conflict in the climate talks. The principle of #Sustainable_Development links environmental protection with economic and social needs. It is the bridge that allows developing countries to accept environmental duties without accepting limits on growth. Two further ideas matter for the ocean and for space. The #Common_Heritage_Of_Humankind principle treats certain areas and resources as belonging to all humanity and requires that benefits be shared. It appears in the deep seabed provisions of the law of the sea and in the 1979 Moon Agreement. The freedom of access principle, in contrast, treats the commons as open to all who can reach it. The tension between these two ideas is unresolved, and it explains much of the disagreement about seabed mining and about lunar resources. 2.3 The compliance problem Most conventions rely on what scholars call managerial compliance. Instead of punishing a state that fails, the regime helps it to succeed. It provides money, training, and technology, and it makes performance public. Hard sanctions are rare. The trade restrictions in the ozone regime and in the wildlife trade convention are the main exceptions, and it is not a coincidence that these two regimes are widely regarded as the most effective. Recent work on biodiversity supports this reading. Hughes and Grumbine (2023) argue that the Kunming Montreal framework is weaker than it appears because many of its targets lack clear definitions and measurable indicators, which makes failure hard to prove. McGowan and colleagues (2024) show the same problem for species targets. Lehmann and colleagues (2025) add that the framework leaves the sharing of costs open, so the parties who can least afford action often carry it. 2.4 The literature on space The space law literature has grown quickly since 2020 because the Artemis Accords broke a long silence. Deplano (2021) asks whether the Accords are evolution or revolution, and concludes that they change the method of law making rather than the content of the law, because they replace multilateral negotiation with a set of bilateral political commitments. Tronchetti and Liu (2021) reach a similar conclusion when they examine national measures on space resources. McKeown, Dempster and Saydam (2022) test one specific idea, the safety zone, and ask whether it can work in practice for commercial lunar mining. Schmidt and Svec (2022) argue that the legal debate about space mining is deadlocked and that only a new institutional forum can break it. Prema, Bharathi and Sundar (2024) see the Accords as a new phase in the growth of space law rather than a departure from it. What unites this literature is a worry about fragmentation: the fear that the single legal order created in 1967 is splitting into competing clubs. 2.5 The gap this article addresses The literature treats these regimes separately. Few studies compare the design of a climate treaty with the design of a space treaty, even though both govern a commons, both rely on voluntary national action, and both now depend heavily on private actors. This article fills that gap by placing them side by side. 3. Methodology This is a qualitative, doctrinal, and comparative study. It does not produce new empirical data. It analyses three kinds of material. The first material is primary legal text: the conventions, their protocols, and the decisions of their conferences of the parties. Examples include the Paris Agreement, the Kunming Montreal Global Biodiversity Framework, the Global Mutirao decision adopted at Belem in November 2025, and the Outer Space Treaty. The second material is institutional record: the reports of treaty secretariats, negotiation summaries, and the advisory opinions of international courts, in particular the opinion of the International Court of Justice of 23 July 2025 and the earlier opinion of the International Tribunal for the Law of the Sea of May 2024. The third material is peer reviewed literature published mainly between 2021 and 2026, selected for direct relevance to treaty design, #Compliance, and the specific regimes under study. The comparison uses six criteria, applied to each regime in turn: Clarity of the obligation. Does the treaty tell a state exactly what to do? Measurability. Can performance be counted? Finance. Is there a working financial mechanism, and is it funded? Science to policy channel. Is there a body that turns evidence into decisions? Review and response. What happens when a state does not perform? Coverage of private actors. Does the regime reach the companies that cause the harm? These six criteria are applied consistently, which allows the article to compare a treaty about mercury with a treaty about orbits without forcing them into the same mould. The main limitation of this method is that it studies law on paper and law in institutions rather than outcomes on the ground. Where possible, the article notes the physical trend, such as emissions or debris counts, but it does not attempt to prove causation between a treaty and an outcome. 4. The Historical Development of the Conventions 4.1 Before 1972 Early agreements were narrow and practical. States signed treaties about fisheries, about migratory birds, and about boundary waters because they wanted to protect a shared resource that had commercial value. The famous arbitration between the United States and Canada over a smelter at Trail, decided in 1941, established that a state may not use its territory in a way that causes serious harm to another state. That single sentence became the seed of the modern duty of prevention. The space treaties belong to this early period in spirit. The 1967 Outer Space Treaty was written during the Cold War, and its purpose was strategic as much as environmental. It declared that outer space is the province of all humankind, that it cannot be claimed by any nation, that it must be used for peaceful purposes, and that states bear international responsibility for national activities, including those carried out by private companies. This last provision, in Article VI, is remarkable. It makes governments answerable for their private operators. No environmental treaty of the same period went so far. 4.2 Stockholm 1972 and the birth of a system The Stockholm Conference produced a declaration rather than a treaty, but it created the United Nations Environment Programme and it planted a principle that still governs the field: states have the sovereign right to exploit their own resources and the responsibility to ensure that activities within their control do not damage the environment of other states or of areas beyond national jurisdiction. The decade after Stockholm produced a wave of specific conventions: on wetlands, on world heritage, on wildlife trade, on migratory species, on long range air pollution, and on marine pollution from ships. Each dealt with one problem. Each was, by later standards, modest. Most of them still work. 4.3 Vienna, Montreal, and the model that succeeded The ozone regime deserves separate attention because it is the model that every later regime has tried and failed to copy. The 1985 Vienna Convention for the Protection of the #Ozone_Layer contained almost no substantive duties. It was a shell. Two years later the #Montreal_Protocol filled the shell with schedules that told each party exactly how much of each controlled substance it could produce and consume, and by when. The protocol created a #Multilateral_Fund to pay the extra costs of developing countries. It allowed the schedules to be tightened by later amendments without reopening the whole treaty. It banned trade in controlled substances with non parties, which made staying outside the regime expensive. The 2016 Kigali Amendment extended the same machinery to hydrofluorocarbons, which are not ozone depleting but are powerful greenhouse gases. The lesson is not that ozone was an easy problem. The lesson is that the regime was designed around a small number of chemicals produced by a small number of firms, with substitutes available and with a fund to pay for the switch. Every element of that design is missing from the climate regime. 4.4 Rio 1992 and the framework model The Rio Earth Summit produced the climate convention, the biodiversity convention, the desertification convention, and the Rio Declaration. It also fixed the framework convention plus protocol model as the standard template. The advantage of the template is that it allows states to agree on a goal before they agree on the means. The disadvantage is that the second step, the step where the real duties are written, may never come, or may come in a form so flexible that the duties are self chosen. 4.5 After Rio The three decades since Rio have produced a steady stream of instruments: the Kyoto Protocol in 1997, the Cartagena Protocol on biosafety in 2000, the Stockholm Convention on persistent organic pollutants in 2001, the Nagoya Protocol on access and benefit sharing in 2010, the Minamata Convention on #Mercury in 2013, the Paris Agreement in 2015, the Kunming Montreal framework in 2022, and the agreement on marine biological diversity beyond national jurisdiction in 2023. The direction of travel is clear. Obligations have become more procedural, more flexible, and more dependent on national choice. Participation has become nearly universal. The link between the two is not accidental. 5. The Climate Regime 5.1 Structure The climate regime has three layers. The United Nations Framework Convention on Climate Change, agreed in 1992, sets the objective: to stabilise greenhouse gas concentrations at a level that prevents dangerous interference with the climate system. The Kyoto Protocol, agreed in 1997, set binding emission targets for developed countries only. The #Paris_Agreement, agreed in 2015, replaced that model with a universal one. Under Paris, every party must prepare, communicate, and maintain a nationally determined contribution, one of the #Nationally_Determined_Contributions that form the core of the regime, must pursue domestic measures to achieve it, and must report on progress under an enhanced transparency framework. Successive contributions must represent a progression and reflect the highest possible ambition. The collective goal is to hold the increase in global average temperature well below two degrees above pre industrial levels and to pursue efforts to limit it to one and a half degrees. A #Global_Stocktake takes place every five years. The crucial legal point is that the obligation to prepare and communicate a contribution is binding, while the content of that contribution is not set by the treaty. This is the great compromise of #UNFCCC diplomacy. It bought participation at the price of ambition. 5.2 The recent record The first global stocktake concluded at the twenty eighth conference in Dubai in 2023. It called on parties to contribute to a transition away from fossil fuels in energy systems, in a just, orderly and equitable manner, and to triple renewable energy capacity by 2030. The twenty ninth conference in Baku in 2024 agreed a new collective quantified goal on #Climate_Finance, with developed countries taking the lead in mobilising at least three hundred billion dollars a year by 2035 and a wider call to scale finance from all sources to at least one point three trillion dollars a year by the same date. The thirtieth conference met in Belem, Brazil, in November 2025. It adopted a package known as the Global Mutirao decision, a Portuguese word of indigenous origin meaning collective effort. The decision launched a Global Implementation Accelerator and a Belem Mission to 1.5 to help countries deliver their contributions and their national adaptation plans. It created a mechanism for a #Just_Transition. It also confirmed the operational and replenishment arrangements for the fund on #Loss_And_Damage that was agreed in 2022 and made operational in 2023. It adopted a reduced set of indicators for the global goal on #Adaptation, known as the Belem adaptation indicators. It called for a tripling of adaptation finance by 2035, a weaker outcome than the 2030 date that many developing countries wanted. It established a two year work programme on climate finance and opened a dialogue on trade and climate for the first time. What Belem did not do is as important as what it did. More than eighty countries asked for a roadmap to transition away from fossil fuels, and a similar group asked for a roadmap on halting and reversing #Deforestation. Neither appeared in the adopted text. The presidency announced that it would take both roadmaps forward outside the formal process, under its own authority. This is a significant development. It shows states moving key questions out of the consensus based treaty process into voluntary coalitions, precisely because the consensus rule allows a determined minority to block. 5.3 The intervention of the courts The most important development in climate law since Paris did not happen in a negotiation room. On 23 July 2025 the International Court of Justice delivered a unanimous advisory opinion on the obligations of states in respect of climate change, following a campaign begun by Pacific island law students and carried to the General Assembly by Vanuatu. The Court held that climate treaties create binding obligations, that the customary duty to prevent significant harm applies to the climate system, that this duty must be performed with #Due_Diligence, and that the one and a half degree limit is the reference point for climate policy. The Court also indicated that the production and consumption of fossil fuels, the granting of exploration licences, and the provision of subsidies may, in certain circumstances, amount to internationally wrongful acts (Foster, 2025). The opinion is not binding in itself. Its power lies elsewhere. It gives national courts an authoritative statement of the law, and it strengthens the hand of claimants in #Climate_Litigation everywhere. It also transforms the legal status of the Paris regime. What looked like a set of self chosen political promises now sits inside a framework of due diligence obligations that a court can assess. Together with the 2024 opinion of the International Tribunal for the Law of the Sea, which treated greenhouse gases as marine pollution, it marks the arrival of judicial pressure as a third force in climate governance, alongside diplomacy and markets. 5.4 Assessment against the six criteria Clarity of obligation is low on substance and high on procedure. Measurability is moderate and improving, because the #Reporting and transparency framework now produces comparable national reports. Finance exists but is far below need, and the adaptation gap is severe: reported adaptation flows to developing countries fell rather than rose in the most recent reporting year, against needs estimated in the hundreds of billions. The science to policy channel, the Intergovernmental Panel on Climate Change, is the strongest in any regime. Review exists but has no teeth. Coverage of private actors is indirect and weak. The regime therefore scores well on participation and poorly on delivery. The gap between the sum of national contributions and the emissions pathway consistent with the temperature goal, usually called the ambition gap, remains the central failure of #Mitigation policy. 6. The Nature and Biodiversity Regime 6.1 The Convention on Biological Diversity The 1992 convention has three objectives: the conservation of #Biodiversity, the sustainable use of its components, and the fair and equitable sharing of benefits arising from the use of genetic resources. The third objective is often forgotten by students, but it is the reason that many developing countries joined. The convention treats genetic resources as subject to national sovereignty, which reversed an older assumption that they were the common heritage of humankind. Two protocols followed. The Cartagena Protocol of 2000 governs the transboundary movement of living modified organisms and gives importing states a right to be informed in advance. The Nagoya Protocol of 2010 gives providers of genetic resources a right to prior informed consent and to mutually agreed terms for benefit sharing. 6.2 The Kunming Montreal Global Biodiversity Framework The framework, adopted in December 2022, is the successor to the failed Aichi targets. It contains four long term goals for 2050 and twenty three action targets for 2030. The best known is target three, sometimes called #Thirty_By_Thirty, the commitment to ensure that at least thirty per cent of terrestrial, inland water, coastal and marine areas are effectively conserved and managed by 2030. Others cover restoration of degraded ecosystems, reduction of pollution risk, reform of harmful subsidies, and the mobilisation of at least two hundred billion dollars a year for biodiversity from all sources, including twenty billion a year in international flows by 2025, rising to thirty billion by 2030. The framework is not a treaty. It is a decision of the conference of the parties. Its targets bind no one directly. Everything depends on national biodiversity strategies and action plans, on national targets, and on domestic law. Hughes and Grumbine (2023) argue that many targets fail the test of being specific, measurable, and time bound, and that loopholes in definitions allow states to claim success without delivering it. McGowan and colleagues (2024) show that the species elements of the framework need much clearer interpretation before progress can be assessed. Boran and colleagues (2024) argue that the biodiversity and climate regimes should run a joint work programme, because the two crises share drivers and solutions and the two conventions currently pull in different directions. 6.3 The Cali and Rome decisions The sixteenth conference of the parties met in Cali, Colombia, in late 2024. It agreed a permanent subsidiary body for indigenous peoples and local communities, giving them a formal seat in the machinery of the convention, and it created a mechanism for sharing benefits from the use of digital sequence information on genetic resources. It then ran out of time and lost quorum, leaving central decisions unfinished. The conference resumed in Rome in February 2025 and completed its work. It adopted a strategy for resource mobilisation, the monitoring framework with its full set of indicators, and a mechanism for planning, monitoring, reporting and review. The Cali Fund, which channels contributions from companies that benefit commercially from digital sequence information, was launched in the margins of the same meeting, with a significant share reserved for indigenous peoples and local communities. These decisions matter because they turn a political framework into a system that can be measured. Whether the money follows is another question. The distributive problem identified by Lehmann and colleagues (2025) remains: the framework acknowledges differentiated responsibility in principle but does not say who pays in practice. 6.4 The older nature conventions Four older conventions continue to do quiet and useful work. The #Ramsar_Convention on #Wetlands of 1971 requires parties to designate wetlands of international importance and to promote their wise use. It has no enforcement mechanism, but the listing process creates reputational pressure and gives domestic campaigners a legal hook. The World Heritage Convention of 1972 protects sites of outstanding universal value, including natural sites. Its in danger list has become an unexpectedly powerful tool, because governments dislike the publicity. The Convention on International Trade in Endangered Species, agreed in 1973, is the strongest of the four because it uses trade controls. Species are placed in appendices, and trade in listed specimens requires permits. Parties can suspend trade with a party that fails to implement the convention. This makes #Wildlife_Trade rules genuinely enforceable in a way that most environmental duties are not. The Convention on Migratory Species of 1979 provides a framework for range states to cooperate on species that cross borders, mainly through subsidiary agreements and memoranda of understanding. 6.5 Land and drought The United Nations Convention to Combat #Desertification, adopted in 1994, is the third Rio convention and the least studied. It works through national action programmes and, since 2015, through the concept of #Land_Degradation_Neutrality, which asks states to balance losses of healthy land with gains elsewhere. Its weakness is finance: it has no dedicated fund of its own comparable to the climate or biodiversity mechanisms, and it depends on general development assistance. 7. Ocean and Atmosphere 7.1 The law of the sea The 1982 United Nations Convention on the Law of the Sea is often called the constitution for the oceans. Part XII imposes a general obligation on states to protect and preserve the marine environment. The deep seabed beyond national jurisdiction, and its mineral resources, are declared the common heritage of mankind and are administered by the International Seabed Authority, the body that governs the #Deep_Seabed. The convention left a large hole. It said little about the conservation of living resources in the water column beyond national jurisdiction, which covers roughly two thirds of the ocean. For forty years the high seas were open to anyone who could reach them, with no general mechanism for creating protected areas and no requirement to assess the environmental impact of new activities. 7.2 The BBNJ Agreement The Agreement on the Conservation and Sustainable Use of Marine Biological Diversity of Areas Beyond National Jurisdiction, adopted on 19 June 2023, fills that hole. It has four pillars: marine genetic resources and the fair sharing of benefits; area based management tools, including marine protected areas; environmental impact assessment; and capacity building and the transfer of marine technology. The agreement reached the required sixty ratifications on 19 September 2025 and entered into force on 17 January 2026, with more than eighty parties by that date. Kim (2024) argues that the treaty's main effect will be structural: it creates a conference of the parties with the authority to make decisions that other bodies must respect, which begins to knit together a fragmented ocean governance system built of regional fisheries bodies, the seabed authority, and the shipping organisation. Taghizadeh (2025) reads the agreement through the lens of intergenerational and intra generational #Equity and finds that it goes further than earlier instruments in embedding both. The agreement is important for #Ocean_Governance for a further reason. Without a mechanism to create high seas protected areas, the thirty per cent target of the biodiversity framework was arithmetically impossible, because so much of the ocean lies beyond national borders. The two regimes are now, for the first time, mutually supporting. The open questions concern implementation. The first conference of the parties must be convened within a year of entry into force. Decisions must be made on the clearing house mechanism, on the scientific and technical body, and on how the agreement will interact with existing sectoral organisations that jealously guard their mandates. #Marine_Biodiversity protection will succeed or fail on those institutional details rather than on the treaty text. 7.3 The atmosphere Beyond ozone, the atmosphere is governed mainly by regional instruments. The 1979 Convention on Long Range Transboundary Air Pollution, with its protocols on sulphur, nitrogen oxides, volatile organic compounds, and heavy metals, is a quiet success. It reduced acid rain in Europe and North America substantially. Its method, which combines a scientific assessment body with negotiated emission ceilings and a strong review procedure, is closer to the ozone model than to the climate model. The lesson repeats: when the problem is bounded and the science is agreed, treaty law works. 8. Chemicals, Waste and Pollution 8.1 The chemicals cluster Three conventions form a cluster that now shares meetings and a joint secretariat. The #Basel_Convention of 1989 controls the transboundary movement of #Hazardous_Waste and requires prior informed consent from the importing state. Its 1995 ban amendment, which entered into force in 2019, prohibits exports of hazardous waste from wealthier to developing countries. Amendments adopted in 2019 brought most plastic waste within the control system, which changed global waste flows almost immediately. The Rotterdam Convention of 1998 requires prior informed consent for trade in listed hazardous chemicals and pesticides. The Stockholm Convention of 2001 targets #Persistent_Organic_Pollutants, chemicals that resist degradation, accumulate in tissue, and travel long distances. It requires elimination or restriction of listed substances and provides for new substances to be added over time. The Minamata Convention of 2013 closes the cluster. It controls mercury across its life cycle, from mining to trade to emissions to waste, and it addresses artisanal gold mining, which is the largest single source of mercury releases. These treaties score well on clarity and measurability, because they list substances and set control measures. They score poorly on finance and on capacity: many parties simply lack the laboratories, the customs training, and the disposal facilities that compliance requires. 8.2 The plastics negotiation and its failure In March 2022 the United Nations Environment Assembly decided to negotiate a legally binding instrument on #Plastic_Pollution, including in the marine environment, based on a full #Life_Cycle_Approach. The mandate was ambitious and the timetable was short: five sessions, ending in 2024. The negotiations did not end in 2024. The fifth session in Busan closed without agreement. A resumed session met in Geneva in August 2025 with one hundred and eighty four countries present. It also closed without agreement. The chair produced two draft texts in the final week, and neither attracted sufficient support. The chair resigned in October 2025. A short organisational session in Geneva in February 2026 was convened mainly to elect a new chair, with substantive talks postponed to a later session in 2026. The dispute is simple to state. A group of more than one hundred countries wants binding global measures that include limits on the production of primary plastic polymers, controls on chemicals of concern, and product design requirements. A smaller group of oil and gas producing states wants a treaty focused on waste management, recycling, and voluntary national measures. Behind this lies the fact that plastic is a fossil fuel product, and that limits on plastic production are limits on future demand for oil and gas. The scientific literature is unusually united. Bergmann and colleagues (2022) argue that a treaty must cap production to be effective. Villarrubia Gomez and colleagues (2024) show that plastic pollution worsens the pressure on multiple planetary boundaries at once, which means it cannot be treated as a simple waste problem. Baztan and colleagues (2024) call for urgent upstream reduction of primary polymers. Farrelly and colleagues (2024) warn that the negotiation needs trusted, independent science, and note the heavy presence of industry lobbyists at the sessions. March and Winton (2025) and Brooks and Havas (2025) both argue that a systems approach, covering design, use, and disposal, is the only route to a workable agreement. The plastics failure is the clearest recent illustration of the structural weakness of consensus decision making. When the negotiating rules allow any state to block, the treaty that emerges is the treaty that the least willing party can accept. 9. The Space Conventions 9.1 The five instruments Space law rests on five United Nations treaties, all negotiated between 1967 and 1979. The Outer Space Treaty of 1967 is the constitution of the field. It provides that the exploration and use of outer space shall be carried out for the benefit and in the interests of all countries and shall be the province of all mankind; that space is free for exploration and use by all states; that space is not subject to national appropriation by claim of sovereignty, by use, by occupation, or by any other means; that states shall bear international responsibility for national activities, whether carried out by governmental agencies or by non governmental entities, and that such non governmental activities require authorisation and continuing supervision by the appropriate state; that states shall retain jurisdiction over objects they register; and that states shall avoid harmful contamination of space and adverse changes in the environment of the Earth. The Rescue Agreement of 1968 requires states to assist astronauts in distress and to return them and recovered space objects. The #Liability_Convention of 1972 creates absolute liability for damage caused on the surface of the Earth or to aircraft in flight, and fault based liability for damage elsewhere, including in orbit. The Registration Convention of 1975 requires launching states to register objects with the Secretary General. The Moon Agreement of 1979 declares the Moon and its natural resources the common heritage of mankind and requires an international regime to govern their exploitation. It attracted very few ratifications and no major spacefaring state, which left the resource question unsettled. 9.2 Environmental protection in orbit The environmental provisions of the space regime are thin. Article IX of the #Outer_Space_Treaty requires due regard for the corresponding interests of others and asks states to avoid harmful contamination, but it does not define the terms and it creates no mechanism to apply them. The gap has been filled by #Soft_Law. The Inter Agency Space Debris Coordination Committee published debris mitigation guidelines that the United Nations Committee on the Peaceful Uses of Outer Space endorsed in 2007. In 2019 the same committee adopted twenty one Guidelines for the Long Term Sustainability of Outer Space Activities, covering registration practice, conjunction assessment, the sharing of orbital data, and the design of spacecraft for safe disposal. These instruments are voluntary. They are also, in the view of most operators, sensible. The problem is that a single non compliant actor can impose costs on everyone. #Space_Debris is a classic negative externality with a long tail. A fragment created today can remain in orbit for decades or centuries. Destructive #Anti_Satellite_Tests, the largest single source of long lived debris, remain lawful under treaty law, although a growing number of states have made unilateral political commitments not to conduct them. The scale of the change is important for students to grasp. The regime was designed for a world in which a small number of governments launched a small number of large satellites. Today, commercial constellations place thousands of small satellites in low Earth orbit, and several more constellations of comparable size are planned. Under the Liability Convention, fault must be proved for damage in orbit, and there is no agreed standard of care against which fault could be assessed. The result is a liability rule that is unlikely to be used. 9.3 The Artemis Accords and the fragmentation problem The Artemis Accords, launched in October 2020 with eight original signatories, are a set of non binding principles for civil space exploration. They affirm peaceful purposes, transparency, interoperability, emergency assistance, registration, the release of scientific data, the preservation of heritage sites, debris mitigation, and the disposal of spacecraft. They also assert two contested positions: that the extraction and use of #Space_Resources does not, in itself, constitute national appropriation prohibited by the Outer Space Treaty, and that operators may establish safety zones around their activities to avoid harmful interference. Growth has been rapid. By mid 2026 the #Artemis_Accords had sixty eight signatories, drawn from every region. The legal debate is sharp. Deplano (2021) argues that the significance of the Accords lies in the method: a leading state has created a set of bilateral political commitments that operate outside the multilateral forum where space law has traditionally been made, and this may pull the interpretation of the Outer Space Treaty in a particular direction without any formal amendment. Tronchetti and Liu (2021) trace the same trend in national legislation on space resources. McKeown, Dempster and Saydam (2022) test the safety zone concept against the practical needs of lunar mining and find real difficulties. Schmidt and Svec (2022) argue that the deadlock between the free access and common heritage positions can only be broken by building a new institution rather than by re reading old texts. Prema, Bharathi and Sundar (2024) take a more positive view and see the Accords as a stage in the growth of the law. Wang (2023) analyses the parallel International Lunar Research Station initiative, which shows that the alternative to one club is not universality but two clubs. For #Space_Sustainability the risk is obvious. If different groups of states operate under different interpretations of the same treaty, then the norms that protect the orbital environment become a matter of club membership rather than of law. 9.4 Assessment against the six criteria Clarity is high on principle and very low on operational detail. Measurability is weak: there is no agreed method for counting compliance with debris rules. Finance is absent, because there is no fund. The science to policy channel exists in the technical subcommittee of the United Nations committee but has no power to set rules. Review is voluntary. Coverage of private actors is, on paper, the strongest of any regime studied here, because of Article VI, but it depends entirely on national licensing, and licensing standards vary widely, which creates an incentive for operators to register where rules are loosest. The space regime therefore has the opposite profile from the climate regime. It has strong principles and weak machinery, where climate has weak principles and increasingly strong machinery. 10. Comparative Analysis 10.1 A common architecture Every regime studied here uses the same institutional parts. There is a framework instrument that states the goal. There is a #Conference_Of_The_Parties that meets at intervals and takes decisions by consensus. There is a secretariat. There is a scientific or technical body. There is, usually, a financial mechanism. There is a reporting duty. This architecture is now so standard that it is copied automatically, and the copying is not always wise. A regime that governs a small number of chemicals does not need the same machinery as a regime that governs an entire economy. 10.2 What separates the effective from the ineffective The comparison across the six criteria suggests four conditions for effectiveness. The first condition is a bounded problem. The ozone regime controls a defined list of substances. The wildlife trade regime controls a defined list of species. The chemicals conventions control defined lists of chemicals. In each case, a state can know exactly whether it has complied. The climate regime and the biodiversity framework, by contrast, ask states to change the whole structure of their economies and their land use, and no list can capture that. The second condition is money that arrives. The multilateral fund of the ozone regime worked because it paid the incremental costs of the transition in developing countries. Climate finance and biodiversity finance are both far below the level that the parties themselves have declared necessary. The adaptation finance figures reported before the Belem conference showed a decline rather than an increase, which damages trust in every other part of the bargain. The third condition is a trusted scientific channel. The Intergovernmental Panel on Climate Change is the strongest example, and its assessments were treated as the best available science by the International Court of Justice, which shows the value of a strong #Science_Policy_Interface in 2025. Where the channel is weak or contested, as in the plastics negotiations, industry funded claims compete with independent research and the negotiation slows. The fourth condition is consequence. Not necessarily punishment, but consequence: publication, peer review, trade measures, loss of standing, or litigation. The regimes with trade measures, ozone and wildlife trade, are the strongest. The regime now acquiring consequence most rapidly is climate, not through its own compliance committee but through the courts. 10.3 The consensus problem Almost all of these bodies decide by consensus, and this #Consensus_Rule shapes every outcome. Consensus protects sovereignty and produces universal membership, but it hands a veto to the least willing participant. The plastics negotiation shows the cost. The Belem outcome shows an emerging response: when the consensus process blocks an issue, willing states move it outside the process into coalitions and presidency led roadmaps. This is understandable, but it carries the same fragmentation risk that the space regime now faces. A world of coalitions is a world in which the states causing the most harm can simply stay out. 10.4 Private actors The environmental conventions were written for states, but most environmental harm is caused by companies. The climate regime reaches companies only through national policy. The biodiversity framework asks parties to require large businesses to disclose their impacts, but the target is soft. The plastics negotiation would have reached producers directly, which is exactly why it stalled. The space regime, uniquely, makes states internationally responsible for their private operators. Yet even there the duty is discharged through national licensing, and the standard of licensing is a national choice. The lesson is that the treaty system has not solved the problem of the private actor. Until it does, the regimes will remain one step removed from the source of the harm. 10.5 Fragmentation and interplay The regimes overlap. Plastic is a waste problem, a chemicals problem, a marine problem, and a climate problem. Forests are a climate sink, a biodiversity habitat, and a livelihood. Satellites monitor climate and biodiversity, so the sustainability of orbit is now a condition of environmental #Monitoring on Earth. This overlap is generally treated as a defect and given the name #Fragmentation, but it can also be an opportunity. Boran and colleagues (2024) propose a joint work programme between the climate and biodiversity conventions. The high seas agreement is deliberately designed to sit under the law of the sea while serving the biodiversity framework. Interplay, managed well, is a source of strength. 11. Discussion 11.1 A regime in transition The evidence assembled here suggests that treaty based environmental governance is entering a new phase. In the first phase, from 1972 to 1992, states wrote narrow treaties about specific problems and largely kept them. In the second phase, from 1992 to 2015, states wrote broad framework treaties with universal membership and weak duties. In the third phase, which is beginning now, three things are happening at once. First, courts are supplying the consequence that the treaties lack. The 2025 advisory opinion of the International Court of Justice reframed voluntary contributions as obligations of due diligence. The 2024 opinion of the law of the sea tribunal treated greenhouse gases as marine pollution. National courts now have a settled interpretation to apply. #ICJ_Advisory_Opinion reasoning will appear in domestic judgments for the next decade. Second, #Coalitions_Of_The_Willing are supplying the ambition that consensus blocks. The Belem presidency roadmaps, the group of more than one hundred states in the plastics talks, and the Artemis Accords in space are all examples of the same manoeuvre. The manoeuvre works, but it fragments. Third, measurement is improving faster than money. The biodiversity monitoring framework, the Belem adaptation indicators, and the climate transparency framework all produce better data. Better data makes failure visible, which is useful. But visibility without finance produces cynicism, and #Transparency without #Capacity_Building simply exposes the countries that can least afford to act. 11.2 The space warning The space regime is worth studying because it is where environmental law was in 1970. There are grand principles, no operational rules, voluntary guidelines, rapid commercial growth, and a physical resource, the orbital environment, that can be degraded irreversibly. The environmental field learned, slowly and expensively, that voluntary guidelines do not restrain a competitive industry. There is no reason to expect a different outcome in orbit. The prudent step, well supported in the literature, is to convert the long term sustainability guidelines into binding national licensing conditions, to agree a common standard of care so that the fault rule in the Liability Convention becomes usable, and to place a binding prohibition on debris creating anti satellite tests. 11.3 What students should take from this Three practical lessons follow. First, read the operational articles, not the preamble. The preamble of a treaty tells you what states wanted to be seen to believe. The operational articles tell you what they agreed to do. In most conventions the real duties are procedural, and the procedural duties are where litigation and pressure actually bite. First among the procedural duties is reporting, and reporting is the foundation of everything else. Second, follow the money. A treaty without a funded financial mechanism is a treaty that will be implemented only by states that would have acted anyway. The single best predictor of whether a developing country party performs is whether the regime pays the incremental cost. Third, watch the decision rule. Consensus produces universality and weakness. Voting produces strength and defection. Every regime studied here sits somewhere on that trade off, and most of the political drama in any conference of the parties is really an argument about where on that line the regime should sit. 12. Limitations This study is doctrinal and comparative. It does not measure environmental outcomes, and it cannot prove that any treaty caused any change in emissions, in species numbers, or in debris counts. It relies on the texts, on institutional records, and on secondary literature. Its selection of regimes is deliberate but not exhaustive: nuclear safety, transboundary watercourses, and the regional seas programmes are omitted for reasons of space. The most recent developments, especially those of 2025 and 2026, are assessed close to the events and may look different with more distance. Finally, the six criteria used for comparison are a simplification. They capture legal design well and political economy poorly. 13. Conclusion and Recommendations International conventions on the environment, climate, nature and space are the best tools that humanity has built so far for governing what everyone shares. They have real achievements. The ozone layer is recovering. Acid rain in Europe fell sharply. Trade in many endangered species is controlled. Persistent organic pollutants are being eliminated. In January 2026 a treaty entered into force that, for the first time, allows the protection of life in the half of the planet that lies beyond any national border. In July 2025 the world's highest court told every government that protecting the climate system is a legal duty, not a policy choice. They also have deep failures. Emissions continue to rise. Biodiversity continues to decline. The plastics negotiation has collapsed twice. Adaptation finance is falling when it should be tripling. Orbit is filling with debris under a set of rules that nobody is obliged to follow. The comparison in this article suggests that the difference between success and failure is mostly a matter of design. Bounded problems, funded transitions, trusted science, and real consequences produce results. Broad ambitions, unfunded promises, contested evidence, and consensus vetoes do not. Six recommendations follow. Design for measurability. Every target should have a defined indicator and a baseline before it is adopted, not years afterwards. The delay between the adoption of the #Kunming_Montreal_Framework in 2022 and the completion of its monitoring framework in 2025 cost three years of accountability. Fund the transition or expect non compliance. Financial mechanisms should be assessed against the incremental cost of compliance in the poorest parties, and #Technology_Transfer should be treated as a condition of the bargain rather than a favour. Reform the decision rule. Regimes should adopt qualified majority voting for technical annexes and lists, keeping consensus for constitutional questions only. This single change would have delivered a plastics treaty. Reach private actors directly. Treaties should require parties to impose specific licensing and disclosure duties on operators, following the model of Article VI of the Outer Space Treaty rather than the model of general national policy. Harden space sustainability now. The long term sustainability guidelines should be converted into binding obligations through national law, a common standard of care should be agreed so that the #Liability rule becomes usable, and destructive anti satellite testing should be prohibited by treaty. Manage interplay deliberately. Joint work programmes between the climate and biodiversity conventions, and formal coordination between the high seas agreement and existing ocean bodies, should be built into the institutions rather than left to goodwill. The next decade will decide whether treaty based cooperation can still solve planetary problems, or whether it will be replaced by coalitions, courts, and markets. On the evidence gathered here, the most likely outcome is a mixture of all four. Students entering this field should therefore learn to read a treaty, to read a balance sheet, and to read a judgment, because the future of #Environmental_Justice will be written in all three languages. Hashtags #Environmental_Conventions #Climate_Conventions #Nature_Conventions #Space_Conventions #International_Environmental_Law #Global_Environmental_Governance #Climate_Treaties #Biodiversity_Law #Space_Law_And_Policy #Planetary_Commons #Treaty_Compliance_Research #Sustainability_Studies #Ocean_And_Space_Governance #STULIB_Research References Baztan, J., Jorgensen, B., Almroth, B. C., Bergmann, M., Farrelly, T., Muncke, J., and Wagner, M. (2024). Primary plastic polymers: urgently needed upstream reduction. Cambridge Prisms: Plastics, 2, e7. https://doi.org/10.1017/plc.2024.8 Bergmann, M., Almroth, B. C., Brander, S. M., Dey, T., Green, D. S., Gundogdu, S., and Walker, T. R. (2022). A global plastic treaty must cap production. Science, 376(6592), 469-470. https://doi.org/10.1126/science.abq0082 Boran, I., and colleagues (2024). The Kunming-Montreal Global Biodiversity Framework and the Paris Agreement need a joint work programme for climate, nature and people. Journal of Applied Ecology, 61(9). https://doi.org/10.1111/1365-2664.14721 Brooks, A. L., and Havas, V. (2025). Strengthening global plastic policy with systems analysis. Nature Sustainability, 8(6), 714-723. https://doi.org/10.1038/s41893-025-01554-4 Deplano, R. (2021). The Artemis Accords: evolution or revolution in international space law? International and Comparative Law Quarterly, 70(3), 799-819. https://doi.org/10.1017/S0020589321000142 Farrelly, T., and colleagues (2024). Global plastics treaty needs trusted science. Science, 384(6693), 281. https://doi.org/10.1126/science.adp4264 Foster, C. E. (2025). The 2025 International Court of Justice advisory opinion on obligations of states in respect of climate change. International and Comparative Law Quarterly, 1-35. https://doi.org/10.1017/S0020589325101292 Hughes, A. C., and Grumbine, R. E. (2023). The Kunming-Montreal Global Biodiversity Framework: what it does and does not do, and how to improve it. Frontiers in Environmental Science, 11, 1281536. https://doi.org/10.3389/fenvs.2023.1281536 Kim, R. E. (2024). The likely impact of the BBNJ Agreement on the architecture of ocean governance. Marine Policy, 165, 106190. https://doi.org/10.1016/j.marpol.2024.106190 Lehmann, I., Kok, M., Immerzeel, R., and Marques, A. (2025). Implications of global distributive justice principles for implementation of the Kunming-Montreal Global Biodiversity Framework. Conservation Biology. https://doi.org/10.1111/cobi.70167 March, A., and Winton, S. (2025). Unlocking opportunities to fulfil the Global Plastics Treaty. Nature Reviews Earth and Environment, 6, 435-436. https://doi.org/10.1038/s43017-025-00689-4 McGowan, P. J. K., Hutchinson, A., Brooks, T. M., Elliott, W., Hoffmann, M., Mair, L., McDougall, A., Raimondo, D. C., and Butchart, S. H. M. (2024). Understanding and achieving species elements in the Kunming-Montreal Global Biodiversity Framework. BioScience, 74(9), 614-623. https://doi.org/10.1093/biosci/biae065 McKeown, B., Dempster, A. G., and Saydam, S. (2022). Artemis Accords: are safety zones practical for long term commercial lunar resource utilisation? Space Policy, 62, 101504. https://doi.org/10.1016/j.spacepol.2022.101504 Prema, E., Bharathi, R. V., and Sundar, V. S. (2024). The Artemis Accords: a new phase in the development of outer space law. In M. A. Pozza (Ed.), Ascending to Space. Space Law and Policy. Singapore: Springer. https://doi.org/10.1007/978-981-97-0714-0_4 Rajamani, L., and Peel, J. (2021). Reflections on a decade of change in international environmental law. Cambridge International Law Journal, 10(1), 6-31. https://doi.org/10.4337/cilj.2021.01.01 Richardson, K., Steffen, W., Lucht, W., and colleagues (2023). Earth beyond six of nine planetary boundaries. Science Advances, 9(37), eadh2458. https://doi.org/10.1126/sciadv.adh2458 Schmidt, N., and Svec, M. (2022). Breaking the deadlock in the space mining legal debate. New Space, 10(2), 115-126. https://doi.org/10.1089/space.2021.0049 Taghizadeh, Z. (2025). Intergenerational and intra-generational equity under the BBNJ Agreement: advancing accountability towards sustainable management of the marine environment. Environmental Management. https://doi.org/10.1007/s00267-025-02256-5 Tronchetti, F., and Liu, H. (2021). The White House Executive Order on the recovery and use of space resources: pushing the boundaries of international space law? Space Policy, 57, 101448. https://doi.org/10.1016/j.spacepol.2021.101448 Tronchetti, F., and Liu, H. (2021). Australia's signing of the Artemis Accords: a positive development or a controversial choice? Australian Journal of International Affairs, 75(3), 243-251. https://doi.org/10.1080/10357718.2021.1893268 Villarrubia-Gomez, P., Almroth, B. C., Eriksen, M., Ryberg, M., and Cornell, S. E. (2024). Plastics pollution exacerbates the impacts of all planetary boundaries. One Earth, 7(12). https://doi.org/10.1016/j.oneear.2024.10.017 Wang, X. (2023). The International Lunar Research Station: China's new era of space cooperation and its new role in the space legal order. Space Policy, 65, 101537. https://doi.org/10.1016/j.spacepol.2022.101537

The United Nations Framework Convention on Climate Change, adopted in New York in May 1992 and opened for signature at the Rio Earth Summit in June of that year, is the founding legal instrument of the international response to a warming planet. This article offers a structured, student oriented but scholarly review of the Convention: why it was created, how it was designed, what it actually obliges states to do, and how its institutions have shaped three decades of climate diplomacy. Drawing on recent legal, political science and policy scholarship published mainly between 2021 and 2026, the article traces a line from the scientific alarm of the 1980s through the Kyoto Protocol of 1997, the Copenhagen breakdown of 2009, the Paris Agreement of 2015, and on to the first global stocktake at Dubai in 2023, the Baku finance decision of 2024, the Belem package of 2025 and the 2025 advisory opinion of the International Court of Justice. Four analytical claims are advanced. First, the Convention was deliberately designed as a framework treaty, trading immediate stringency for near universal participation and durable institutions. Second, its principles, above all common but differentiated responsibilities and respective capabilities, are not decorative preambular language but the operative grammar of every subsequent negotiation. Third, the machinery it created, the Conference of the Parties, the subsidiary bodies, the secretariat, the reporting system and the financial mechanism, has proved remarkably resilient even when substantive ambition has stalled. Fourth, the regime is now entering a phase in which its soft procedural obligations are being hardened from outside, through domestic and international courts, rather than from within through negotiation. The article concludes that the Convention should be read less as a failed emissions treaty than as a constitutional settlement that made subsequent, more demanding law possible, and that its principal unfinished business is the gap between agreed procedure and delivered outcome. Keywords: international environmental law, climate regime, multilateral negotiation, treaty design, environmental governance, climate finance, equity, compliance 1. Introduction Few treaties have been asked to carry as much weight as the #UNFCCC. Adopted on 9 May 1992 and opened for signature at the United Nations Conference on Environment and Development in Rio de Janeiro, the Convention entered into force on 21 March 1994 and now counts 198 Parties, a level of participation that is effectively universal in the international system. It is the instrument from which the #Kyoto_Protocol of 1997 and the #Paris_Agreement of 2015 descend. It created the annual meetings that the world now knows simply as the climate summits. It gave the world a stated objective for climate policy, a set of principles for allocating effort, a reporting system, a financial mechanism and a permanent bureaucracy. Everything that has followed in #global_climate_governance has been built inside, on top of, or in reaction against, the frame it laid down. Yet the Convention is also routinely misunderstood, and misunderstood in two opposite directions. One misreading treats it as a strong environmental treaty that simply failed, pointing to the fact that global #greenhouse_gases emissions were roughly forty per cent higher in the mid 2020s than they were when the treaty was signed. The other misreading treats it as an empty declaration, a piece of diplomatic theatre with no legal content at all. Neither view survives close contact with the text. The Convention was never designed to cap emissions by itself. It was designed to do something both more modest and more ambitious: to establish an agreed problem definition, an agreed objective, an agreed set of principles, and a permanent institutional venue in which harder commitments could be negotiated later, repeatedly, and by nearly every state on earth. This is the classic architecture of a #framework_convention, borrowed self consciously from the ozone regime and from earlier regional agreements on transboundary air pollution. This article is written for students who need to understand not only what the Convention says but why it says it that way, and for researchers who want a current map of the scholarly debate. It has four aims. The first is descriptive: to set out the origins, the legal architecture and the institutional machinery of the 1992 treaty with enough precision to be usable. The second is historical: to follow the regime through its major phases, from Kyoto to Paris to the present, showing how each phase responded to the perceived failures of the last. The third is analytical: to assess what the framework has and has not achieved, using the criteria that the scholarly literature itself proposes, and to examine the structural critiques that have been levelled at it. The fourth is prospective: to identify where the regime appears to be heading, particularly in light of the turn toward #climate_litigation and the 2025 advisory opinion of the #International_Court_of_Justice, and to suggest a research agenda. The argument that runs through the article is a simple one. The Convention purchased #universal_participation at the price of substantive weakness, and that bargain was defensible in 1992 and remains defensible now, because a treaty that a small number of ambitious states sign and a large number of major emitters ignore is not a climate regime at all. But the bargain has a bill attached, and the bill is now falling due. Three decades of procedure have produced institutions of impressive sophistication and an emissions curve that has bent but not broken. The central question for the next decade is whether the framework can convert its procedural strength into delivered #mitigation and delivered support, or whether that conversion will increasingly be forced upon it by courts, markets and domestic politics operating outside the negotiating rooms. 2. Approach and Scope This article is a structured narrative review rather than an empirical study. It synthesises three bodies of work. The first is international legal scholarship on the design, interpretation and legal character of the climate treaties. The second is political science and international relations scholarship on regime effectiveness, negotiation dynamics, coalitions and legitimacy. The third is applied policy analysis produced by research institutes and by the intergovernmental scientific and reporting bodies themselves, principally the Intergovernmental Panel on Climate Change and the United Nations Environment Programme. The literature base is drawn overwhelmingly from work published between 2021 and 2026, on the reasoning that the regime changed materially with the completion of the Paris rulebook in 2021 and again with the first #global_stocktake in 2023, so that older assessments, however classic, describe an institution that no longer exists in quite the same form. Where older sources are indispensable, the argument is reconstructed through recent work that engages with them. Primary legal materials, meaning the text of the Convention itself, the decisions of the #Conference_of_the_Parties, and the 2025 advisory opinion of the International Court of Justice, are treated as sources in their own right. Two limitations should be stated at the outset. First, the article is regime centred. It says relatively little about the vast domestic and transnational activity, from municipal climate plans to corporate net zero pledges, that now constitutes a large share of real world climate action, except where that activity intersects with the Convention's own machinery. A fuller account would situate the treaty within the wider landscape of #polycentric_governance. Second, the article does not attempt a quantitative estimate of the Convention's causal effect on emissions. Such an estimate is methodologically fraught, because the counterfactual world without a climate regime is unobservable, and because the treaty's principal effects are plausibly indirect, working through expectations, domestic legislation and investment signals rather than through direct compulsion. 3. Origins: How the World Arrived at a Framework Treaty 3.1 The scientific prelude The physics underlying the Convention long predates it. The warming effect of atmospheric carbon dioxide was described in the nineteenth century and quantified, crudely, at the turn of the twentieth. What changed in the second half of the twentieth century was the accumulation of observational evidence and the emergence of a scientific community organised well enough to speak to governments with a single voice. Continuous atmospheric measurement from the late 1950s established that carbon dioxide concentrations were rising, not merely that they might. Through the 1970s and 1980s, improvements in general circulation modelling, together with a growing palaeoclimate record, converted a theoretical concern into a projection. The institutional turning point came in 1988, when the World Meteorological Organization and the United Nations Environment Programme established the #IPCC. This move mattered enormously, and its significance is easy to underestimate. By creating a body whose assessments would be produced by scientists but formally approved, line by line in its summaries, by governments, the international community solved a problem that has crippled other environmental negotiations: the problem of contested facts. Governments could not easily disown in a negotiating room a scientific conclusion that their own delegations had accepted in a plenary. The First Assessment Report of 1990 concluded that emissions from human activity were substantially increasing the atmospheric concentration of greenhouse gases and that this would enhance the greenhouse effect. That conclusion was the direct trigger for the negotiating mandate that produced the Convention. 3.2 The political prelude Science alone does not produce treaties. Several political currents converged in the late 1980s and early 1990s. The end of the Cold War created, briefly, an unusual appetite for multilateral problem solving and freed diplomatic bandwidth that had been consumed by superpower confrontation. The Montreal Protocol of 1987 on ozone depleting substances had just demonstrated, in a way that was fresh in every negotiator's mind, that a framework convention followed by successively tightened protocols could actually work, and could do so while accommodating differences between industrialised and developing economies through a dedicated fund and delayed compliance schedules. The ozone precedent shaped climate negotiators' assumptions about what was possible, and arguably shaped them too optimistically, since the economics of replacing refrigerants bears little resemblance to the economics of replacing the energy basis of industrial civilisation. At the same time, small island states, whose territorial existence is directly threatened by sea level rise, organised themselves into an effective diplomatic bloc and injected a note of existential urgency that has never left the process. Developing countries more broadly approached the negotiations with a settled and coherent position: the accumulated stock of atmospheric greenhouse gases was overwhelmingly the product of industrialisation in the North; the South's overriding priority was development and poverty eradication; any climate regime that constrained Southern growth without addressing Northern #historical_responsibility would be neither fair nor acceptable. That position, articulated with great consistency by the Group of 77 and China, is the single most important political fact about the Convention's design. The negotiations themselves were conducted by an Intergovernmental Negotiating Committee established by the General Assembly in December 1990, which met over five sessions in fifteen months. The compression of that timetable was driven by the political deadline of the #Rio_Earth_Summit: the text had to be ready for signature in June 1992. This deadline explains a great deal about the resulting instrument. It is a treaty that had to be finished, and being finishable took priority over being stringent. 3.3 The core bargain The central disagreement was over targets. The European Community, together with several other industrialised states, pressed for a binding commitment by developed countries to stabilise carbon dioxide emissions at 1990 levels by the year 2000. The United States, then the largest emitter, refused any legally binding quantified target and made clear that its participation depended on that refusal being respected. The developing world, for its part, would accept no quantified obligations of its own. The outcome was a compromise that has been read in different ways ever since. Article 4.2 of the Convention commits developed country Parties, those listed in #Annex_I, to adopt national policies and take corresponding measures on mitigation, with the stated aim of returning emissions to earlier levels, and it contains a widely noted reference to the year 2000 and to 1990 levels. But the linguistic construction is deliberately indirect. The aim is recited as an object of the policies rather than imposed as a hard result obligation on each Party. The resulting provision is, in the honest assessment of the scholarship, an obligation of conduct dressed in the vocabulary of a target. This was not an accident or a drafting failure. It was the price of American signature and ratification, and American ratification, achieved by unanimous consent of the Senate in October 1992, was the price of the treaty mattering at all. It is worth pausing on this, because it is the pattern that repeats. Again and again over three decades, the climate regime has faced a choice between a stronger instrument with fewer significant participants and a weaker instrument with nearly all of them, and it has, with only one major exception, chosen breadth. The exception was Kyoto, and the consequences of that exception are instructive. 4. The Legal Architecture of the Convention 4.1 The objective: Article 2 Article 2 is the hinge of the entire regime. It sets the ultimate objective as the stabilisation of greenhouse gas concentrations in the atmosphere at a level that would prevent #dangerous_anthropogenic_interference with the climate system, and it adds that such a level should be achieved within a time frame sufficient to allow ecosystems to adapt naturally, to ensure that food production is not threatened, and to enable economic development to proceed in a sustainable manner. Three features of this formulation deserve attention. First, it is framed in terms of concentrations, not emissions or temperature. This is scientifically coherent but politically awkward, because concentrations are a global stock that no single state controls, whereas emissions are a flow that states can be held to account for. Second, the operative concept, dangerous interference, is left undefined. This was a conscious deferral: the negotiators could agree that danger should be avoided without agreeing on where danger begins, and they left the specification of the threshold to future science and future politics. The subsequent history of the regime can be read as a slow, contested, and still incomplete process of filling in that blank, first with the two degree formulation that gained currency after 2009, and then with the more demanding 1.5 degree ambition that the Paris Agreement records and that the Court in 2025 read as the operative #temperature_goal. Third, the article explicitly qualifies environmental protection with development and food security, embedding #sustainable_development in the treaty's very objective rather than treating it as a competing consideration. 4.2 The principles: Article 3 Article 3 sets out the principles that are to guide Parties in achieving the objective and implementing the Convention. Legally, principles of this kind are interpretive rather than directly obligation creating, but in the climate regime they function as something closer to constitutional norms: they structure what arguments are admissible and what outcomes are defensible. The first and most consequential is #common_but_differentiated_responsibilities and respective capabilities. The formulation contains two distinct ideas that are frequently conflated. Responsibility looks backward, to the contribution a state has made to the problem. Capability looks at present circumstances, to the resources a state can bring to bear. The Convention's operationalisation of the principle, through the division of Parties into Annex I and #non_Annex_I categories fixed at the moment of drafting, effectively froze a snapshot of the 1992 world economy into treaty structure. That freezing became the central legal and political controversy of the next twenty three years, because the world economy did not stand still: several non Annex I economies grew into major emitters and major economies, and several Annex I economies stagnated. Recent scholarship has cautioned against a simplistic reading of this history, noting that the early regime in fact contained multiple layers of differentiation beyond the binary annex structure, but the binary was nonetheless what dominated political argument. The second principle is that of #equity, which appears alongside differentiation and gives it moral content. The third is the #precautionary_principle, expressed in Article 3.3 in a form that has become standard: a lack of full scientific certainty should not be used as a reason for postponing measures, where there are threats of serious or irreversible damage, and policies should be cost effective. The fourth is the right to, and duty to promote, sustainable development. The fifth concerns the promotion of a supportive and open international economic system, a provision inserted at the insistence of states anxious that climate policy could become a vehicle for disguised protectionism, and a provision that has returned to prominence in the 2020s as carbon border measures have proliferated and as the trade and climate interface has forced its way onto the COP agenda. 4.3 The commitments: Article 4 Article 4 distinguishes between commitments that apply to all Parties and commitments that apply only to specified groups. All Parties are required, among other things, to develop, periodically update and publish national inventories of anthropogenic emissions by sources and removals by sinks; to formulate and implement national and, where appropriate, regional programmes containing measures to mitigate climate change and to facilitate adequate #adaptation; to promote and cooperate in the development and diffusion of technologies; to promote sustainable management of sinks and reservoirs; to cooperate in preparing for adaptation; to take climate considerations into account in relevant social, economic and environmental policies; to cooperate in research and systematic observation; to promote education, training and public awareness; and to communicate information related to implementation. The inventory obligation deserves special notice: in 2025 the International Court of Justice explicitly identified it as a legally binding obligation, which is a reminder that even the general commitments of the framework treaty carry legal force. Annex I Parties, the industrialised states and the economies in transition, carry the additional mitigation commitments of Article 4.2 discussed above, together with more demanding reporting obligations. A further subset, listed in Annex II and comprising the members of the OECD as of 1992 but excluding the transition economies, carries the financial commitments: to provide new and additional financial resources to meet the agreed full costs incurred by developing country Parties in complying with their reporting obligations, to provide resources including for #technology_transfer needed to meet the agreed full incremental costs of implementing measures, and to assist particularly vulnerable developing country Parties in meeting the costs of adaptation. The vocabulary of new and additional, and of agreed full incremental costs, has generated three decades of accounting disputes that remain unresolved. Article 4.7 states, in terms that developing countries have invoked in every negotiation since, that the extent to which developing country Parties will effectively implement their commitments will depend on the effective implementation by developed country Parties of their commitments related to financial resources and technology transfer, and that economic and social development and poverty eradication are the first and overriding priorities of developing country Parties. This is the conditionality clause, and it is the legal anchor of the #climate_finance debate. 4.4 The legal character of the instrument It is a mistake to conclude from the softness of Article 4.2 that the Convention creates no meaningful legal obligations. It creates a large number of them, but they are predominantly obligations of conduct and of procedure rather than obligations of result. Parties must report, must plan, must cooperate, must take climate change into account. What they need not do, under the 1992 text standing alone, is achieve any particular emissions outcome by any particular date. This is the characteristic structure of a framework convention, and it is a structure with a theory behind it. The theory holds that where a problem is scientifically uncertain, economically pervasive and politically explosive, the first task is not to solve it but to institutionalise it: to establish that it is a legitimate subject of international regulation, to create a permanent venue, to standardise information, and to build the habits of cooperation from which stronger rules can later emerge. Judged against that theory, and not against the theory of an emissions cap, the 1992 treaty performed well. Its architects understood exactly what they were building. The critique that it failed to reduce emissions is a critique of the world, not of the drafters. 5. The Institutional Machinery 5.1 The Conference of the Parties The Convention establishes the Conference of the Parties as its supreme body, charged with keeping implementation under regular review and making the decisions necessary to promote effective implementation. In practice the COP is the beating heart of the regime, and its annual meetings have become the single most visible ritual of global environmental politics, drawing tens of thousands of delegates, observers, journalists, business representatives and campaigners. The COP takes its decisions by consensus. This is one of the most consequential facts about the entire regime and one of the least well understood. The Convention's draft rules of procedure were never formally adopted precisely because Parties could not agree on a voting rule, and the process has operated ever since by applying the draft rules provisionally, with the voting rule bracketed. The practical consequence is that #consensus_decision_making prevails: a decision is adopted when no Party formally objects. This gives every Party, including the smallest and the most obstructive, an effective veto, and it slows the regime to the pace of its most reluctant participant. It also, and this is the counterargument that defenders of the rule advance, gives every Party a stake in the outcome and denies any coalition the ability to impose rules on an unwilling minority, which in a regime whose whole value lies in universality is not a trivial consideration. Presiding officers have developed a repertoire of techniques for managing the rule, including gavelling decisions through over stated objections, and these techniques have themselves become a source of contestation about the regime's #legitimacy. 5.2 Subsidiary bodies and the secretariat Two permanent subsidiary bodies serve the COP. The Subsidiary Body for Scientific and Technological Advice provides assessments of the state of scientific knowledge and advises on methodological questions, functioning as the interface between the IPCC and the negotiations. The Subsidiary Body for Implementation assists in the assessment and review of implementation, and handles a large share of the regime's practical business, including reporting, finance arrangements and administrative matters. Both meet twice a year, once in Bonn in the middle of the year and once alongside the COP, and it is in these sessions rather than in the glare of the summits that most technical work actually gets done. The Convention also establishes a secretariat, headquartered in Bonn. Its formal functions are modest and largely administrative, but its real influence has been considerable. A permanent bureaucracy with institutional memory, drafting capacity and convening power is an asset in any regime, and the climate secretariat has used those assets to shape agendas, to broker text, to build relationships with #non_state_actors and to sustain continuity across presidencies that rotate annually. The influence of successive executive secretaries on the trajectory of the process is a recognised theme in the literature, and former officials of the secretariat have themselves contributed some of the most candid assessments of what the multilateral process can and cannot deliver. 5.3 Reporting and review The Convention requires all Parties to submit #national_communications and inventories, with the frequency and detail differentiated by category. This system, unglamorous and technically demanding, is arguably the treaty's most underrated achievement. Before 1992 there was no standardised, internationally comparable, regularly updated global accounting of greenhouse gas emissions. There now is. Every subsequent instrument, including the Paris Agreement's #enhanced_transparency_framework and its #biennial_transparency_reports, is an elaboration of the reporting infrastructure that the Convention created. Without common accounting rules, targets are meaningless, markets are impossible, and #accountability is rhetorical. The Convention built the measuring apparatus first, which in hindsight was the correct order of operations. 5.4 The financial mechanism Article 11 establishes a mechanism for the provision of financial resources on a grant or concessional basis, including for technology transfer, functioning under the guidance of and accountable to the COP. Its operation was entrusted on an interim basis to the #Global_Environment_Facility, an arrangement that became permanent in practice and that has been a persistent source of Southern complaint, principally because the Facility's governance is weighted toward donors in a way that Parties from the developing world regard as inconsistent with COP accountability. Over time the financial architecture has multiplied: the Special Climate Change Fund and the Least Developed Countries Fund were created in 2001, the Adaptation Fund under the Kyoto Protocol, the #Green_Climate_Fund at Cancun in 2010, and most recently the Fund for responding to Loss and Damage. The result is an architecture that is comprehensive on paper and fragmented in practice, and the fragmentation itself imposes real access costs on the countries least equipped to bear them. The empirical literature on climate finance policy has repeatedly found that the effectiveness of support depends less on headline volumes than on predictability, concessionality and the ability of recipients to actually access it. 6. From Framework to Protocol: The Kyoto Era 6.1 Berlin, Kyoto and the logic of targets The Convention's first COP, held in Berlin in 1995, concluded that the Article 4.2 commitments were inadequate and launched a negotiating mandate for a protocol containing quantified emission limitation and reduction objectives for Annex I Parties, explicitly excluding any new commitments for developing countries. The Berlin Mandate is the moment at which the regime's differentiation became rigid, and it is the moment that determined the fate of the protocol that followed. The #Kyoto_Protocol, adopted in December 1997, established legally binding quantified targets for Annex I Parties, averaging a reduction of about five per cent below 1990 levels over the commitment period 2008 to 2012, with individual targets varying by country. It introduced three flexibility mechanisms, emissions trading between Annex I Parties, joint implementation, and the Clean Development Mechanism, which allowed Annex I Parties to earn credits from emission reduction projects in developing countries. These mechanisms were the intellectual ancestor of every subsequent debate about #carbon_markets, and the Clean Development Mechanism in particular generated a body of practical experience, much of it cautionary, about additionality, baselines and environmental integrity that has directly informed the design of #Article_6 of the Paris Agreement. Kyoto also created something the Convention conspicuously lacked: a #compliance system with teeth, at least on paper, including an enforcement branch empowered to impose consequences on Parties that exceeded their assigned amounts. That such a system could be agreed at all, in the #Marrakesh_Accords of 2001 that constituted the Kyoto rulebook, is a remarkable fact about the era. 6.2 What Kyoto actually was, and was not It has become conventional to describe Kyoto as a top down treaty in contrast to a bottom up Paris Agreement, and to attribute Kyoto's difficulties to its top down character. Recent scholarship has subjected this framing to sustained criticism and found it wanting. The Kyoto targets were not imposed on governments by any supranational authority; they were the product of intergovernmental bargaining in which each state negotiated its own number, exactly as Parties later determined their own contributions under Paris. The genuinely distinctive features of Kyoto were that the resulting targets were internationally binding, that they were subject to a compliance mechanism, and that they applied only to one group of countries. The top down and bottom up metaphor obscures more than it reveals, and the case for abandoning it is strong. The real problem with Kyoto was coverage. The United States signed but never ratified. Canada withdrew. The protocol's second commitment period, agreed at Doha in 2012, covered a shrinking share of a growing global total. By the time Kyoto's first period ended, the Parties with binding targets accounted for a minority of world emissions, and the fastest growth was occurring in countries that had none. A treaty can be stringent or it can be comprehensive; Kyoto tried to be stringent and discovered the cost. 6.3 Bali, Copenhagen and the crisis of 2009 The #Bali_Road_Map of 2007 launched a two track negotiation intended to produce an agreed outcome by 2009, encompassing both a second Kyoto period and long term cooperative action under the Convention, including for the first time the notion of nationally appropriate mitigation actions by developing countries. The destination was Copenhagen, and Copenhagen was, by common consent, a failure of process even if it was not entirely a failure of substance. The conference produced no legally binding instrument. It produced instead the #Copenhagen_Accord, a short political document negotiated by a small group of heads of state and government in the final hours, which the COP was unable to adopt and could only take note of. The Accord nonetheless contained the seeds of everything that followed: a recognition of the two degree limit, a system in which countries would inscribe their own pledges in appendices rather than have targets assigned to them, a commitment by developed countries to mobilise one hundred billion dollars a year by 2020, and provision for international consultation and analysis of developing country actions. Copenhagen's importance lies in what it taught. It demonstrated that a universal, legally binding, top down allocation of emission entitlements was not politically attainable, and that attempting it risked destroying the process. It also demonstrated the reputational cost of exclusionary, small group bargaining in a regime whose authority rests on inclusiveness. The #Cancun_Agreements of 2010 rescued the process by bringing the substance of the Accord formally inside the Convention, and the #Durban_Platform of 2011 set the mandate for a new instrument applicable to all Parties, to be adopted by 2015. The word all, in that mandate, marked the end of the Berlin Mandate's rigid firewall and the beginning of the road to Paris. 7. The Paris Agreement and the Turn to Pledge and Review 7.1 Design The #Paris_Agreement, adopted in December 2015 under the Convention and entering into force with unusual speed in November 2016, resolved the Copenhagen dilemma by relocating stringency. Instead of negotiating targets internationally and binding them legally, Paris requires each Party to prepare, communicate and maintain successive #nationally_determined_contributions, and to pursue domestic measures with the aim of achieving them. The obligation to submit and to pursue is binding. The content of the contribution is not. Successive contributions are to represent a progression and to reflect the Party's highest possible ambition. Around this core the Agreement builds a system of oversight designed to generate ambition over time through transparency and peer pressure rather than through enforcement. Article 13 establishes the enhanced transparency framework, with common reporting requirements and technical expert review for all Parties, with flexibility for those developing countries that need it in light of their capacities. Article 14 establishes the global stocktake, a five yearly collective assessment of progress toward the Agreement's purpose and long term goals, conducted in the light of equity and the best available science, whose outcome is to inform Parties in updating their contributions. This is the #ambition_cycle, and its logic is the logic of #pledge_and_review: no Party is told what to do, but every Party's performance is measured, published, compared and collectively assessed, and the expectation is that this creates a #ratchet_mechanism that turns only in one direction. Article 2 sets the purpose: holding the increase in global average temperature to well below two degrees above pre industrial levels and pursuing efforts to limit it to 1.5 degrees, increasing the ability to adapt, and making finance flows consistent with a pathway toward low emissions and climate resilient development. Article 7 establishes a #global_goal_on_adaptation. Article 8 addresses #loss_and_damage, though the accompanying COP decision records that it does not involve or provide a basis for liability or compensation, a caveat that has shaped every subsequent discussion of the subject. Article 9 sets out finance obligations, with developed countries to provide and other Parties encouraged to do so voluntarily. Article 6 provides for voluntary cooperative approaches, including a mechanism for internationally transferred mitigation outcomes and a centralised crediting mechanism. Article 15 establishes a compliance committee that is expressly facilitative and non punitive. 7.2 What Paris changed and what it did not Paris changed the political economy of the regime in three ways. It made participation nearly universal by making the substance self determined. It shifted the regime's coercive centre of gravity from international enforcement to domestic accountability, since a contribution that a government has publicly committed to becomes a domestic political fact that opposition parties, courts, investors and civil society can hold it to. And it gave the world a focal number, 1.5 degrees, which has proved to be an extraordinarily effective mobilising device, more effective than its drafters can have anticipated. What Paris did not change is the arithmetic. The contributions submitted in the first round were, and were known at the time to be, collectively inconsistent with the temperature goal. Assessments of the aggregate effect of pledges have consistently found a substantial #emissions_gap between the trajectory implied by national commitments and the trajectory required for the stated goals, although careful modelling has also found that full implementation of announced pledges, including long term net zero targets, would represent a very significant improvement on the pre Paris trajectory and could bring projected warming close to the two degree mark. Both findings are true, and holding them together is the beginning of an honest assessment: the regime has bent the curve without breaking it, and the wave of #net_zero targets that spread through the late 2010s and early 2020s opened a window that has not yet been closed and has not yet been walked through. 7.3 The rulebook Paris was a treaty with most of its operational detail deferred. That detail, the Paris rulebook, was substantially completed at Katowice in 2018 and finished at Glasgow in 2021, with the guidance on Article 6 finally settled in stages through 2024. The rulebook is where the abstract commitments of the Agreement acquired operational meaning: what information Parties must provide with their contributions, how the transparency reports are structured and reviewed, how the stocktake is conducted, and how cooperative approaches avoid double counting. It is technical, unloved and indispensable, and the balance it struck between international prescriptiveness and national discretion is the central question in assessing whether Paris can work. 8. Implementation: Transparency, Stocktake, Markets 8.1 The enhanced transparency framework Under the enhanced transparency framework, all Parties, developed and developing alike, submit biennial transparency reports containing a national inventory and information necessary to track progress in implementing and achieving their contributions. The first of these reports fell due at the end of 2024. This represents the effective end of the strict Annex based firewall in the reporting system, though flexibility provisions preserve differentiation in practice. The framework does not punish. It exposes. Its underlying wager is that exposure, in a world of active media, mobilised publics and increasingly climate literate financial markets, is a form of power. Whether that wager pays depends on #capacity_building. Many developing country Parties lack the institutional infrastructure to produce inventories of the required quality, and the regime's own support programmes have been chronically underfunded relative to the demand. A transparency system that shames the poorly resourced for failing to report, while the well resourced report fully and act inadequately, would be a perverse outcome, and avoiding it is a live policy problem rather than a hypothetical one. 8.2 The first global stocktake The first global stocktake ran from 2021 to 2023, comprising an information collection phase, a technical dialogue and a political consideration of outputs. The technical dialogue produced a synthesis report in September 2023 that was notably direct: the world is not on track, the window is narrowing, and much greater ambition is required across mitigation, adaptation and means of implementation. The political outcome, adopted at Dubai at the end of 2023, is the single most quoted COP decision since Paris. It calls on Parties to contribute to global efforts including transitioning away from fossil fuels in energy systems, in a just, orderly and equitable manner, accelerating action in this critical decade so as to achieve net zero by 2050 in keeping with the science; tripling renewable energy capacity globally and doubling the global average annual rate of energy efficiency improvement by 2030; and accelerating efforts toward the phase down of unabated coal power. The phrase transitioning away from fossil fuels was the first explicit reference of its kind in the regime's history, and its inclusion, in a decision adopted by consensus at a conference presided over by a major oil producing state, was a genuine diplomatic achievement. It is also a call, not a command, and it is not allocated to any Party in particular. The #fossil_fuel_transition thus entered the regime's normative vocabulary without entering its obligation structure, which is a characteristic move. The stocktake's design has attracted a substantial scholarly literature concerned with whether a collective, non individualised assessment can actually drive national ambition, and with how the process might be made more effective, including by disaggregating the assessment to the level of individual economic sectors so that the signals it sends are actionable by the actors who must act on them. 8.3 Article 6 and carbon markets After nearly a decade of negotiation, the operational rules for Article 6 were completed at Baku in 2024, covering both the bilateral cooperative approaches of Article 6.2 and the centralised crediting mechanism of Article 6.4. The stakes are high in both directions. Well governed markets can lower the cost of a given quantity of abatement and channel finance to developing countries. Poorly governed markets can manufacture the appearance of mitigation without the substance, and the Clean Development Mechanism's history furnishes ample evidence that this risk is not theoretical. The governance problem is compounded by the coexistence of the compliance market under the Agreement with a large, fragmented and lightly regulated voluntary market, and the interaction between the two is one of the more urgent unresolved questions in the field. The wider evidence base on carbon pricing suggests that such instruments do reduce emissions, but typically by modest amounts, largely because the prices actually adopted have been low and the coverage partial. 9. Finance, Adaptation and Loss and Damage 9.1 The hundred billion and its successor The pledge made at Copenhagen and formalised at Cancun, that developed countries would mobilise one hundred billion dollars a year by 2020 for developing countries, became the most politically charged number in the regime. It was not met on time. Assessments differ on precisely when and whether it was subsequently met, in part because what counts as climate finance is contested, with disputes over whether loans should be counted at face value, whether finance is genuinely new and additional to existing development assistance, and whether over reporting is systemic. The failure to deliver a headline promise, on time and transparently, did lasting damage to trust, and the scholarship on the subject has argued forcefully that the credibility of the entire regime depends on repairing it. At Baku in 2024 the Parties agreed a #New_Collective_Quantified_Goal to replace it. The decision sets a goal of at least three hundred billion dollars per year by 2035 for developing country climate action, with developed countries taking the lead but with the sum drawn from a wide range of public and private, bilateral and multilateral sources, and it calls on all actors to work together to scale finance to developing countries to at least 1.3 trillion dollars per year by 2035. Developing country Parties, many of whom had sought a core provision goal an order of magnitude larger and composed of public grant equivalent funds, reacted with open anger, and several accused the presidency of pushing the text through without proper consent. The episode is a compact illustration of the regime's central tension: a decision that is simultaneously a genuine advance on the previous goal and a bitter disappointment relative to assessed need. 9.2 Adaptation Adaptation was, for most of the regime's history, the poor relation of mitigation. This has changed, though the change is more evident in text than in money. The Cancun Adaptation Framework, the national adaptation plan process, the global goal on adaptation established by Paris and elaborated through the framework adopted at Dubai in 2023 and the indicator work concluded at Belem in 2025, together constitute a real institutional architecture. What has not followed is proportionate finance. Adaptation finance flows remain far below assessed needs, and the goal agreed at Glasgow of doubling adaptation finance by 2025 was not met on the available evidence. At Belem the Parties called for efforts to at least triple adaptation finance by 2035, situating that call within the new collective goal. Whether this represents progress depends on the baseline and the date one measures against, and the analytical commentary has been sceptical that the target is as ambitious as its headline suggests. 9.3 Loss and damage The demand that the regime address harms that are neither prevented by mitigation nor avoided by adaptation dates back to the negotiation of the Convention itself, when small island states proposed an insurance mechanism. It was resisted for decades, principally because industrialised states feared that acknowledging the category would open the door to liability and compensation. The Warsaw International Mechanism was established in 2013, the Santiago Network for technical assistance in 2019, and at Sharm el Sheikh in 2022 the Parties agreed to establish a fund. That fund, the Fund for responding to Loss and Damage, was operationalised at Dubai in 2023, its arrangements were finalised at Baku in 2024, and it began moving toward disbursement thereafter. Initial pledges, however, have been measured in hundreds of millions of dollars against estimated needs measured in hundreds of billions. The establishment of the fund is nonetheless a structural change in the regime and not merely an incremental one. It concedes the principle that the harms of climate change are a legitimate object of international financial solidarity, and it establishes an institution through which that solidarity can be organised. The scholarly literature on the third pillar of climate policy has argued that the political battle over recognition is now largely won and that the battle over resourcing has only begun. 10. Assessing the Framework: Effectiveness and Its Measures 10.1 Against what standard? Assessment requires a standard, and the choice of standard determines the verdict. If the standard is the trajectory of global emissions, the regime has failed, since emissions have risen substantially since 1992 and atmospheric concentrations have risen with them. If the standard is whether the regime has produced the outcomes that its own objective requires, it has not yet done so. But a more discriminating literature proposes to assess the regime against the functions that multilateral processes can realistically perform. Former senior officials of the secretariat, reviewing thirty years of negotiation, assessed the process against a set of such functions, including the development of international law, the setting of goals, the establishment of rules and standards, the provision of transparency and accountability, and the mobilisation of support to developing countries, and reached a mixed but not dismissive verdict: the regime has performed well on law, institutions and norms, and poorly on the mobilisation of implementation at the necessary scale. Related work has proposed a similar functional inventory, distinguishing guidance and signalling, rule setting, transparency and accountability, means of implementation, and the mobilisation of non state actors, and has argued that the COP's future utility depends on whether it can move from regime building, a task it has essentially completed, to implementation, a task it has barely begun. 10.2 What the framework has achieved Five achievements can be defended with confidence. First, universality. The Convention is one of the most widely ratified treaties in existence. It established, definitively, that climate change is a matter of international legal concern and not a purely domestic policy question, and that every state has obligations in respect of it. Second, an agreed objective and an agreed set of principles. The regime has a shared normative language. This sounds thin until one considers the alternative, in which each state argues from its own moral premises with no common vocabulary. Third, information infrastructure. The global system of emissions accounting, reporting and review exists because the Convention created it and its progeny extended it. This infrastructure is a precondition for every other form of climate action, including action wholly outside the regime. Fourth, a permanent venue. Annual COPs, twice yearly subsidiary body sessions, a standing secretariat and a dense web of expert bodies mean that climate change is permanently on the international agenda, with predictable moments of decision that focus political attention and around which domestic policy processes, corporate announcements and civil society mobilisation are organised. The convening function is not a side effect; it is a substantial part of the regime's output. Fifth, normative diffusion. The temperature goal, the concept of net zero, the expectation that states will have national climate plans and long term strategies: all of these originated in or were consolidated by the regime and have propagated far beyond it, into domestic legislation, financial regulation, corporate strategy and, increasingly, judicial reasoning. 10.3 Where it has fallen short Three shortfalls are equally clear. The ambition shortfall is the most obvious. Aggregate national contributions remain insufficient for the stated goals, and even the most recent synthesis of contributions indicates a reduction in global emissions by 2035 relative to 2019 that is far short of what the temperature goal requires. The stocktake at Dubai acknowledged this. The Belem decision acknowledged, for the first time, the risk of overshooting 1.5 degrees, and committed Parties to limiting both the magnitude and the duration of that overshoot, which is a candid but sobering formulation. The implementation shortfall follows. Even the contributions that have been made are not always fully implemented, and the regime has no mechanism, and by deliberate design will have no mechanism, to compel implementation. The support shortfall is the most corrosive, because it undermines the conditional bargain on which developing country participation rests. Finance has been late, smaller than promised, weighted toward loans rather than grants, weighted toward mitigation rather than adaptation, and difficult to access. Article 4.7 of the Convention makes developing country implementation expressly contingent on developed country delivery. When delivery falls short, the legal and political basis for expecting Southern ambition weakens correspondingly, and the resulting cycle of mistrust is visible in every negotiating session. 11. Structural Critiques 11.1 Consensus and the pace of the process The consensus rule protects the reluctant. It has been used, at various points, by fossil fuel exporting states to block or dilute language on fossil fuels, and by others to block procedural reforms. Proposals for majority voting on defined categories of decision surface regularly and go nowhere, for the excellent reason that adopting them would itself require consensus. The literature is divided: some argue that the rule is the principal brake on the regime and must be reformed; others argue that its removal would simply drive dissenting states into non participation, converting a slow universal regime into a fast partial one, which is precisely the trade that Kyoto made and lost. There is no way to test the counterfactual. 11.2 Sovereignty and the limits of soft law The regime cannot compel. Withdrawal is permitted, and has been exercised. The United States withdrew from the Paris Agreement, rejoined, and withdrew again, and the possibility of a further departure from the Convention itself has been raised in policy debate. A regime whose central obligations are procedural and whose enforcement is facilitative is structurally vulnerable to the electoral cycles of its most powerful members. Whether the design could have been otherwise, given the constitutional realities of the states involved, is doubtful; but the vulnerability is real, and it means that the regime's continuity depends on it having built enough institutional and economic momentum to survive the defection of any single Party, however large. 11.3 Equity and the differentiation dispute The differentiation question is unresolved and probably unresolvable in the abstract. The Annex structure of 1992 no longer maps onto the economic and emissions geography of the 2020s. Any attempt to revise it directly is met with the entirely reasonable objection that responsibility for the accumulated stock of atmospheric carbon has not changed merely because current flows have, and that per capita emissions and per capita wealth remain grossly unequal. Paris finessed the question rather than settling it, replacing categorical differentiation with self differentiation. Work on fair shares, which attempts to derive national obligations from the principles of international environmental law rather than from political bargaining, has become one of the most active areas of scholarship, and it has acquired new practical significance now that courts are being asked to assess whether a given national target is consistent with a state's obligations. #climate_justice has moved from the margins of the discourse to a place where it is beginning to have legal consequences. 11.4 Participation, capture and the COP as spectacle The COP has grown enormously. Attendance now runs to the tens of thousands, and a substantial share of participants represent industry, including sectors whose commercial interests are directly at stake in the outcomes. Concerns about the influence of vested interests and about the proportionality of fossil fuel industry representation are widely voiced, and proposals for conflict of interest rules have been formally raised. At the same time, the openness of the process to observers, including environmental organisations, indigenous peoples' organisations, research institutions and subnational governments, is one of its genuine democratic strengths, and the growth of #non_state_actors within and around the process has been studied as a significant development in the authority structure of global governance. The reconciliation of openness with integrity is a live institutional design problem, not a rhetorical one. 11.5 The equity of vulnerability Finally, there is the persistent structural asymmetry between the states that caused the problem and the states that bear its worst consequences. #small_island_developing_states and #least_developed_countries have negotiating capacity that is minuscule relative to the largest delegations, must cover an agenda of extraordinary technical breadth with a handful of officials, and are asked to accept outcomes that fall far short of what their own survival requires. The literature on coalitions in the negotiations documents how these states have partially compensated through alliance building, narrative framing and moral authority, and how coalition dynamics have become steadily more complex as the traditional blocs have fragmented. 12. The Legalisation Turn: Courts Enter the Regime Perhaps the most significant development of the 2020s is that the interpretation of the climate treaties has begun to move, in part, from negotiating rooms to courtrooms. Domestic climate litigation has grown rapidly in volume and in ambition, with cases brought against governments for inadequate targets and against companies for their contribution to harm. The tracking literature records a steady expansion in the number and geographic spread of cases, and a growing proportion grounded in human rights arguments. This litigation frequently invokes the Convention and the Paris Agreement, not as directly enforceable domestic law, but as evidence of the standard of conduct that a state has accepted and against which its actions may be measured. In this way the regime's soft obligations acquire hard consequences through domestic legal systems, and the Paris Agreement exercises a power in litigation that its own compliance mechanism was designed not to exercise. The culmination, to date, came on 23 July 2025, when the International Court of Justice delivered its unanimous advisory opinion on the obligations of states in respect of climate change, requested by the General Assembly following a campaign initiated by law students in the Pacific. The Court held that the climate treaties, the Convention, the Kyoto Protocol and the Paris Agreement, are complementary and mutually reinforcing and establish binding legal obligations. It held that the inventory obligation under Article 4 of the Convention is legally binding. It rejected the argument, advanced by a number of major emitters, that the climate treaties constitute a self contained special regime displacing general international law, holding instead that states are also bound by the customary duty to prevent significant transboundary and environmental harm, by duties of cooperation, and by obligations arising under human rights law. It held that the discretion of Parties in preparing their contributions under the Paris Agreement is limited rather than absolute, and that the standard of due diligence applicable to the prevention of climate harm is stringent. It treated the 1.5 degree limit as the operative benchmark. And it indicated that failure to act may constitute an internationally wrongful act, engaging the law of state responsibility. An advisory opinion is not binding in the manner of a judgment between parties. Its practical effect will depend on how national courts, negotiators and litigants use it. But it is a striking development in the life of a regime that was constructed precisely to avoid liability, and its significance for the interpretation of the 1992 Convention is difficult to overstate: the framework treaty's general commitments, long treated as aspirational scaffolding, have been read by the principal judicial organ of the United Nations as legal obligations with content. 13. The Contemporary Negotiation Landscape: Dubai, Baku, Belem The three most recent conferences illustrate both the regime's capacity and its constraints. At Dubai in 2023, the first global stocktake concluded with the fossil fuel transition language described above, together with the goals of tripling renewable capacity and doubling efficiency improvement rates, and with the operationalisation of the loss and damage fund on the conference's opening day. It was, by the standards of the process, a substantial outcome. At Baku in 2024, the Parties adopted the new collective finance goal and completed the Article 6 rules, but failed to agree how the fossil fuel commitments of the stocktake should be carried forward, deferring the question. The conference is widely regarded as having delivered its two headline items while leaving the political core of the mitigation agenda untouched, and it left a residue of resentment among developing country Parties over both the quantum of the finance goal and the manner of its adoption. At Belem in 2025, the Parties adopted the #Global_Mutirao decision, named for a Brazilian term denoting collective effort. The package included a call to at least triple adaptation finance by 2035 within the framework of the new collective goal, the adoption of a reduced set of indicators for the global goal on adaptation, the creation of a mechanism on #just_transition, a mission to strengthen national contributions, an acknowledgement of the road map toward 1.3 trillion dollars a year in climate finance developed with the preceding presidency, and a first formal acknowledgement of the risk of temperature overshoot together with a commitment to limit its magnitude and duration. What the package did not include was a road map for the transition away from fossil fuels; that proposal was removed from the negotiated text, and was instead taken up by a voluntary coalition of more than eighty countries outside the formal decision, with a separate international conference on the subject announced for 2026. The Belem outcome is therefore a precise miniature of the whole regime. Consensus was preserved. Institutions were extended. Finance was promised on a horizon distant enough to be politically survivable. And the single measure that would most directly address the physical cause of the problem was displaced from the binding text into a coalition of the willing operating alongside it. Whether such coalitions strengthen the multilateral core by pioneering commitments that later migrate into it, or weaken it by allowing states to substitute voluntary declarations for negotiated obligation, is one of the sharpest open questions in the field, and the debate over climate clubs speaks directly to it. 14. Reform Debates and Future Directions Several reform agendas are live in the scholarship and, increasingly, in the process itself. Sectoralisation. One influential strand argues that the COP should organise a larger share of its work around specific economic sectors, energy, steel, cement, transport, agriculture, rather than around aggregate national emissions, on the grounds that the actors who must actually change behaviour operate within sectors, that sectoral targets and road maps generate actionable signals, and that the stocktake in particular would be far more useful if it assessed progress sector by sector. The counterargument is that sectoralisation risks fragmenting the regime and diluting the national accountability that the transparency system is designed to produce. Implementation focus. A related argument holds that the COP has completed the regime building phase and must now reorganise itself around implementation, shifting from the negotiation of texts to the facilitation of delivery, with corresponding changes to agendas, formats and the role of non state participants. Proposals for reform of the meetings themselves, including changing their frequency, size and structure, recur. Institutional reform of the process. Proposals include revisiting the consensus rule for defined categories of decision, adopting conflict of interest rules, imposing eligibility criteria on host states, streamlining an agenda that has grown unwieldy, and strengthening the mandate of the secretariat. Each faces the same obstacle: reform of the process requires the agreement of the process. Finance architecture. The proliferation of funds has created an access problem, and reform proposals focus on simplification, on direct access for vulnerable states, on the grant equivalent share of support, on reform of the multilateral development banks, and on innovative sources including levies on aviation, shipping and fossil fuel profits, and on debt for climate instruments. The Belem package gestured at several of these. Hardening the ratchet. If the ambition cycle is the regime's central mechanism, its weakness is that nothing happens when a Party's contribution is inadequate. Proposals to strengthen it range from developing an operational standard for what highest possible ambition requires, so that inadequacy can be identified against a benchmark rather than merely asserted, to using the stocktake to make individualised rather than purely collective findings. The advisory opinion of 2025, by holding that discretion in preparing contributions is limited and that due diligence is stringent, supplies exactly the kind of external benchmark that such proposals require, and its interaction with the regime's internal processes is likely to be the defining institutional story of the late 2020s. 15. Implications for Study and Research For students, five analytical habits follow from the foregoing. First, read the text. A great deal of confident commentary about the climate regime dissolves on contact with the actual wording of Articles 2, 3 and 4 of the Convention and Articles 2, 4, 6, 13 and 14 of the Paris Agreement. The distinction between an obligation of conduct and an obligation of result is not pedantry; it is the whole architecture. Second, distinguish the treaty from the decisions. Most of what people call the climate regime consists of decisions of the COP and of the Paris Agreement's meeting of the Parties, which are not treaties, whose legal status is contested, and which are nonetheless where the operational content lives. Third, resist the standard narratives. The top down Kyoto and bottom up Paris story is, on the evidence, misleading. So is the claim that the Convention did nothing, and so is the claim that Paris solved anything. Good analysis in this field is almost always an exercise in resisting a tidy story. Fourth, follow the money and the accounting rules. Finance disputes and methodological disputes look technical and are in fact where the politics is transacted. Whoever writes the accounting rules writes the outcome. Fifth, watch the courts. The most interesting developments in the interpretation of the 1992 Convention in the last five years have come from judicial bodies, not from negotiators. For researchers, several questions are conspicuously open. What is the causal effect, if any, of the transparency framework on national ambition, and how would one identify it? Under what conditions do voluntary coalitions strengthen rather than substitute for multilateral obligation? Does the individualisation of the stocktake improve or degrade the political feasibility of the ambition cycle? How does the advisory opinion propagate into domestic legal systems, and does it change the behaviour of states or merely the vocabulary of their litigation? What governance arrangements would make the Article 6 markets deliver real, additional mitigation at scale, and what is the appropriate relationship between the compliance and voluntary markets? And, most fundamentally, what institutional design would allow a consensus based universal regime to deliver outcomes at a pace set by physics rather than by its most reluctant member? 16. Conclusion The United Nations Framework Convention on Climate Change is best understood not as a solution but as a constitution. It did not, and was never designed to, reduce emissions by itself. It defined the problem in terms that nearly every state on earth accepted. It established an objective, a set of principles for allocating effort, an information system, a financial mechanism and a permanent forum. It made possible the Kyoto Protocol, and when Kyoto's architecture proved unsustainable it made possible the Paris Agreement, and when Paris required elaboration it supplied the machinery through which the rulebook, the transparency framework, the stocktake and the loss and damage fund were built. Thirty three years after Rio, the institution it created remains the only forum in which every state on earth is legally engaged with the problem, and the normative vocabulary it established, from #dangerous_anthropogenic_interference to differentiated responsibility, still structures how the world argues about the subject. The indictment against it is nonetheless serious and must be stated plainly. Emissions are higher than in 1992. Concentrations are higher. Warming has continued, and the year 2024 was the warmest on record and the first calendar year to exceed 1.5 degrees above pre industrial levels, even if a single year above the threshold is not the same thing as breaching the long term goal. Promised finance has been late and short. The most consequential measure, a managed and equitable exit from fossil fuels, remains outside the binding text. Three decades of impeccable procedure have not yet produced a physical result commensurate with the objective set out in Article 2. Both of these things are true, and the honest conclusion holds them together. The framework's institutions are real and they work; the political will to use them at the necessary scale has not yet materialised. What is changing is where the pressure to supply that will comes from. For thirty years, ambition was expected to emerge from negotiation. It is now emerging, in part, from courts applying the very obligations that the negotiators wrote, from investors pricing the transition, from technologies that have become cheaper than the incumbents, and from domestic publics who increasingly experience the problem directly rather than reading about it in an assessment report. The Convention's ultimate contribution may prove to be that it wrote down, in 1992, the commitments that everyone assumed were merely aspirational, and thereby created the legal material from which, decades later, others would build something with teeth. Whether that proves sufficient, and sufficient in time, is the open question of our era, and it is not one that #multilateralism will answer alone. Hashtags #UNFCCC_1992 #Climate_Change_Convention #Rio_1992 #Global_Climate_Governance #International_Environmental_Law #Climate_Diplomacy #Paris_Agreement_2015 #Kyoto_Protocol_1997 #COP_Negotiations #Climate_Policy_Studies #Environmental_Treaty_Law #Climate_Finance_Debate #Loss_And_Damage_Fund #Climate_Justice_Research #Sustainable_Development_Goals References Ahonen, H. M., Kessler, J., Michaelowa, A., Espelage, A., and Hoch, S. (2022). Governance of fragmented compliance and voluntary carbon markets under the Paris Agreement. Politics and Governance, 10(1), 235 to 245. Allan, J. I. (2021). The New Climate Activism: NGO Authority and Participation in Climate Change Governance. Toronto: University of Toronto Press. Aykut, S. C., and Maertens, L. (2021). The climatization of global politics: introduction to the special issue. International Politics, 58(4), 501 to 518. Backstrand, K., Kuyper, J., and Nasiritousi, N. (2021). From collaboration to contestation? Perceptions of legitimacy and effectiveness in post Paris climate governance. Earth System Governance, 9, 100115. Bhandary, R. R., Gallagher, K. S., and Zhang, F. (2021). Climate finance policy in practice: a review of the evidence. Climate Policy, 21(4), 529 to 545. Broberg, M., and Martinez Romera, B. (Eds.). (2021). The Third Pillar of International Climate Change Policy: On Loss and Damage. Abingdon: Routledge. Depledge, J. (2022). The top down Kyoto Protocol? Exploring caricature and misrepresentation in literature on global climate change governance. International Environmental Agreements: Politics, Law and Economics, 22(4), 673 to 692. doi: 10.1007/s10784-022-09580-9 Dubash, N. K. (2021). Varieties of climate governance: the emergence and functioning of climate institutions. Environmental Politics, 30(sup1), 1 to 25. Falkner, R., and Buzan, B. (Eds.). (2022). Great Powers, Climate Change, and Global Environmental Responsibilities. Oxford: Oxford University Press. Falkner, R., Nasiritousi, N., and Reischl, G. (2022). Climate clubs: politically feasible and desirable? Climate Policy, 22(4), 480 to 487. Green, J. F. (2021). Does carbon pricing reduce emissions? A review of ex post analyses. Environmental Research Letters, 16(4), 043004. Hale, T. (2024). Long Problems: Climate Change and the Challenge of Governing across Time. Princeton: Princeton University Press. Hermwille, L., Dzebo, A., Iacobuta, G., and Obergassel, W. (2023). Global stocktake and the SDG mid term review as opportunities for integration. Nature Climate Change, 13, 1002 to 1004. Hohne, N., Gidden, M. J., den Elzen, M., Hans, F., Fyson, C., Geiges, A., Jeffery, M. L., Gonzales Zuniga, S., Mooldijk, S., Hare, W., and Rogelj, J. (2021). Wave of net zero emission targets opens window to meeting the Paris Agreement. Nature Climate Change, 11, 820 to 822. Intergovernmental Panel on Climate Change. (2022). Climate Change 2022: Mitigation of Climate Change. Contribution of Working Group III to the Sixth Assessment Report. Cambridge: Cambridge University Press. Intergovernmental Panel on Climate Change. (2023). Climate Change 2023: Synthesis Report. Contribution of Working Groups I, II and III to the Sixth Assessment Report. Geneva: IPCC. International Court of Justice. (2025). Obligations of States in respect of Climate Change. Advisory Opinion of 23 July 2025. The Hague: International Court of Justice. Kinley, R., Zammit Cutajar, M., de Boer, Y., and Figueres, C. (2021). Beyond good intentions, to urgent action: former UNFCCC leaders take stock of thirty years of international climate change negotiations. Climate Policy, 21(5), 593 to 603. doi: 10.1080/14693062.2020.1860567 Klock, C., Castro, P., Weiler, F., and Blaxekjaer, L. O. (Eds.). (2021). Coalitions in the Climate Change Negotiations. Abingdon: Routledge. Meinshausen, M., Lewis, J., McGlade, C., Gutschow, J., Nicholls, Z., Burdon, R., Cozzi, L., and Hackmann, B. (2022). Realization of Paris Agreement pledges may limit warming just below 2 degrees Celsius. Nature, 604, 304 to 309. Obergassel, W., Bauer, S., Hermwille, L., Aykut, S. C., Boran, I., Chan, S., Fraude, C., Klein, R. J. T., Mar, K. A., Schroeder, H., and Simeonova, K. (2022). From regime building to implementation: harnessing the UN climate conferences to drive climate action. WIREs Climate Change, 13(6), e797. doi: 10.1002/wcc.797 Obergassel, W., Lah, O., and Rudolph, F. (2021). Driving towards transformation? To what extent does global climate governance promote decarbonisation of land transport? Earth System Governance, 8, 100098. doi: 10.1016/j.esg.2021.100098 Oberthur, S., Hermwille, L., and Rayner, T. (2021). A sectoral perspective on global climate governance: analytical foundation. Earth System Governance, 8, 100104. doi: 10.1016/j.esg.2021.100104 Raiser, K., Cali, B., and Flachsland, C. (2022). Understanding pledge and review: learning from analogies to the Paris Agreement review mechanisms. Climate Policy, 22(6), 711 to 727. doi: 10.1080/14693062.2022.2059436 Rajamani, L., Jeffery, L., Hohne, N., Hans, F., Glass, A., Ganti, G., and Geiges, A. (2021). National fair shares in reducing greenhouse gas emissions within the principled framework of international environmental law. Climate Policy, 21(8), 983 to 1004. Rayner, T., Oberthur, S., and Hermwille, L. (2021). A sectoral perspective on international climate governance: key findings and research priorities. Earth System Governance, 8, 100105. doi: 10.1016/j.esg.2021.100105 Roberts, J. T., Weikmans, R., Robinson, S. A., Ciplet, D., Khan, M., and Falzon, D. (2021). Rebooting a failed promise of climate finance. Nature Climate Change, 11(3), 180 to 182. Savaresi, A., and Setzer, J. (2022). Rights based litigation in the climate emergency: mapping the landscape and new knowledge frontiers. Journal of Human Rights and the Environment, 13(1), 7 to 34. Setzer, J., and Higham, C. (2024). Global Trends in Climate Change Litigation: 2024 Snapshot. London: Grantham Research Institute on Climate Change and the Environment, London School of Economics and Political Science. United Nations Environment Programme. (2024). Emissions Gap Report 2024. Nairobi: United Nations Environment Programme. United Nations Environment Programme. (2024). Adaptation Gap Report 2024. Nairobi: United Nations Environment Programme. United Nations Framework Convention on Climate Change. (2023). Technical Dialogue of the First Global Stocktake: Synthesis Report by the Co Facilitators. Bonn: UNFCCC Secretariat. United Nations Framework Convention on Climate Change. (2025). Nationally Determined Contributions under the Paris Agreement: Synthesis Report by the Secretariat. Bonn: UNFCCC Secretariat. van Asselt, H., Obergassel, W., Hall, C., Hermwille, L., Oberthur, S., Otto, S., Sharmina, M., and Bataille, C. (2023). Leveraging the Global Stocktake for effective sectoral climate governance. PLOS Climate, 2(6), e0000229. doi: 10.1371/journal.pclm.0000229 Voigt, C. (2023). The power of the Paris Agreement in international climate litigation. Review of European, Comparative and International Environmental Law, 32(2), 237 to 249. World Meteorological Organization. (2025). State of the Global Climate 2024. Geneva: World Meteorological Organization.

The two peace conferences held at The Hague in 1899 and 1907 produced the first broad, written, and generally accepted rules on how states may fight one another. Before these conferences, the conduct of hostilities was governed mainly by custom, by military manuals of individual armies, and by scattered bilateral agreements. After them, states possessed a shared text that told them which weapons were forbidden, how captured soldiers must be treated, what an occupying army may and may not do, and which acts placed a belligerent outside the law. This article examines the Hague Conventions as the founding layer of what is now called international humanitarian law, and as the source from which the modern concept of the war crime later grew. It traces the intellectual and diplomatic background of the conferences, analyses the content of the Regulations Respecting the Laws and Customs of War on Land, gives extended attention to the Martens Clause, and follows the path by which treaty rules without penalties became the basis for individual criminal responsibility at Nuremberg and, eventually, before the International Criminal Court. The article also takes the criticisms of the Hague system seriously. The general participation clause, the absence of any enforcement machinery, the failure of the disarmament agenda, the language of civilised nations, and the exclusion of colonised peoples from the protective scope of the rules are all examined. The conclusion argues that the Hague Conventions matter less as a finished code than as the moment when the idea that war has legal limits was written down, accepted by states, and made available for later generations to enforce. Keywords: Hague Conventions; laws of war; war crimes; international humanitarian law; Martens Clause; belligerent occupation; customary international law; individual criminal responsibility 1. Introduction Every legal system needs a starting text. For the rules that govern armed conflict, that text is not a single document but a cluster of treaties adopted at two diplomatic gatherings in the Dutch city of The Hague, the first in 1899 and the second in 1907. Together these instruments are known as the #Hague_Conventions. They are the point at which the customary restraints that soldiers had observed unevenly for centuries were turned into written obligations binding on states as a matter of #international_law. The claim that the Hague instruments were the first formal statements of the #laws_of_war requires care. They were not the first attempt to write down rules of restraint. Religious traditions, codes of chivalry, the writings of jurists such as Grotius and Vattel, and the practice of surrender and ransom had all produced norms long before 1899. Nineteenth century documents such as the #Lieber_Code of 1863, the St Petersburg Declaration of 1868, the #Brussels_Declaration of 1874, and the Oxford Manual of 1880 came closer still. What the Hague conferences added was something these earlier efforts lacked. They produced multilateral treaties, negotiated by the great powers together with smaller states, signed and ratified, and framed as legally binding rather than as advice, aspiration, or the internal instruction of a single army. That combination of generality, formality, and consent is what makes the Hague texts the true beginning of the modern law of armed conflict. The second half of the claim, that the Hague Conventions were also a first statement on #war_crimes, needs even more care. Nowhere in the 1899 or 1907 texts does the phrase war crime appear. Nowhere is a court established to try individuals. Nowhere is a penalty specified for a soldier who kills a prisoner or a commander who orders a town to be sacked. The Hague instruments imposed duties on states, and Article 3 of the 1907 Convention IV made a state liable to pay compensation for violations committed by members of its armed forces. Criminal liability of the individual came later, and it came through practice, through the judgments of tribunals, and through the recognition that the Hague rules had hardened into #customary_international_law. Yet the connection is real, not rhetorical. When the International Military Tribunal at #Nuremberg convicted German defendants of war crimes in 1946, the substantive definitions it applied were drawn overwhelmingly from the Hague Regulations. When the Rome Statute of the International Criminal Court listed war crimes in 1998, a substantial part of Article 8 reproduced, in modern language, prohibitions that were first written at The Hague. In this sense the Hague Conventions supplied the content of the crime long before the international community supplied the court. This article is written for students who need to understand not only what the Hague Conventions say, but why they were written, what they achieved, what they failed to achieve, and why lawyers still cite them more than a century later. The argument advanced here has three parts. First, the Hague conferences represented a genuine break in legal method, because they made the regulation of hostilities a matter of treaty law negotiated among states rather than a matter of custom, honour, or unilateral instruction. Second, the substantive rules adopted at The Hague, especially the Regulations annexed to Convention II of 1899 and Convention IV of 1907, remain the structural core of the law on the conduct of hostilities and of #belligerent_occupation. Third, the weaknesses of the Hague system, and above all its silence on enforcement, are as instructive as its achievements, because the entire subsequent history of #IHL can be read as a series of attempts to repair those weaknesses. 2. Research Aims, Questions, and Method 2.1 Aims This study has four aims. It seeks to reconstruct the historical and doctrinal context in which the Hague Conventions were negotiated. It seeks to analyse their substantive content with enough precision that a reader can identify which modern rules descend from which Hague provisions. It seeks to trace the transformation of Hague obligations from interstate duties into the foundations of #individual_criminal_responsibility. And it seeks to assess the continuing legal relevance of the Hague texts in contemporary conflicts and in debates about new technologies of war. 2.2 Research questions The article is organised around five questions. What legal and political conditions made the codification of the laws of war possible at the end of the nineteenth century, and why did earlier attempts fail? What did the Hague Conventions actually regulate, and what did they deliberately leave unregulated? In what sense, if any, can the Hague instruments be described as the first formal statement of war crimes, given that they contain no criminal provisions? What were the structural defects of the Hague system, and how were those defects exploited or exposed in the wars that followed? Which parts of the Hague framework remain legally operative today, and which have been superseded? 2.3 Method The method is doctrinal and historical. It reads the treaty texts closely, situates them within the diplomatic record of the two conferences, and follows their reception in later treaties, in the jurisprudence of international courts, and in the scholarly literature. The article relies on secondary analysis by leading contemporary writers in the field, and on the case law of the International Military Tribunal, the ad hoc criminal tribunals of the 1990s, the International Court of Justice, and the International Criminal Court. It does not conduct empirical or quantitative research. Its purpose is interpretive: to explain a body of law, to evaluate the standard account of its significance, and to identify where that account requires qualification. Two limitations should be stated at the outset. First, the article concentrates on the law of land warfare, and in particular on Convention II of 1899 and Convention IV of 1907 with their annexed Regulations. The naval conventions of 1907, the rules on neutrality, and the law of maritime blockade are treated more briefly, although they were an important part of the 1907 programme. Second, the article does not attempt a full political history of the peace movement or of the arms control diplomacy of the period. Those subjects are addressed only where they bear on the legal outcome. 3. The Road to The Hague: Restraint Before Codification 3.1 Custom, chivalry, and the limits of unwritten rules For most of recorded history, restraint in war depended on reciprocity, religion, and reputation. Armies spared certain categories of persons because they expected the same treatment for their own, or because a religious authority commanded it, or because the honour of the commander was thought to require it. These constraints were real, and they sometimes saved lives, but they were unstable. They varied by region, by faith, and by the perceived status of the enemy. They usually collapsed when one side regarded the other as barbarian, heretic, or rebel. And because they were unwritten, disputes about their content could not be settled by reference to an agreed text. The philosophical literature of the early modern period, from Vitoria and Suarez to Grotius, Pufendorf, and Vattel, gave these customs a theoretical shape. Writers distinguished between the justice of going to war and the justice of conduct within war, argued about which acts were permitted against an enemy, and debated whether the innocent could ever be targeted deliberately. This tradition supplied the moral vocabulary that the codifiers of the nineteenth century would later borrow. What it did not supply was an authoritative instrument that a government could sign, a soldier could be trained in, and a court could apply. 3.2 The nineteenth century turn to writing Three developments in the nineteenth century changed the picture. The first was industrial warfare. Rifled artillery, breech loading firearms, railways, telegraphy, and mass conscription made war deadlier, faster, and larger. The Crimean War, the American Civil War, the wars of German unification, and the Franco Prussian War produced casualty figures that shocked European publics and made the humanitarian consequences of conflict a matter of political debate. The second was the growth of an organised humanitarian movement. The founding of the International Committee of the Red Cross in the early 1860s and the adoption of the first Geneva Convention in 1864 demonstrated that states could be persuaded to accept binding obligations concerning the treatment of the wounded. The Geneva instrument was narrow, but it was a treaty, and its existence proved that #codification was politically feasible. The third was the professionalisation of military law. Armies increasingly wanted clear instructions for their officers, partly for humanitarian reasons and partly because indiscipline in the field is militarily costly. The Instructions for the Government of Armies of the United States in the Field, drafted by Francis Lieber and issued in 1863 during the American Civil War, was the outstanding example. The Lieber Code was a domestic instrument, binding only on Union forces, but its structure anticipated the Hague Regulations with remarkable closeness. It addressed #military_necessity, the treatment of #prisoners_of_war, the protection of civilians and their property, the status of spies and guerrillas, and the duties of an occupying force. 3.3 St Petersburg, Brussels, and Oxford Three international efforts between 1868 and 1880 prepared the ground directly. The St Petersburg Declaration of 1868 prohibited the use of certain explosive projectiles of small weight. Its preamble is more important than its operative text, because it stated the principle that the only legitimate object of war is to weaken the military forces of the enemy, and that this object is exceeded by weapons that uselessly aggravate the suffering of disabled men. This is the origin of the prohibition on means of warfare causing #unnecessary_suffering, a rule that passed into the Hague Regulations and from there into every subsequent treaty on weapons. The Brussels Conference of 1874, convened at Russian initiative, produced a draft international declaration on the laws and customs of war. It was never ratified. States could not agree on the central political question of the period, which was the legal status of civilians who took up arms against an invading army. Great powers with large armies wanted such persons treated as criminals. Smaller states, fearing invasion, wanted them recognised as lawful #combatants. The deadlock killed the Brussels project, but the text survived as a blueprint. The Oxford Manual of 1880, prepared by the Institute of International Law, restated much of the same material in the form of a model for national legislation and reached the same impasse. The lesson of these failures shaped what happened at The Hague. Codification would only succeed if the hardest question, the status of resistance fighters, could be resolved by a compromise that neither side had to admit was a defeat. That compromise, when it came, took the form of a preambular clause proposed by the Russian delegate Fyodor Martens, and it is discussed at length below. 4. The First Hague Peace Conference of 1899 4.1 Origins and agenda The First Peace Conference was convened on the initiative of Tsar Nicholas II of Russia, through a circular issued by his foreign minister in 1898 and a more detailed programme early in 1899. The Russian motives were mixed. Public documents emphasised the burden of the arms race and the danger that competitive rearmament posed to peace. Contemporary observers, and most historians since, have also noted that Russia was struggling to keep pace with Austrian and German artillery modernisation and had a fiscal interest in slowing the competition. Motive does not determine outcome, and the invitation was accepted. Twenty six states attended, meeting from May to July 1899. The agenda had three main branches. The first was #disarmament, or more precisely a freeze on the growth of armaments and military budgets. The second was the extension of humanitarian rules, including the application of the Geneva principles to war at sea and the revival of the Brussels project on land warfare. The third was the peaceful settlement of disputes through #arbitration and mediation. 4.2 Outcomes On disarmament the conference failed almost completely. No state was willing to bind itself to limits on the size of its army or the growth of its budget, and the proposals were referred to future study, which is the diplomatic form of refusal. On the other two branches the conference succeeded to a degree that surprised many participants. It adopted three conventions and three declarations. Convention I established machinery for the pacific settlement of international disputes and created the #Permanent_Court_of_Arbitration, an institution that still sits at The Hague. Convention II contained the Regulations Respecting the Laws and Customs of War on Land, the heart of the Hague system. Convention III applied the principles of the Geneva Convention of 1864 to maritime warfare. The three declarations prohibited, respectively, the launching of projectiles and explosives from balloons, the use of projectiles whose sole object is the diffusion of asphyxiating or deleterious gases, and the use of bullets that expand or flatten easily in the human body, commonly called dum dum bullets. Two features of this list deserve emphasis for students. First, the conference regulated conduct without abolishing war, and indeed without questioning the legal right of states to resort to war. The prohibition of aggressive war would not arrive until the twentieth century, through the Kellogg Briand Pact and the Charter of the United Nations. The Hague project accepted war as a permitted instrument of policy and sought to civilise its conduct. Second, the ban on weapons was selective and technologically specific. The delegates prohibited the weapons they could imagine and could afford to give up, and they left untouched the weapons they expected to need. 5. The Second Hague Peace Conference of 1907 5.1 A larger and more representative gathering The Second Peace Conference met from June to October 1907. It was originally proposed by the President of the United States and formally convened, once again, by the Russian Emperor. Forty four states attended, including a large delegation of Latin American republics that had been absent in 1899. The broader participation mattered. It meant that the Hague texts could plausibly claim to reflect the practice of a wider community than the European great powers, and it introduced into the debates the concerns of states that feared intervention by stronger neighbours. 5.2 The thirteen conventions The 1907 conference adopted thirteen conventions and one declaration. Convention I revised and expanded the machinery for pacific settlement. Convention II, often associated with the Argentine jurist Luis Drago and known through the Drago Porter formula, restricted the use of armed force to recover contract debts owed to nationals of one state by the government of another, unless the debtor state refused arbitration. Convention III required a formal warning before the opening of hostilities, in the form of a declaration of war or an ultimatum with a conditional declaration of war. #Convention_IV revised the Regulations on land warfare. Convention V dealt with the rights and duties of neutral powers in land war, and Convention XIII with neutrality in naval war. Conventions VI through XII addressed a series of maritime questions, including enemy merchant ships at the outbreak of war, the conversion of merchant ships into warships, the laying of automatic submarine contact mines, bombardment by naval forces, the application of Geneva principles at sea, restrictions on the right of capture, and the establishment of an international prize court, which never came into existence. The single declaration renewed the prohibition on discharging projectiles from balloons. 5.3 What changed in the Regulations The 1907 Regulations on land warfare are, in substance, a revision rather than a replacement of those of 1899. The changes were modest but not trivial. The revised text clarified the treatment of prisoners, refined the rules on bombardment, and strengthened certain provisions on occupied territory. Most significantly, Article 3 of the 1907 Convention IV introduced an obligation of compensation. A belligerent party that violates the Regulations is liable to pay compensation, and is responsible for all acts committed by persons forming part of its armed forces. This provision is the ancestor of the modern law of #state_responsibility for violations of humanitarian law and of the contemporary debate on #reparations for victims. Because both the 1899 and the 1907 versions remain formally in force between different combinations of states, lawyers occasionally have to determine which text applies. In practice this rarely matters, because the substantive content of the Regulations has long been regarded as declaratory of custom and therefore binding on all states regardless of ratification. 6. The Hague Regulations: A Structural Analysis The Regulations annexed to Convention IV consist of sixty articles arranged in three sections. Section I concerns belligerents. Section II concerns hostilities. Section III concerns military authority over the territory of a hostile state. This structure is worth memorising, because it corresponds to three enduring questions of the discipline: who may lawfully fight, how may they fight, and what happens when one army controls the territory of another. 6.1 Who may fight: the definition of the belligerent Articles 1 to 3 define the persons entitled to the rights of a belligerent. The laws, rights, and duties of war apply not only to armies but also to militia and volunteer corps, provided they satisfy four conditions. They must be commanded by a person responsible for subordinates. They must have a fixed distinctive emblem recognisable at a distance. They must carry arms openly. And they must conduct their operations in accordance with the laws and customs of war. These four conditions are the compromise that the Brussels Conference could not reach. They give irregular fighters a route into lawful combatancy, but they set the entry price at a level that a mass popular resistance would find hard to pay. Article 2 adds a narrow additional category, the levee en masse, covering the inhabitants of a territory not yet occupied who spontaneously take up arms to resist an invader, provided they carry arms openly and respect the laws of war. The conditions of Articles 1 and 2 passed almost unchanged into the Third Geneva Convention of 1949 and continue to structure the debate about who qualifies as a prisoner of war. The importance of these articles for students lies in what they establish, which is a legal frontier. On one side stands the lawful combatant, who may kill without committing murder, who may not be punished for the mere act of fighting, and who is entitled to prisoner of war status on capture. On the other side stands the person who fights without meeting the conditions, who enjoys no combatant immunity and may be prosecuted under the ordinary criminal law of the capturing state. The entire modern controversy about unprivileged belligerency, unlawful combatants, and the status of members of armed groups is a continuation of the argument settled provisionally in Articles 1 to 3. 6.2 Prisoners of war Articles 4 to 20 set out an early charter for #prisoners_of_war. Prisoners are in the power of the hostile government, not of the individuals who capture them. They must be humanely treated. Their personal belongings, apart from arms, horses, and military papers, remain their property. They may be interned but may not be confined except as an indispensable measure of safety. Their labour may be used, but not for work connected with military operations, and they must be paid. They must be maintained by the government that holds them, on the same footing as its own troops. Escape, if unsuccessful, may be punished only by disciplinary measures, and a prisoner who is recaptured after rejoining his own army may not be punished for the earlier escape at all. Prisoners are subject to the laws and regulations of the detaining army, and an information bureau must be established to record their identity and communicate with their families. Read today, these provisions look thin. They contain nothing on the interrogation of prisoners, little on medical care, and no supervision mechanism apart from the information bureau and the good offices of relief societies. The point, however, is not their completeness. It is that they exist. The Hague articles converted the treatment of captives from an act of grace by the victor into an obligation owed by the state. Every later development, from the Prisoner of War Convention of 1929 to the Third Geneva Convention of 1949 and its modern commentary, builds on that conversion. 6.3 Means and methods of warfare Section II opens with the sentence that carries more weight than any other in the Regulations. Article 22 provides that the right of belligerents to adopt means of injuring the enemy is not unlimited. Everything else in the law on the conduct of hostilities is an application of that proposition. Article 23 then lists specific prohibitions. It is especially forbidden to employ poison or poisoned weapons. It is forbidden to kill or wound treacherously individuals belonging to the hostile nation or army, a rule that underlies the modern prohibition of #perfidy. It is forbidden to kill or wound an enemy who has laid down his arms or, having no longer means of defence, has surrendered at discretion. It is forbidden to declare that no quarter will be given. It is forbidden to employ arms, projectiles, or material calculated to cause unnecessary suffering. It is forbidden to make improper use of a flag of truce, of the enemy uniform or insignia, or of the distinctive badges of the Geneva Convention. It is forbidden to destroy or seize enemy property unless imperatively demanded by the necessities of war. And it is forbidden to declare abolished, suspended, or inadmissible in a court of law the rights and actions of the nationals of the hostile party. Two observations follow. First, the list mixes rules of humanity with rules of good faith. The prohibition of poison and of unnecessary suffering protects the individual soldier from cruelty. The prohibition of perfidy and of the misuse of protective emblems protects the system of restraint itself, because a party that fakes surrender or hides behind a red cross destroys the trust on which every protective rule depends. Second, the qualifying words matter. Property may be destroyed when imperatively demanded by the necessities of war. The law does not forbid destruction, it forbids gratuitous destruction. The tension between #military_necessity and #humanity, which runs through the whole of humanitarian law, is written into the text. 6.4 Attack, siege, and bombardment Articles 25 to 27 address the protection of places. The attack or bombardment of towns, villages, dwellings, or buildings which are undefended is prohibited, by whatever means. The commander of an attacking force must, before commencing a bombardment except in cases of assault, do all in his power to warn the authorities. In sieges and bombardments all necessary steps must be taken to spare, as far as possible, buildings dedicated to religion, art, science, or charitable purposes, historic monuments, hospitals, and places where the sick and wounded are collected, provided they are not being used at the time for military purposes. Article 28 prohibits the pillage of a town or place, even when taken by assault, and Article 47 repeats the prohibition of pillage in occupied territory. These provisions are the seed of two modern doctrines. The first is the principle of #distinction between military objectives and civilian objects, which was given its full and general expression only in 1977 in Additional Protocol I. The Hague text does not state the principle abstractly. It protects undefended places, which is a much narrower rule, and it says nothing about the deliberate targeting of civilians in a defended city. The gap would be exposed catastrophically by the aerial bombardment campaigns of the Second World War. The second doctrine is the protection of #cultural_property, which was developed through the Roerich Pact of 1935 and then through the Hague Convention for the Protection of Cultural Property in the Event of Armed Conflict of 1954, an instrument that consciously carries the Hague name and the Hague inheritance. 6.5 Spies, parlementaires, capitulations, and armistices Articles 29 to 41 deal with a set of practical questions that a modern reader may find quaint but that mattered greatly to nineteenth century armies. A spy is defined as a person who, acting clandestinely or on false pretences, obtains or seeks to obtain information in the zone of operations of a belligerent with the intention of communicating it to the hostile party. A spy taken in the act may not be punished without previous trial, and a spy who rejoins his own army and is subsequently captured is treated as a prisoner of war with no responsibility for his previous acts of espionage. Soldiers in uniform who penetrate enemy lines to gather information are not spies. Parlementaires, that is persons authorised to negotiate under a white flag, are inviolable, although the commander to whom they are sent is not obliged to receive them and may take steps to prevent them from using their mission to gather intelligence. Capitulations must take into account the rules of military honour and must be scrupulously observed. Armistices suspend military operations by mutual agreement, must be clearly notified, and a serious violation by one party entitles the other to denounce them and even to resume hostilities immediately in urgent cases. Individual violation of an armistice entitles the injured party to demand punishment of the offenders and compensation for losses sustained. That last sentence is important and is often overlooked. It shows that the drafters were aware of the problem of individual misconduct and had a remedy in mind, but the remedy is punishment by the offender's own state at the demand of the injured state. The idea that an international body might itself try the individual was still far away. 6.6 Belligerent occupation Section III, Articles 42 to 56, is the most influential part of the Regulations, and it is the part that international courts cite most often today. Article 42 defines occupation by reference to fact rather than to claim. Territory is considered occupied when it is actually placed under the authority of the hostile army, and the occupation extends only to the territory where such authority has been established and can be exercised. This is a factual test, and it means that an occupying power cannot escape the law by denying that it is an occupant. Article 43 is the central provision. The authority of the legitimate power having in fact passed into the hands of the occupant, the latter shall take all the measures in his power to restore and ensure, as far as possible, public order and safety, while respecting, unless absolutely prevented, the laws in force in the country. Two obligations sit inside this sentence. The occupant must maintain order, and the occupant must leave the existing legal order alone unless absolutely prevented from doing so. The occupant is a temporary administrator, not a sovereign. It may not annex, it may not transform the political and legal identity of the territory, and it holds public property as an administrator and usufructuary, under Article 55, bound to safeguard the capital of public buildings, real estate, forests, and agricultural estates. The remaining articles fill in the detail. Family honour and rights, the lives of persons, private property, and religious convictions must be respected. Private property may not be confiscated. Collective penalties may not be inflicted on the population for acts of individuals for which they cannot be regarded as jointly and severally responsible. Taxes may be collected only under the rules in force and, as far as possible, for the needs of the administration of the territory. Requisitions in kind and services may be demanded only for the needs of the army of occupation, must be in proportion to the resources of the country, and must as far as possible be paid for in cash. The property of municipalities and of institutions dedicated to religion, charity, education, and the arts and sciences is treated as private property, and seizure or destruction of historic monuments and works of art is forbidden and should be made the subject of legal proceedings. The doctrinal significance of Section III is difficult to overstate. It established that #sovereignty is not transferred by conquest, that the occupant's powers are limited and functional, and that the population of an occupied territory retains rights that the occupant must respect. Almost every contemporary controversy about occupation, whether concerning settlement policy, changes to local law, the exploitation of natural resources, or the duration of an occupation, is argued with Articles 42, 43, 46, 47, 52, and 55 in hand. The International Court of Justice has repeatedly treated these provisions as declaratory of customary law and has applied them in its advisory and contentious jurisprudence on occupied territory. 6.7 Article 3 of Convention IV and the question of remedy Article 3 of the 1907 Convention stands outside the Regulations and deserves separate treatment. It provides that a belligerent party which violates the provisions of the Regulations shall, if the case demands, be liable to pay compensation, and that it shall be responsible for all acts committed by persons forming part of its armed forces. Three points follow. First, the article establishes attribution. The state answers for the acts of its soldiers, including acts committed contrary to orders, which is a rule of considerable practical severity. Second, the article establishes a remedy, and the remedy is monetary. Third, and controversially, the article says nothing about who may claim. Whether individual victims can rely on Article 3 to demand compensation directly, or whether the right belongs only to the injured state, has been debated for a century and litigated in national courts. The dominant view has been that the classical structure of the provision is interstate, but the growth of victim centred approaches in modern international law has kept the argument alive. What Article 3 conspicuously does not do is criminalise. It converts a violation into a debt, not into an offence. The absence of a penal dimension is the great structural gap of the Hague system, and it is the gap that the twentieth century would spend a hundred years trying to fill. 7. The Martens Clause 7.1 Origin The most quoted sentence to come out of The Hague was not a rule at all. It was a preambular compromise, drafted under pressure by Fyodor Fyodorovich Martens, the Russian jurist and diplomat who chaired the subcommission on the laws of war in 1899. The problem, as noted above, was the status of civilians who resist an invader. Belgium and other smaller states refused to accept a text that treated their populations as criminals for defending their homes. Germany and other great military powers refused to accept a text that gave irregular fighters the protections of soldiers. The conference came close to collapse on the point. Martens proposed that the delegates should not decide the question at all, and should instead declare that whatever the Regulations failed to cover remained governed by the principles of the law of nations. The formula allowed each side to believe that the residual law favoured its own position, and the conference was saved. 7.2 The text In the preamble to Convention II of 1899, and again in Convention IV of 1907, the clause states that until a more complete code of the laws of war has been issued, the High Contracting Parties think it right to declare that, in cases not included in the Regulations they have adopted, populations and belligerents remain under the protection and the rule of the principles of the law of nations, as they result from the usages established among civilised peoples, from the laws of humanity, and the dictates of the public conscience. The #Martens_Clause has been restated, in modified form, in the denunciation provisions of the four Geneva Conventions of 1949, in Article 1(2) of Additional Protocol I of 1977, in the preamble of Additional Protocol II, and in the preamble of the Convention on Certain Conventional Weapons. In the 1977 formulation the beneficiaries are civilians and combatants rather than populations and belligerents, and the reference to civilised peoples is replaced by established custom. 7.3 Interpretations The literature identifies at least three readings of the clause, and the difference between them is not academic. The narrow reading treats the clause as a saving provision. It says only that the adoption of a treaty does not extinguish the customary rules that existed before it. On this view the clause adds nothing to the law and simply prevents a bad argument, namely the argument that whatever is not prohibited by the treaty is therefore permitted. The intermediate reading treats the clause as an interpretive principle. Where a treaty rule is ambiguous, it should be read in the light of the principles of humanity and the dictates of public conscience. This reading gives the clause work to do without letting it create new obligations by itself. The expansive reading treats the principles of humanity and the dictates of public conscience as independent sources of legal obligation, capable of prohibiting conduct that no treaty and no settled custom addresses. On this view a weapon might be unlawful because it offends the #public_conscience, even in the absence of a specific ban. International courts have used the clause repeatedly but have never fully resolved this disagreement. The International Court of Justice, in its advisory opinion on the legality of the threat or use of nuclear weapons, described the clause as an effective means of addressing the rapid evolution of military technology and confirmed its customary status, without endorsing the expansive reading. Criminal tribunals have invoked it to reinforce conclusions reached on other grounds rather than as a free standing prohibition. 7.4 Why the clause still matters The practical importance of the clause today lies in the regulation of new technologies. Debates about #autonomous_weapons, about #cyber_warfare, and about the environmental effects of hostilities all confront the same structural problem that Martens faced. The treaty texts were written for a world that did not contain the object being regulated. The clause supplies an answer to the argument that a gap in the treaty is a licence. Campaigners for a ban on lethal autonomous weapons rely on it heavily, arguing that delegating the decision to kill to a machine offends the principles of humanity whether or not a specific treaty prohibits it. Sceptics reply that public conscience is not a measurable legal standard and that treating it as one invites courts and advocacy groups to legislate. Both positions are defensible, and students should be able to state each of them fairly. 8. The Principles Crystallised at The Hague Although the Regulations are drafted as a list of specific rules, four general principles can be extracted from them. These principles were not invented at The Hague, but The Hague is where they acquired treaty form, and modern textbooks still organise the subject around them. 8.1 Military necessity Military necessity permits measures that are actually necessary to accomplish a legitimate military purpose and that are not otherwise prohibited by law. The definition contains its own limit. Necessity is not a licence to do whatever helps to win. It operates inside the law, not above it. The doctrine of Kriegsraison, which held that necessity in war overrides the rules of war, was argued by defendants after both world wars and was rejected by the tribunals. The Hague Regulations acknowledge necessity, in Article 23(g) and elsewhere, but they acknowledge it as an exception written into particular rules rather than as a general escape clause. 8.2 Humanity The principle of humanity forbids the infliction of suffering, injury, or destruction that is not necessary for a legitimate military purpose. It is the counterweight to necessity, and the two exist in permanent tension. The prohibitions on poison, on weapons causing unnecessary suffering, and on denying quarter are all expressions of it, as is the Martens Clause itself. 8.3 Distinction The Hague Regulations contain distinction in embryo rather than in mature form. The protection of undefended places, the prohibition of pillage, the respect owed to private property, and the safeguards for hospitals and cultural buildings all presuppose that #civilians and civilian objects are not legitimate targets. But the text never says so in general terms, and it does not regulate the bombardment of defended cities in which civilians live. The full principle, with its definition of military objective and its prohibition of indiscriminate attack, had to wait for Additional Protocol I. Students should read this as an example of a wider pattern. The Hague drafters legislated for the war they knew, which was fought by armies on land within reach of one another, and they could not legislate for the war that was coming, in which aircraft would place whole populations within reach. 8.4 Limitation on means Article 22, with its statement that the right to choose means of injuring the enemy is not unlimited, is the foundation of weapons law. Every weapons treaty since 1899, from the gas protocol of 1925 to the conventions on biological and chemical weapons, on certain conventional weapons, on anti personnel mines, and on cluster munitions, is an elaboration of that sentence. Article 36 of Additional Protocol I, which requires states to review the legality of new weapons, is its procedural expression. The 1899 declarations on gas and on expanding bullets show the method in miniature: identify a specific weapon, assess it against the principles of humanity and unnecessary suffering, and prohibit it by name. 8.5 The proportionality question Proportionality in attack, the rule that expected incidental harm to civilians must not be excessive in relation to the concrete and direct military advantage anticipated, does not appear in the Hague Regulations. This is one of the most important gaps in the 1907 text. #Proportionality was developed through state practice, through military manuals, and above all through Additional Protocol I. Its absence from the Regulations is a reminder that the Hague texts are a foundation and not a completed building. 9. From Treaty Violation to War Crime 9.1 The problem stated The Hague Conventions created duties for states and a remedy in compensation. They did not create crimes, courts, or penalties. How then did they become the first formal statement of war crimes in international law, as they are so often described? The answer lies in a chain of developments over fifty years, in which the substantive prohibitions written at The Hague were progressively attached to mechanisms of #accountability that the drafters never contemplated. The Hague texts supplied the content. Later practice supplied the sanction. 9.2 The First World War and its aftermath The Great War tested the Hague system immediately and found it wanting. Poison gas was used on a large scale, notwithstanding the 1899 declaration. Occupied Belgium and northern France were subjected to deportation, forced labour, requisition, and collective punishment, in direct conflict with Section III of the Regulations. Merchant ships were sunk without warning. Cities were shelled and bombed from the air. The response after 1918 was the first serious attempt to link violations to individual punishment. The Commission on the Responsibility of the Authors of the War, established by the victorious powers in 1919, compiled a list of offences against the laws and customs of war, drawing directly on the Hague Regulations, and recommended prosecutions. The Treaty of Versailles provided in Article 227 for the arraignment of the former German Emperor before a special tribunal, and in Article 228 for the surrender of persons accused of violating the laws and customs of war for trial before Allied military tribunals. Almost none of it worked. The Netherlands refused to surrender the Kaiser. The Allies eventually allowed Germany to try a small number of its own nationals before the Supreme Court in Leipzig. The Leipzig trials produced a handful of convictions, mostly with light sentences, and became a byword for the failure of national courts to police their own armed forces. The episode was not wholly barren. It established two ideas that would return. The first is that violations of the laws and customs of war, as defined in the Hague Regulations, can be the basis of criminal charges against individuals. The second is that leaving prosecution to the state of the accused is unlikely to produce justice. 9.3 Nuremberg and the customary law argument The decisive moment came in 1945 and 1946. The Charter of the International Military Tribunal gave the Tribunal jurisdiction over crimes against peace, war crimes, and crimes against humanity. War crimes were defined as violations of the laws or customs of war, and the examples given, including murder or ill treatment of prisoners of war, plunder of public or private property, wanton destruction of cities not justified by military necessity, and deportation of civilian populations from occupied territory, correspond closely to the prohibitions of the Hague Regulations. The defendants raised two objections that go to the heart of this article's subject. The first was that Germany could not be bound by the 1907 Convention in the Second World War because the general participation clause, discussed below, excluded its application when a non party took part in the conflict. The second was that the Hague Regulations contained no penal provisions and could not therefore be the basis of criminal liability, so that punishing individuals under them amounted to retroactive law. The Tribunal rejected both objections. On the first, it held that by 1939 the rules laid down in the Hague Convention were recognised by all civilised nations and were regarded as declaratory of the laws and customs of war, so that the question whether the Convention applied as a treaty had ceased to matter. On the second, it held that international law imposes duties on individuals as well as on states, that crimes against international law are committed by men and not by abstract entities, and that only by punishing individuals who commit such crimes can the provisions of international law be enforced. This is the moment at which the Hague Regulations were converted from an interstate bargain into the substantive law of #war_crimes. The conversion was accomplished by judicial reasoning, not by treaty amendment. It rested on the proposition that the Hague rules had become #customary_international_law binding on every state and every soldier, whether or not the relevant convention applied by its own terms. Critics at the time and since have argued that this amounted to retroactive criminalisation dressed in the language of custom. The better view, and the one that has prevailed, is that the prohibitions themselves were undoubtedly law before the war, that they were known to every professional officer, and that what was new in 1946 was the forum and the sanction rather than the standard of conduct. 9.4 The Geneva system and the grave breaches regime The four #Geneva_Conventions of 1949 supplied what The Hague had lacked. Each contains a list of grave breaches, and each requires states parties to enact penal legislation, to search for persons alleged to have committed grave breaches, and either to bring them before their own courts or to hand them over to another state party for trial. This is the principle of universal jurisdiction over grave breaches, and it is the first treaty based enforcement mechanism in the law of armed conflict. The relationship between the two bodies of rules is often described as the distinction between Hague law, which governs the conduct of hostilities, and Geneva law, which protects the victims of conflict. The distinction is useful for teaching but has been substantially dissolved in practice. Additional Protocol I of 1977 regulates the conduct of hostilities in detail and is a Geneva instrument. The Fourth Geneva Convention of 1949 regulates occupied territory and therefore overlaps directly with Section III of the Hague Regulations, supplementing it without replacing it. Courts today apply both bodies together. 9.5 The tribunals of the 1990s and the Rome Statute The ad hoc tribunals for the former Yugoslavia and for Rwanda revived international criminal justice after four decades of dormancy. Their statutes and their case law again drew on the Hague inheritance. The Yugoslav Tribunal's jurisdiction included violations of the laws or customs of war, an open textured category which the Tribunal filled by reference to customary law, including Hague derived rules on the destruction of property, the plunder of public and private property, and attacks on undefended towns and on institutions dedicated to religion, charity, education, and the arts and sciences. Its jurisprudence confirmed that serious violations of customary rules entail individual criminal responsibility even in the absence of an express penal provision in the underlying treaty. The Rome Statute of the International Criminal Court, adopted in 1998, completed the process. Article 8 lists war crimes in four categories, and the second of these, other serious violations of the laws and customs applicable in international armed conflict within the established framework of international law, is in substance a modernised and expanded Hague list. It includes intentionally directing attacks against civilians, attacking undefended towns, killing or wounding a combatant who has surrendered, improper use of a flag of truce, employing poison or poisoned weapons, employing asphyxiating gases, employing expanding bullets, declaring that no quarter will be given, pillaging, and destroying or seizing enemy property unless imperatively demanded by the necessities of war. A student who reads Article 8(2)(b) of the #Rome_Statute alongside Articles 23 to 28 of the Hague Regulations will recognise most of the material immediately. The #ICC did not invent the war crime. It supplied the court that the conferences of 1899 and 1907 were unable to create. 10. The Failures and Blind Spots of the Hague System An honest assessment must give the criticisms as much space as the achievements. The Hague Conventions failed in several respects, and some of those failures were not accidents of drafting but reflections of the political order that produced them. 10.1 The general participation clause Article 2 of Convention IV of 1907 provides that the Convention and its Regulations do not apply except between contracting powers, and then only if all the belligerents are parties. This is the clausula si omnes, the #si_omnes clause or general participation clause. Its effect is close to absurd. If a single belligerent in a multi party war is not a party to the Convention, the Convention ceases to bind any of them. In the First World War the participation of states that had not ratified meant that, on a strict reading, the Convention did not formally apply to the very conflict it had been written for. The clause was a concession to formalism and to the reluctance of states to accept obligations towards non parties. It has been abandoned in modern treaties, and the Geneva Conventions of 1949 apply between parties even when one belligerent is not a party. But the damage the clause did to the credibility of the Hague system in its first great test was severe. The only reason it did not destroy the system altogether is that tribunals afterwards relied on custom rather than on the treaty. 10.2 No enforcement The Hague Conventions contain no monitoring body, no dispute settlement clause specific to violations of the Regulations, no penal provisions, and no obligation to prosecute. The Permanent Court of Arbitration created in 1899 was available only if both parties agreed to submit a dispute to it, and no belligerent proposed to arbitrate its own conduct in the middle of a war. #Enforcement was left to reciprocity, to reprisals, and to public opinion. Reciprocity works only while both sides find compliance useful. Reprisals, which the law of the time permitted, are a mechanism of enforcement that operates by committing further violations. Public opinion is not a legal remedy. 10.3 The disarmament failure The original purpose of the 1899 conference, and a substantial part of the 1907 agenda, was the limitation of armaments. Both conferences failed entirely on this point. The naval arms race between Britain and Germany accelerated after 1907. The conferences regulated the conduct of a war they could not prevent, and the war came seven years after the second of them. Some critics have therefore described the Hague project as a humanitarian consolation prize for a peace project that never had any real chance of success. The criticism has force, although it is also true that the humanitarian achievements were real and durable while the disarmament proposals were never politically serious. 10.4 Technology outran the text The 1899 declaration on projectiles launched from balloons was renewed in 1907 for a limited period, and the prohibition then lapsed. Within a decade of the second conference, aircraft were bombing cities. The Regulations, drafted for land warfare conducted by armies in contact, had almost nothing to say about strategic aerial bombardment, and the attempt to fill the gap through the draft Hague Rules of Air Warfare in the early 1920s failed to attract ratification. The result was that the most destructive form of attack on civilians in the twentieth century developed in a legal vacuum that was filled only in 1977. The same structural lesson applies to submarine warfare then, and to cyber operations and autonomous systems now. 10.5 Civilised nations and the colonial exclusion The Hague texts speak of civilised peoples and of the practice of civilised nations. This was not casual language. The international law of the period was understood by its practitioners as the law of a community of European and European descended states, into which other polities might be admitted if they satisfied a standard of civilisation. The Ottoman Empire, Persia, China, Japan, and Siam attended the conferences, and Japan in particular was concerned to demonstrate compliance with European legal standards. But large parts of the world were not represented at all, because they were colonies, protectorates, or territories controlled by the powers sitting at the table. The consequence was that the restraints agreed at The Hague were applied selectively. Colonial warfare, and campaigns of pacification against peoples classified as uncivilised, were widely regarded as falling outside the Regulations or as governed by a diluted version of them. Expanding bullets prohibited for use against European soldiers were considered acceptable for use elsewhere. Critical and postcolonial scholarship on international law has documented this pattern in detail, and any honest account of the Hague achievement must include it. The law of war was born with a geography. 10.6 The rebel and the resistance fighter Finally, the Regulations apply only to war between states. Civil war, insurrection, and rebellion fall outside their scope. A government fighting its own population owed no Hague obligations to the rebels, who remained criminals under domestic law. This gap was addressed only in 1949, through Common Article 3 of the Geneva Conventions, and was then developed in Additional Protocol II of 1977. Since most armed conflicts since 1945 have been non international in character, the limitation is a serious one, and it explains why the Hague Regulations, though foundational, cannot be the whole of the modern law. 11. The Living Legacy of the Hague Conventions Despite these failures, the Hague texts remain in force and in use. Four areas illustrate their continuing operation. 11.1 Customary status and universal application The most important single fact about the Hague Regulations today is that they bind everyone. Their customary status, affirmed at Nuremberg, has been reaffirmed by the International Court of Justice, by the ad hoc tribunals, and by the major scholarly and institutional studies of customary humanitarian law. A state that has never ratified Convention IV is bound by the Regulations. A soldier of such a state can be prosecuted for violating them. This is why the general participation clause no longer matters, and it is why a treaty drafted for the armies of Kaiser Wilhelm and Tsar Nicholas can be applied to conflicts fought with drones and precision munitions. 11.2 Occupation law in contemporary practice Section III has never been more actively litigated. The #ICJ has relied on Articles 42 to 56 in its jurisprudence concerning occupied territory, treating them as customary and applying them alongside the Fourth Geneva Convention. The same provisions have been invoked in relation to occupations in the Middle East, in Africa, and in Europe, including in the context of the conflict in #Ukraine, where questions of annexation, of transfer of population, of exploitation of resources, and of the administration of justice in occupied areas are argued in the vocabulary of Article 43 and Article 55. The Hague definition of occupation as a question of fact rather than of title is what allows the law to be applied to a state that denies that it is an occupant. 11.3 Weapons law and new technologies Article 22 and Article 23(e) continue to supply the general standard against which new weapons are measured, and the Martens Clause continues to supply the argument that a gap in treaty coverage is not a permission. The debate over autonomous weapons is conducted substantially in Hague terms. Can a system comply with distinction and proportionality, does its use cause unnecessary suffering, and does the delegation of lethal decisions to software offend the dictates of public conscience. The debate over cyber operations asks whether such an operation is an attack, whether data is an object, and whether the prohibition of perfidy extends to digital deception. These are new questions, but the conceptual apparatus used to answer them was assembled in 1899 and 1907. 11.4 The protection of cultural property The 1954 Hague Convention for the Protection of Cultural Property in the Event of Armed Conflict, together with its two protocols, elaborated Articles 27 and 56 of the Regulations into a specialised regime with its own definitions, marking system, and penal provisions. The destruction of cultural heritage in armed conflict has since been prosecuted as a war crime before the International Criminal Court, an outcome that descends directly from the modest sentence in Article 27 requiring commanders to spare buildings dedicated to religion, art, and science. 12. Discussion 12.1 What kind of achievement was it The Hague Conventions are best understood not as a code but as a threshold. Before them, the proposition that war has legal limits was a claim made by moralists and by some soldiers. After them it was a proposition to which states had formally agreed in writing. That change of status made everything else possible. It gave later reformers a text to build on, gave tribunals a standard to apply, gave armies a syllabus to teach, and gave victims a language in which to state a grievance. At the same time the Conventions were a product of their moment. They accepted war as lawful. They protected soldiers more thoroughly than they protected civilians. They assumed that armies fight armies. They confined their protective community to states recognised as civilised. They provided no penalty and no policeman. To praise them without noting these features is to write hagiography rather than history. 12.2 The paradox of enforcement There is a paradox at the centre of the Hague story. The treaties that contained no criminal provisions became the source of the law of war crimes, while the disarmament proposals backed by an emperor achieved nothing at all. The reason is instructive. The Hague drafters wrote rules that armies could actually follow and that stated with reasonable clarity what a soldier must not do. Once such rules exist and are widely accepted, the demand for enforcement follows almost automatically, because a rule that everyone acknowledges and that is openly violated creates political pressure for punishment. Disarmament, by contrast, required states to surrender capability against an uncertain future, and no amount of drafting could induce them to do so. The lesson for students of international law is that sequence matters. Clear standards can generate institutions. Institutions rarely generate standards from nothing. 12.3 Hague law and Geneva law today Textbooks still divide the field into Hague law and Geneva law, and the division has pedagogical value. But the two streams merged in 1977 and have been treated as a single body ever since. Additional Protocol I contains rules on the conduct of hostilities that are Hague in character and Geneva in provenance. The Rome Statute lists war crimes drawn from both traditions in the same article. Students should learn the distinction and then learn why it no longer determines outcomes. 12.4 Implications for teaching and research For teaching, the Hague Conventions offer an unusually clear case study in how codification works, how customary law forms, and how a rule without a sanction can nevertheless become the basis of criminal liability. For research, several questions remain open. The proper interpretation of the Martens Clause is unresolved and increasingly consequential. The status of individual claims to #reparations under Article 3 of Convention IV is contested. The application of occupation law to prolonged occupations, to occupations conducted through proxies, and to territories administered without any formal claim of occupation remains difficult. And the historical scholarship on the colonial and racial limits of the original Hague settlement is still developing. 13. Conclusion The Hague Conventions of 1899 and 1907 deserve their standard description as the first formal statements of the laws of war and of war crimes in international law, provided that the description is understood precisely. They were the first multilateral treaties that told states, in binding written form, how they might and might not fight. They set out the conditions of lawful combatancy, the treatment of prisoners, the limits on weapons and methods, the protection of undefended places and cultural buildings, and the duties of an occupying power. They contained the Martens Clause, which insisted that the absence of a rule is not the absence of law. They created the first standing institution for the settlement of interstate disputes. They did not create crimes, and they did not create courts. What they created was the substance from which crimes were later defined. At Nuremberg the Hague Regulations were held to be declaratory of custom and therefore binding on individuals, and from that holding runs a straight line to the Geneva grave breaches regime, to the ad hoc tribunals, and to Article 8 of the Rome Statute. The prohibitions written in 1899 by delegates who had never seen an aircraft in combat are the prohibitions that an international prosecutor applies today. The Hague system also failed, and its failures were consequential. The general participation clause was an absurdity. The absence of enforcement left compliance to reciprocity and reprisal. The disarmament agenda collapsed. The rules said nothing about aerial bombardment, nothing about civil war, and nothing that constrained the conduct of colonial powers against peoples they did not regard as civilised. The twentieth century paid for those gaps in lives. The correct conclusion is neither celebration nor dismissal. The Hague Conventions were the moment at which the international community wrote down, for the first time and in a form that states accepted, the idea that the right of belligerents to injure the enemy is not unlimited. Everything achieved in the law of armed conflict since then has been an extension, a repair, or an enforcement of that single sentence. Hashtags #Hague_Conventions_1899_1907 #laws_of_war_history #international_humanitarian_law #war_crimes_law #Martens_Clause_debate #Hague_Regulations_1907 #belligerent_occupation_law #conduct_of_hostilities #history_of_international_law #Nuremberg_legacy #customary_IHL #Hague_Peace_Conferences #law_of_armed_conflict #legal_research_students #war_crimes_accountability References Casey-Maslen, S. (2020). Weapons. In B. Saul and D. Akande (Eds.), The Oxford guide to international humanitarian law (pp. 261-276). Oxford University Press. Clapham, A. (2021). War. Oxford University Press. Crawford, E., and Pert, A. (2020). International humanitarian law (2nd ed.). Cambridge University Press. Dinstein, Y. (2022). The conduct of hostilities under the law of international armed conflict (4th ed.). Cambridge University Press. https://doi.org/10.1017/9781009106191 Duerr, B. (2024). De droom van Den Haag. Atlas Contact. Fleck, D. (Ed.). (2021). The handbook of international humanitarian law (4th ed.). Oxford University Press. Grimberg, S. (2025). Modern interpretations of international humanitarian law's Martens Clause: Opening the door to strategies to better protect the environment and Indigenous Peoples during armed conflict. Asia Pacific Journal of International Humanitarian Law. Henckaerts, J. M. (2020). History and sources. In B. Saul and D. Akande (Eds.), The Oxford guide to international humanitarian law (pp. 1-28). Oxford University Press. International Committee of the Red Cross. (2021). Commentary on the Third Geneva Convention: Convention (III) relative to the treatment of prisoners of war. Cambridge University Press. Ivanenko, V. S. (2022). The origins, causes and enduring significance of the Martens Clause: A view from Russia. International Review of the Red Cross, 104(920-921), 1708-1724. https://doi.org/10.1017/S1816383122000273 Kalmanovitz, P. (2020). The laws of war in international thought. Oxford University Press. Leisure, P. (2021). The Martens Clause, global pandemics, and the law of armed conflict. Harvard International Law Journal, 62(2), 279-334. Nijs, M. (2024). Review of International humanitarian law: Rules, controversies, and solutions to problems arising in warfare, by M. Sassoli. Journal of Conflict and Security Law, 29(3), 391-397. https://doi.org/10.1093/jcsl/krae010 Robertson, G. (2024). Crimes against humanity: The struggle for global justice (5th ed.). Penguin Books. Sassoli, M. (2024). International humanitarian law: Rules, controversies, and solutions to problems arising in warfare (2nd ed.). Edward Elgar Publishing. Saul, B., and Akande, D. (Eds.). (2020). The Oxford guide to international humanitarian law. Oxford University Press. Schmitt, M. N. (2020). International humanitarian law and the conduct of hostilities. In B. Saul and D. Akande (Eds.), The Oxford guide to international humanitarian law (pp. 147-174). Oxford University Press. Sivakumaran, S., and Burne, C. R. (Eds.). (2024). Making and shaping the law of armed conflict. Oxford University Press. Vite, S. (2020). Occupation. In B. Saul and D. Akande (Eds.), The Oxford guide to international humanitarian law (pp. 299-318). Oxford University Press.

The Rome Statute, adopted on 17 July 1998 and in force since 1 July 2002, created the first permanent #international_criminal_court with authority to try individuals for genocide, crimes against humanity, war crimes, and, since 2018, the crime of aggression. This article examines the treaty as a legal instrument, an institution, and a political project. It traces the historical path from the post-1945 military tribunals to the #ad_hoc_tribunals of the 1990s and then to Rome; explains the Court's structure, jurisdiction, trigger mechanisms, and admissibility rules; analyses the definitions of the core crimes; and assesses the Court's practice from 2002 to mid-2026. The article gives particular attention to complementarity, to the tension between Article 27 (irrelevance of official capacity) and Article 98 (immunities of third states), to victim participation and reparations, and to the enforcement gap created by the Court's lack of a police force. It then evaluates the political pressures now facing the Court, including United States sanctions imposed under Executive Order 14203 in February 2025, Russian retaliation against Court officials, and withdrawal notices from Burkina Faso, Mali, and Niger. The analysis finds that the Statute has succeeded as a normative framework and as a catalyst for domestic prosecution, but that its enforcement record remains modest and unevenly distributed. The article argues that the Court's future depends less on further doctrinal refinement than on #state_cooperation, financial stability, and a credible answer to the charge of selectivity. The study is doctrinal and qualitative, based on the Statute, ICC jurisprudence, and peer reviewed scholarship published mainly between 2021 and 2026. Keywords: Rome Statute; International Criminal Court; complementarity; crime of aggression; state cooperation; reparations; international criminal law. INTRODUCTION For most of recorded history, the people who ordered mass killing were rarely punished. Wars ended, treaties were signed, and leaders retired. The idea that an individual, and not only a state, could be held criminally responsible under international law is young. It became institutional reality only in the twentieth century, and it became permanent only with the #Rome_Statute of 1998. The Statute is the founding treaty of the #International_Criminal_Court, an independent judicial body seated in The Hague. It was adopted at a diplomatic conference in Rome on 17 July 1998, and it entered into force on 1 July 2002 after sixty ratifications. As of mid-2026, 125 states are parties. The Court has authority over four categories of conduct: #genocide, #crimes_against_humanity, #war_crimes, and the #crime_of_aggression. It tries #natural_persons, not states, and it acts only when national courts do not. This last point is the heart of the system and is often misunderstood. The ICC is not a supreme court sitting above national judiciaries, and it is not an appeal court for human rights complaints. It is a #court_of_last_resort. Under the principle of #complementarity, a case is inadmissible if a state with jurisdiction is genuinely investigating or prosecuting it. The Court steps in when a state is unwilling or unable to act genuinely, a test set out in Article 17. The Statute matters for three reasons. First, it codified the definitions of the core crimes in treaty form, so that a lawyer in Manila, Nairobi, or Bogota can consult the same text. Second, it built an institution with a #Prosecutor who can open investigations on her or his own initiative, subject to judicial authorisation, rather than waiting for a political body to act. Third, it gave victims a formal place in proceedings, including the right to participate and to seek #reparations. The Statute is also, and increasingly, a site of conflict. Between 2023 and 2026 the Court issued arrest warrants against a sitting Russian head of state, a sitting Israeli prime minister, and a Hamas commander; took custody of a former Philippine president; and delivered its first judgment on crimes in the Central African Republic. In the same period, the United States sanctioned the Prosecutor, deputy prosecutors, and several judges; Russia convicted Court officials in absentia; Hungary announced withdrawal and then reversed course; and three Sahel states began the process of leaving. The Court, in other words, is being tested at exactly the moment when it is acting most ambitiously. This article asks a simple question with a complicated answer: what has the Rome Statute actually achieved, and what has it not? It is written for students who need a clear map of a technical field, but it does not simplify to the point of distortion. Sections 2 and 3 set out the literature and method. Sections 4 and 5 give the historical background. Sections 6 to 11 explain the legal architecture. Sections 12 to 16 assess practice and impact. Section 17 turns to reform. Section 18 states the limits of the study, and Section 19 concludes. LITERATURE REVIEW Scholarship on the ICC falls into four broad streams, and it is useful to see them as separate conversations that sometimes talk past one another. The first stream is doctrinal. It analyses the text of the Statute, the Elements of Crimes, the Rules of Procedure and Evidence, and the Court's case law. The leading reference works are the article-by-article commentary edited by Ambos (2022) and the textbook by Robinson, Vasiliev, van Sliedregt, and Oosterveld (2024), the fifth edition of a work long associated with Robert Cryer. Doctrinal writing asks what the law is: what counts as a widespread or systematic attack, when command responsibility attaches, how #admissibility is assessed. It is indispensable, but it tends to treat political resistance as background noise rather than as data. The second stream is institutional and sociological. It studies the Court as an organisation embedded in a field of states, non-governmental organisations, and donors. Wierda (2023) is a useful example, examining the local impact of the Court in Afghanistan, Colombia, Libya, and Uganda and arguing that the distance between legal outputs and felt justice is wide. This stream is sceptical of the assumption that a judgment in The Hague translates automatically into meaning in an affected village. The third stream is empirical and quantitative, largely produced by political scientists. It asks whether the Court changes behaviour. Miller (2023) finds that ICC indictments are associated with reductions in atrocities, arguing that armed groups linked to an indictee often reduce attacks because they perceive that restraint may lower the costs imposed on them. McAllister (2023), writing in a special issue of the Journal of Human Rights marking the Court's first twenty-five years, is more cautious, finding that #deterrence effects are conditional and strongest where the Court takes repeated, visible action during an ongoing conflict. Broache, Cronin-Furman, Mendeloff, and McAllister (2023) introduce that special issue and stress that outcomes depend heavily on state cooperation. Garcia Iommi (2023), in the same issue, argues that patterns of state support have remained remarkably stable since 1998, which complicates the popular narrative of a sudden collapse in support. The fourth stream is critical and post-colonial. It asks who is prosecuted and who is not. Writers in this tradition point out that for the Court's first fifteen years, every situation under formal investigation concerned an African state. Ba (2023) situates African resistance, including the Malabo Protocol project for an African criminal chamber, within a longer history of contested legal authority. Phooko (2023) examines what he calls existential threats to the Court arising from the convergence of geopolitical interests. The critical stream is sometimes dismissed as polemical, but its central empirical claim, that enforcement is uneven, is not seriously contested. Two further clusters deserve mention because they dominate current debate. On the #crime_of_aggression, Dannenbaum (2022) and McDougall (2023) analyse why the ICC cannot exercise jurisdiction over Russian aggression in Ukraine and why a special tribunal was proposed instead; Kress (2025) reports on the July 2025 special session of the Assembly of States Parties, where a proposal to align the jurisdictional regime for aggression with that of the other crimes attracted wide support but the final decision was deferred to 2029. On #immunity and cooperation, Sabel (2024) presents the objections to the warrants against Israeli officials, while Sterio (2026) analyses the United States sanctions regime and its operational effects on the Court. The gap this article addresses is integrative. Doctrinal work rarely engages the empirical findings on deterrence; empirical work often treats legal categories as fixed. This article reads the Statute's architecture and its practice together, and it updates the picture through mid-2026, a period in which the Court's docket and its political environment both changed sharply. METHODOLOGY This is a doctrinal and qualitative study. It does not generate new statistical data. Its method has three components. First, primary legal analysis. The Rome Statute is read as a treaty, using the ordinary meaning of its terms in context and in light of its object and purpose. Where the text is contested, the article examines how ICC chambers have interpreted it, notably on #jurisdiction, admissibility, and immunities. Second, case analysis. The article draws on the Court's public record of situations, arrest warrants, confirmation decisions, judgments, and reparations orders through mid-2026. Cases are used illustratively, to show how abstract rules operate, rather than exhaustively. Third, a structured review of recent scholarship. Sources were selected on three criteria: peer reviewed or published by an academic press; published mainly between 2021 and 2026; and directly concerned with the Statute, the Court, or the enforcement of atrocity crimes. Older foundational work is acknowledged in substance but not relied upon for current claims. The article's limitations follow from this design and are stated in Section 18. In particular, doctrinal analysis cannot establish causation, and claims about #deterrence are reported from the empirical literature rather than independently verified. HISTORICAL ROOTS: FROM VERSAILLES TO ARUSHA The Rome Statute did not appear from nothing. It is the product of roughly eighty years of trial and error. The first modern attempt at international criminal accountability came after the First World War. The Treaty of Versailles contemplated the trial of the German Kaiser for a supreme offence against international morality. He was never tried; the Netherlands refused to surrender him. A small number of trials took place before German courts at Leipzig, and they are generally regarded as a failure. The episode is instructive because it shows the pattern that recurs throughout this history: the law can be written faster than states are willing to enforce it. The decisive moment was 1945. The International Military Tribunal at #Nuremberg, and its counterpart in Tokyo, tried senior German and Japanese leaders for crimes against peace, war crimes, and crimes against humanity. Nuremberg established a principle that the Rome Statute later codified in Article 27: official position, including that of head of state, is not a defence and does not exempt a person from criminal responsibility. It also established that individuals, not abstract entities, commit crimes. Nuremberg has always carried a shadow. It was justice imposed by victors on the vanquished, and the Allied powers were not themselves subject to the tribunal's authority. That objection, sometimes called the problem of #selective_justice, has followed international criminal law ever since, and it is the sharpest weapon in the hands of the Court's present critics. The Cold War froze the project. The United Nations General Assembly asked the International Law Commission to draft a statute for a permanent court, and drafts were produced, but the political conditions did not exist. For roughly four decades, international criminal justice was largely dormant. It revived suddenly in the 1990s, in response to two catastrophes. In 1993, the Security Council created the International Criminal Tribunal for the former Yugoslavia. In 1994, after the Rwandan genocide, it created the International Criminal Tribunal for Rwanda. These ad hoc tribunals proved that international trials of senior figures were practically possible. They developed a substantial jurisprudence on genocide, on sexual violence as an international crime, and on modes of liability. They also revealed serious weaknesses: they were expensive, slow, geographically limited, and created by the Security Council, which meant their existence depended on the agreement of five permanent members. The lesson drawn by many states was that a permanent, treaty-based court would be fairer and more efficient than a series of ad hoc bodies created after the fact. A permanent court would also avoid the accusation that tribunals are created for some victims and not for others. That reasoning drove the negotiations that culminated in Rome. THE ROME CONFERENCE AND THE ADOPTION OF THE STATUTE The United Nations Diplomatic Conference of Plenipotentiaries on the Establishment of an International Criminal Court met in Rome from 15 June to 17 July 1998. More than 160 states attended, along with a large coalition of non-governmental organisations that played an unusually visible role. Several issues divided the conference. The first was the power of the Prosecutor. Some states wanted investigations to begin only when referred by a state party or by the Security Council. Others insisted that the Prosecutor be able to act on her or his own initiative, on the basis of information from any source. The compromise, in Article 15, allows the Prosecutor to open an investigation #proprio_motu, but only with the authorisation of a Pre-Trial Chamber. This was a significant concession to judicial control, and it has since been used in situations including Kenya, Georgia, Bangladesh and Myanmar, Afghanistan, and the Philippines. The second was the relationship with the Security Council. The Council obtained two powers: it may refer a situation to the Court under Chapter VII of the United Nations Charter, even where the state concerned is not a party, and it may defer an investigation for twelve months, renewably, under Article 16. Referral has been used twice, for Darfur in 2005 and for Libya in 2011. Deferral has never been successfully invoked, though it has been requested. The third was the crime of aggression. Many delegations wanted it included; others feared that criminalising the use of force would draw the Court into the political heart of interstate relations. The result was a placeholder. Aggression was listed in Article 5 as a crime within the Court's jurisdiction, but jurisdiction was suspended until a definition and conditions were agreed. That agreement came only in 2010 at the Kampala Review Conference, and the amendments were activated in July 2018 with jurisdictional restrictions so tight that they largely explain the Court's inability to address the invasion of Ukraine. The fourth was #complementarity itself. Developing and developed states alike were unwilling to hand primary jurisdiction to an international body. The Preamble and Articles 1, 17, and 18 therefore state that the Court is complementary to national criminal jurisdictions. States keep the primary right and duty to prosecute. The Statute was adopted by a vote of 120 in favour, 7 against, and 21 abstentions. The votes were not officially recorded by name, but the United States, China, and Israel publicly opposed the text. The United States signed in December 2000 and then, in 2002, informed the depositary that it did not intend to become a party, a step often described as unsigning. INSTITUTIONAL ARCHITECTURE OF THE COURT The Court is composed of four organs, and understanding their separation is essential to understanding how criticism of the Court is often misdirected. The #Presidency consists of the President and two Vice-Presidents, elected by the judges. It handles the administration of the Court, other than the Office of the Prosecutor, and external relations. The Chambers are organised in three divisions: Pre-Trial, Trial, and Appeals. There are eighteen judges, elected by the #Assembly_of_States_Parties for non-renewable terms of nine years, with rules designed to secure geographical representation, a fair representation of men and women, and expertise in criminal law and procedure or in relevant areas of international law. The Pre-Trial Chamber issues arrest warrants and summonses, authorises proprio motu investigations, and decides whether to confirm charges. The Trial Chamber conducts the trial and, if there is a conviction, sentences and orders reparations. The Appeals Chamber reviews decisions of law and fact. The #Office_of_the_Prosecutor is independent. It receives referrals and information, conducts preliminary examinations, decides whether there is a reasonable basis to proceed, investigates, and prosecutes. It is important to note that the Prosecutor does not decide who is convicted. Journalistic language such as the Court has charged someone often blurs the distinction between the Prosecutor's application and the judicial decision of a Pre-Trial Chamber to issue a warrant. The Chamber must find reasonable grounds to believe that the person committed the crime. The #Registry provides judicial support: courtroom management, defence and victim support, witness protection, detention, and outreach. Two further bodies sit alongside the Court. The #Assembly_of_States_Parties is the management and legislative body. It elects judges and the Prosecutor, adopts the budget, and amends the Rules. It cannot instruct chambers. The #Trust_Fund_for_Victims is a separate institution with two mandates: implementing reparations ordered by chambers, and providing assistance to victims in situations under investigation using voluntary contributions, independently of any conviction. The distinction between the reparations mandate and the assistance mandate matters practically. Convicted persons before the ICC have almost always been declared indigent, which means that court-ordered reparations depend on voluntary contributions to the Trust Fund. This is the structural reason why reparations awards, though large on paper, are slow to reach victims. JURISDICTION AND TRIGGER MECHANISMS The Court's authority is limited in four dimensions, and most public confusion about the ICC arises from ignoring one of them. Subject-matter jurisdiction (ratione materiae) covers genocide, crimes against humanity, war crimes, and aggression. It does not cover terrorism as such, drug trafficking, corruption, or environmental destruction as free-standing offences, although conduct in those categories can sometimes be prosecuted if it satisfies the elements of a core crime. Temporal jurisdiction (ratione temporis) begins on 1 July 2002. For a state that joins later, jurisdiction begins when the Statute enters into force for that state, unless the state makes a declaration under Article 12(3) accepting jurisdiction for earlier conduct after 1 July 2002. Personal and territorial jurisdiction (ratione personae and ratione loci) is set by Article 12. The Court may exercise jurisdiction if the conduct occurred on the territory of a state party, or if the accused is a national of a state party. Either link suffices. This #territoriality principle is not exotic; it is the ordinary basis on which #national_criminal_law operates. A French national who commits a crime in Germany is tried under German law, and no one considers this an assault on French sovereignty. The same logic explains why the nationals of non-party states can fall within the Court's reach when they act on the territory of a state party, and it underlies the Court's jurisdiction in the situations of Afghanistan, Bangladesh and Myanmar, Ukraine, and Palestine. It is also the point most vigorously contested by states that are not parties. There is one exception to the Article 12 preconditions: a Security Council referral under Article 13(b), which can bring within the Court's reach conduct on the territory of, or by nationals of, any state. Three trigger mechanisms open a situation. First, a state party may refer a situation, including its own, which is known as a self-referral and was used by Uganda, the Democratic Republic of the Congo, the Central African Republic, and Mali. #Self_referral has been criticised because it allows a government to invite the Court to investigate its opponents. Second, the Security Council may refer a situation. Third, the Prosecutor may act proprio motu with Pre-Trial Chamber authorisation. Finally, Article 17(1)(d) requires that a case be of sufficient #gravity. This is a filter, not a formality. The Prosecutor has used it to decline cases that formally fall within jurisdiction but involve a small number of incidents. THE CORE CRIMES 8.1 Genocide Article 6 reproduces the definition in the 1948 Genocide Convention. Genocide consists of specified acts, including killing, causing serious bodily or mental harm, and forcibly transferring children, committed with intent to destroy, in whole or in part, a national, ethnical, racial, or religious group as such. Two features make #genocide the hardest crime to prove. The first is the special intent, or dolus specialis: the perpetrator must intend the destruction of the group, not merely intend to kill individuals who belong to it. The second is the closed list of protected groups. Political groups and social classes are excluded, which is why the mass killings associated with political repression are ordinarily charged as crimes against humanity rather than genocide. Students frequently assume that genocide is the most serious crime and that a prosecutor should always charge it. In practice, prosecutors often prefer crimes against humanity because the evidentiary burden is lower and the sentencing range is comparable. There is no formal hierarchy of gravity among the core crimes in the Statute. 8.2 Crimes against humanity Article 7 defines #crimes_against_humanity as certain acts committed as part of a widespread or systematic attack directed against any civilian population, with knowledge of the attack. The listed acts include murder, extermination, enslavement, deportation or forcible transfer, imprisonment, torture, rape and other forms of sexual violence, persecution, enforced disappearance, apartheid, and other inhumane acts. Three elements deserve emphasis. The attack must be widespread or systematic, not necessarily both; the disjunctive is deliberate. There must be a state or organisational policy, though the policy need not be formal or written and may be inferred from a pattern of conduct. And, unlike war crimes, crimes against humanity do not require an armed conflict. This is why the Court can address the killings associated with the so-called war on drugs in the Philippines, a situation with no armed conflict of the classic type. Article 7 is also notable for its treatment of sexual and gender-based crimes. Rape, sexual slavery, enforced prostitution, forced pregnancy, and enforced sterilisation are listed expressly. This was a drafting achievement of the 1990s, built on the jurisprudence of the ad hoc tribunals, and it was given practical effect in the Ongwen case, where the accused was convicted of forced marriage as an other inhumane act and of forced pregnancy, the first conviction of its kind before the Court. 8.3 War crimes Article 8 is the longest provision in the Statute. It divides war crimes into four categories: grave breaches of the 1949 Geneva Conventions in international armed conflict; other serious violations of the laws and customs applicable in international armed conflict; serious violations of Common Article 3 in non-international armed conflict; and other serious violations applicable in non-international armed conflict. The list includes wilful killing, torture, taking hostages, intentionally directing attacks against civilians or civilian objects, attacks on protected buildings including hospitals and places of worship, pillage, conscripting or enlisting children under fifteen, and #starvation of civilians as a method of warfare. The starvation offence, long neglected, became central to the arrest warrants issued in November 2024 in the Palestine situation. The threshold clause in Article 8(1) states that the Court has jurisdiction in particular when war crimes are committed as part of a plan or policy or as part of a large-scale commission. This is not a strict jurisdictional requirement, but it signals that the Court is not intended to try isolated battlefield offences, which remain the responsibility of national military justice systems. 8.4 The crime of aggression Article 8 bis, added at Kampala in 2010, defines the crime of #aggression as the planning, preparation, initiation, or execution by a person in a position effectively to exercise control over or to direct the political or military action of a state, of an act of aggression which by its character, gravity, and scale constitutes a manifest violation of the Charter of the United Nations. It is a leadership crime. A soldier cannot commit it; only those at the apex of the state can. The definition of an act of aggression draws on General Assembly Resolution 3314 of 1974 and includes invasion, bombardment, blockade, and the sending of armed bands. The jurisdictional regime, however, is markedly narrower than for the other three crimes. Under Article 15 bis, where a situation is not referred by the Security Council, the Court may not exercise jurisdiction over aggression committed by nationals of, or on the territory of, a state that is not a party to the Statute. States parties may also opt out by declaration. The practical result is that the Court cannot try Russian leaders for aggression against Ukraine, because Russia is not a party and a Security Council referral would be vetoed. Section 15 returns to this problem. COMPLEMENTARITY AND ADMISSIBILITY Complementarity is the organising principle of the whole system, and it is worth stating carefully because it is often described incorrectly. The Preamble affirms that the Court shall be complementary to national criminal jurisdictions, and Article 1 repeats the point. Article 17 operationalises it. A case is inadmissible if it is being investigated or prosecuted by a state which has jurisdiction, unless that state is unwilling or unable genuinely to carry out the proceedings; if it has been investigated and the state decided not to prosecute, subject to the same caveat; if the person has already been tried for the same conduct, reflecting the principle of ne bis in idem in Article 20; or if the case is not of sufficient gravity. The key words are unwilling and unable. Unwillingness is assessed by asking whether proceedings are being conducted for the purpose of shielding the person, whether there has been unjustified delay inconsistent with an intent to bring the person to justice, and whether proceedings are independent and impartial. Inability is assessed by asking whether, due to a total or substantial collapse or unavailability of the national judicial system, the state cannot obtain the accused, the evidence, or otherwise carry out proceedings. Two points follow. First, complementarity is assessed case by case, not situation by situation. The fact that the Court is investigating a situation does not deprive national authorities of the right to prosecute other cases arising from the same events. Second, a national proceeding must cover substantially the same conduct and substantially the same person to displace the Court. A sham trial on lesser charges will not. Complementarity has a second, softer dimension often called #positive_complementarity: the idea that the Court and its Office of the Prosecutor should actively encourage and support national prosecutions, through information sharing, capacity building, and cooperation with national authorities. The Office issued a dedicated policy on complementarity and cooperation in 2024, and the direction of travel is clear. The Court, with an annual budget of a few hundred million euros and a handful of trials at a time, cannot possibly try more than a tiny fraction of atrocity crimes. Its realistic ambition is to be the apex of a decentralised system, not its whole substance. The evidence on whether this works is mixed. Uganda enacted an International Criminal Court Act in 2010 and created an International Crimes Division in its High Court, which convicted the LRA commander Thomas Kwoyelo of war crimes in 2024. In the Central African Republic, a hybrid Special Criminal Court operates alongside the ICC and has issued its own verdicts. These are genuine catalytic effects. But Wierda (2023) cautions that domestic legal reform driven by international pressure can remain formal, producing statutes and courtrooms without producing prosecutions. INDIVIDUAL RESPONSIBILITY, IMMUNITY, AND FAIR TRIAL 10.1 Modes of liability Article 25 sets out individual criminal responsibility. A person may be liable for committing a crime individually, jointly with another, or through another person, regardless of whether that other person is criminally responsible; for ordering, soliciting, or inducing; for aiding, abetting, or otherwise assisting; for contributing to a crime committed by a group acting with a common purpose; and, in the case of genocide, for direct and public incitement. The Court has developed a doctrine of #co_perpetration based on control over the crime, drawn substantially from German criminal theory. A person who does not personally pull a trigger may nonetheless be a principal if he or she exercises control over the criminal apparatus. This doctrine allows the Court to reach leaders while maintaining that they are not merely accessories to the acts of their subordinates. It has been criticised as textually strained, but it has proved durable. In the Yekatom and Ngaissona judgment of July 2025, the Trial Chamber convicted one accused as a perpetrator combining direct and indirect forms of participation and the other only as an accessory who aided and abetted, which illustrates that the distinction has real consequences for sentence. Article 28 adds #command_responsibility. A military commander is responsible for crimes committed by forces under his effective command and control where he knew or, owing to the circumstances at the time, should have known that the forces were committing or about to commit crimes, and failed to take all necessary and reasonable measures within his power to prevent or repress them or to submit the matter to the competent authorities. A civilian superior is held to a somewhat stricter mental standard. The doctrine is a form of liability for omission, and it is the principal legal bridge between atrocities on the ground and offices in the capital. 10.2 Immunity: Articles 27 and 98 Article 27 provides that the Statute applies equally to all persons without distinction based on official capacity, and that immunities attaching to the official capacity of a person, whether under national or international law, do not bar the Court from exercising jurisdiction. Article 98(1), by contrast, provides that the Court may not proceed with a request for surrender that would require the requested state to act inconsistently with its obligations under international law with respect to the state or diplomatic #immunity of a person or property of a third state, unless the Court first obtains that third state's cooperation for a waiver. The relationship between these provisions is the most contested question in the Court's law. It arose in the Al Bashir litigation, where states parties failed to arrest the then Sudanese president, and it has arisen again in relation to the arrest warrants issued in November 2024 against Israeli officials, since Israel is not a party. In its 2019 Jordan judgment the Appeals Chamber held that there is no head of state immunity before an international court and that Article 27(2) reflects that position; the reasoning drew in part on the Security Council referral of the Darfur situation, which has left commentators uncertain how far the holding extends. In its 2024 decision concerning Mongolia's failure to arrest a visiting head of state, a Pre-Trial Chamber reaffirmed that Article 98 does not displace Article 27 and characterised Article 98 as a procedural provision concerning state and diplomatic immunity rather than a shield for individuals. Sabel (2024) presents the contrary view forcefully: on the traditional understanding, personal immunity attaches to serving senior officials of non-party states, and executing a warrant against them without a waiver would breach the requested state's obligations to that state. The Court has not accepted this view, but many states behave as if they had. Hungary did not arrest a visiting official subject to a warrant in April 2025, and announced its withdrawal from the Statute during the visit, a notification it subsequently withdrew, with effect from 29 May 2026, leaving the number of states parties at 125. 10.3 Fair trial rights The Statute is not only an instrument of prosecution. Articles 55, 66, and 67 protect the rights of suspects and accused persons: the #presumption_of_innocence, the right to counsel, the right to remain silent without inference of guilt, the right to examine witnesses, and the right to disclosure of exculpatory evidence. The Court has taken these obligations seriously enough to acquit. Ngudjolo Chui was acquitted in 2012, Gbagbo and Ble Goude were acquitted in 2019, and Bemba's conviction was reversed on appeal in 2018. Acquittals are often presented in the press as institutional failures. They are better read as evidence that the standard of proof is doing its work. VICTIMS, PARTICIPATION, AND REPARATIONS One of the genuine innovations of the Rome Statute is the position of #victims. Before the ICC, victims of international crimes appeared before international tribunals almost exclusively as witnesses, that is, as instruments of the prosecution case. #Article_68 allows victims to present their views and concerns at stages of the proceedings determined appropriate by the Court, through legal representatives, where their personal interests are affected. #Article_75 empowers the Court to establish principles relating to reparations, including restitution, compensation, and rehabilitation, and to make an order directly against a convicted person. Article 79 establishes the Trust Fund for Victims. The practice has produced a substantial body of reparations law across the Lubanga, Katanga, Al Mahdi, Ntaganda, Ongwen, and Al Hassan cases. The most significant order to date came in the Ongwen case in February 2024, when Trial Chamber IX set the convicted person's liability at EUR 52,429,000 for approximately 49,772 victims, combining collective community-based rehabilitation programmes with a symbolic individual award of EUR 750 per eligible victim. The Appeals Chamber confirmed the order in April 2025, making it final. The order is the largest in the Court's history, and it exposes the structural problem precisely. Ongwen is indigent. The money must come from voluntary contributions to the Trust Fund, which launched a funding appeal for an initial portion of the total. Meanwhile the beneficiaries are ageing, and some have died. The scholarly assessment is sober. Atingo, Allen, and Macdonald (2024), writing from long fieldwork in northern Uganda, describe a betrayal of expectations: the women whose testimony grounded the historic convictions for forced marriage and forced pregnancy report feeling used and under-supported. Fisher (2024) documents the reverse problem in the Barlonyo community, whose massacre was not among the charges, and whose members therefore fell outside the reparations scheme entirely. This is the phenomenon sometimes called a #hierarchy_of_victimhood: because reparations follow the charges, and the charges follow prosecutorial selection, the selection of charges silently determines who counts as a victim. There is no clean solution within the current framework. Widening charges lengthens trials; narrowing them excludes victims. The most that can be said is that the Court's assistance mandate, which does not depend on conviction, partially mitigates the gap, and that honest communication about what reparations can deliver would reduce the disappointment documented in the fieldwork literature. THE PRACTICE OF THE COURT, 2002 TO 2026 A fair assessment requires an accurate record. The Court has been slower and smaller than its founders hoped, but it is not the empty vessel its critics describe. The Court has opened investigations in situations including Uganda, the Democratic Republic of the Congo, Darfur, the Central African Republic, Kenya, Libya, Cote d'Ivoire, Mali, Georgia, Burundi, Bangladesh and Myanmar, Afghanistan, Palestine, the Philippines, Venezuela, and Ukraine. The early concentration on African situations was real, and it did enormous political damage. It is no longer an accurate description of the docket. Convictions for core crimes include Thomas Lubanga Dyilo (2012, conscription and use of child soldiers, fourteen years); Germain Katanga (2014, accessory to war crimes and crimes against humanity, twelve years); Ahmad Al Faqi Al Mahdi (2016, the first conviction for the war crime of intentionally directing attacks against protected cultural buildings in Timbuktu, nine years, following a guilty plea); Bosco Ntaganda (2019, thirty years, the longest sentence to date); Dominic Ongwen (2021, sixty-one counts, twenty-five years, upheld on appeal in December 2022); Al Hassan Ag Abdoul Aziz (2024, war crimes and crimes against humanity in Timbuktu, ten years); and Alfred Yekatom and Patrice-Edouard Ngaissona (24 July 2025, the Court's first judgment on the Central African Republic, fifteen and twelve years respectively, with sentence delivered alongside the judgment for the first time in the Court's history). There have also been convictions for offences against the administration of justice under Article 70, principally witness tampering in the Bemba case. The most consequential recent developments concern #heads_of_state and government. In March 2023, Pre-Trial Chamber II issued arrest warrants against the Russian president and a Russian official in connection with the unlawful deportation of children from occupied Ukraine. In November 2024, Pre-Trial Chamber I issued warrants against the Israeli prime minister and former defence minister for the war crime of starvation as a method of warfare and crimes against humanity including murder, persecution, and other inhumane acts, and against a Hamas commander. In March 2025, the former president of the Philippines was arrested in Manila and surrendered to the Court, the first Asian former head of state in its custody; the confirmation of charges hearing was held in February 2026, all three counts of crimes against humanity were confirmed on 23 April 2026, and trial was scheduled to open on 30 November 2026. Set against this, the enforcement record remains weak. Warrants issued in 2005 against Joseph Kony remain unexecuted; the Court does not try defendants in absentia, so the Kony case has never advanced to trial. Al Bashir was never surrendered by Sudan. Several suspects in the Libya and Darfur situations remain at large. Of the individuals publicly wanted by the Court, a substantial number have never been arrested. The pattern is consistent: the Court can adjudicate whenever it has a suspect in the dock, and it can adjudicate to a high standard. What it cannot do is produce the suspect. COOPERATION AND THE ENFORCEMENT GAP Part 9 of the Statute obliges states parties to cooperate fully with the Court, including in the arrest and surrender of persons, the collection of evidence, and the protection of witnesses. Article 87(7) allows a chamber to make a finding of non-compliance and refer the matter to the Assembly of States Parties or, where the situation was referred by the Security Council, to the Council. This is the Court's structural weakness, and it is worth stating plainly: the ICC has no police, no army, and no enforcement power of its own. It is entirely dependent on states. When states cooperate, as the Philippines did in March 2025, an arrest happens within days. When they do not, warrants sit unexecuted for two decades. Findings of non-compliance have been made against several states, including South Africa, Jordan, and Mongolia in relation to visiting suspects. The consequences have been declaratory. The Assembly of States Parties has no coercive tools, and the Security Council has never acted on a referral of non-compliance in the Darfur situation. The result is what might be called the #enforcement_gap. The Court's jurisprudence has advanced far ahead of its practical reach. Every doctrinal victory on immunity is worth exactly as much as the willingness of the state where the suspect lands to act on it. Three responses have been proposed. The first is legal: states should adopt implementing legislation that makes arrest and surrender automatic and removes domestic immunity defences. Many states parties still lack adequate implementing legislation, a point stressed in the practitioner literature. The second is political: states parties should collectively raise the cost of non-cooperation, for example through diplomatic and travel consequences. The third is operational: the Office of the Prosecutor should sequence its warrants strategically, seeking sealed warrants where the prospect of arrest depends on surprise. The Duterte arrest, executed on a warrant issued under seal and reclassified as public only after his surrender, is the clearest recent illustration that this works. POLITICS, LEGITIMACY, AND THE CHARGE OF SELECTIVITY 14.1 The African critique For much of the Court's first fifteen years, every open investigation concerned an African state. Some of these were self-referrals by African governments, and 33 African states remain parties, more than any other regional group. But the optics were terrible, and the substance of the complaint was never only about optics. The charge, made forcefully by the African Union in the Al Bashir and Kenyatta episodes, was that the Court prosecuted the leaders of weak states while the nationals of powerful states remained untouchable. Ba (2023) traces how this critique fed the Malabo Protocol, an African Union project to give a regional court criminal jurisdiction, notably including an immunity clause for serving heads of state that is incompatible with Article 27. The Protocol has not entered into force, but it stands as a statement of dissatisfaction. #Burundi withdrew in 2017. The Philippines withdrew in 2019, although the Appeals Chamber confirmed in April 2026 that the Court retains jurisdiction over crimes committed before the withdrawal took effect. In September 2025, the three Sahel states of Burkina Faso, Mali, and Niger announced their intention to withdraw, describing the Court as an instrument of neo-colonial repression, and they deposited their instruments in June 2026, with effect one year later under Article 127. Withdrawal does not affect obligations arising while the state was a party, nor does it discontinue proceedings already begun. 14.2 Great power resistance The critique of selectivity did not remain an African argument. It has been adopted, in a very different key, by the states most resistant to the Court. Russia signed but never ratified the Statute, and withdrew its signature in 2016. In December 2025, following the warrants concerning the deportation of Ukrainian children, a Russian court convicted the Prosecutor and eight ICC judges in absentia. The United States has never been a party. Its opposition has been consistent across administrations of both parties, though its intensity has varied. In February 2025, #Executive_Order_14203 authorised asset freezes and entry bans against ICC officials and those materially supporting the Court's investigations of United States or allied nationals. Designations followed against the Prosecutor, deputy prosecutors, and judges, in successive rounds during 2025. Sterio (2026) analyses the regime and argues that its operational effects on the Court are severe and that its broader consequence is to signal that a powerful state may lawfully coerce an independent judicial institution. United Nations human rights experts have described the measures as an assault on judicial independence. Litigation in United States courts has produced partial relief for United States persons on constitutional grounds, but the designations of Court officials remain. 14.3 How should this be evaluated? It is possible to hold two thoughts at once, and students should resist the temptation to collapse them. The first is that the selectivity critique has a factual basis. The Court's jurisdiction is structurally uneven, because the Security Council can extend jurisdiction to non-parties but three permanent members are not parties and can veto referrals against themselves and their allies. That is not an accident of practice; it is a feature of the Statute's design, accepted in Rome as the price of Security Council acquiescence. The second is that structural unevenness is not a reason to abandon accountability where it is possible. The argument that no one should be prosecuted until everyone can be prosecuted has the effect, whatever its intent, of protecting the powerful. Garcia Iommi (2023) offers the useful empirical corrective that overall state support for the Court has been more continuous than the crisis narrative suggests: loud exits and loud attacks coexist with quiet, routine cooperation by most states parties most of the time. THE CRIME OF AGGRESSION AND THE UKRAINE PROBLEM The invasion of Ukraine in February 2022 exposed the Kampala compromise with unusual clarity. The ICC has jurisdiction over war crimes, crimes against humanity, and genocide committed on Ukrainian territory, because Ukraine accepted jurisdiction by declaration and later became a state party. But it has no jurisdiction over the crime of #aggression by Russian leaders, because Article 15 bis excludes aggression committed by nationals of a non-party state, and a Security Council referral would face a Russian veto. This is not a gap in the definition. It is a gap in jurisdiction, deliberately created. Dannenbaum (2022) set out the resulting options and the case for a special tribunal; McDougall (2023) argued that a special international tribunal was the only viable path to prosecuting the aggression, and that failing to pursue it would further weaken the prohibition on the use of force. Heller (2022) argued from the other side that a special tribunal is a bad idea, in part because it revives the charge of victors' justice and further entrenches double standards. Ambos (2022) pressed the double standards point directly, asking why the invasion of Iraq did not produce a comparable tribunal. The political answer arrived anyway. On 25 June 2025 the Council of Europe and Ukraine signed an agreement in Strasbourg to establish a #Special_Tribunal for the Crime of Aggression against Ukraine, with a statute annexed, grounded in Ukraine's own jurisdiction over crimes committed on its territory. In May 2026, an enlarged partial agreement establishing the tribunal's steering committee was adopted, with participation by a group of Council of Europe member states and others. Meanwhile, the states parties addressed the underlying defect. At a special session of the Assembly in July 2025, a proposal to align the jurisdictional regime for aggression with that applicable to the other three crimes attracted wide support, but a minority insisted on delay, and the final decision was deferred to 2029 (Kress, 2025). Two lessons follow. First, the Statute's design choices have consequences measured in decades. The narrow aggression regime, agreed to secure consensus in 2010, is the direct cause of the institutional improvisation of 2025 and 2026. Second, ad hoc tribunals have not been superseded by the ICC. When the permanent court cannot act, states build something else, and the objection that the resulting institution is selective is precisely the objection the ICC was created to answer. IMPACT, DETERRENCE, AND WHAT THE EVIDENCE SHOWS Does the Rome Statute prevent atrocities? This is the question that matters most and that the law itself cannot answer. The empirical literature offers cautious support for a conditional effect. Miller (2023) finds that ICC indictments are associated with reductions in atrocities by groups affiliated with the indictee, and explains the mechanism through the concept of assurance: groups believe they can reduce the costs imposed by an indictment if they stop attacking civilians. McAllister (2023) finds that the more actions the Court takes against suspected perpetrators during an ongoing conflict, the more likely rebel restraint becomes, while cautioning that the overall record is uneven and that deterrence claims should not be overstated. Three qualifications are necessary. First, deterrence requires perceived probability of punishment, and the ICC's probability of punishment is objectively low. Very few suspects are ever tried. Where deterrence appears to operate, it is likely mediated through reputation and through the calculations of groups seeking legitimacy, rather than through fear of arrest. Second, the Court's most important effects may be #expressive rather than deterrent. The Statute changed the vocabulary in which atrocity is discussed. National legislatures adopted implementing statutes; #universal_jurisdiction prosecutions in Europe drew on Rome Statute definitions; military manuals were rewritten. This normative diffusion is difficult to quantify and easy to dismiss, but it is real. Third, there is an unresolved concern that international prosecution may prolong conflict by removing exit options for leaders. The empirical support for this claim is weaker than its rhetorical popularity, but it is not zero, and honest scholarship acknowledges the tension between #peace_versus_justice as competing short-term goods. Wierda (2023) supplies the most useful corrective for students inclined to either extreme. Studying four situations, she finds that the Court's local impact is neither the transformation promised by its advocates nor the irrelevance alleged by its critics. It is partial, mediated by domestic politics, and often disappointing to those it claims to serve. REFORM: WHAT COULD ACTUALLY BE DONE Reform proposals for the ICC are abundant. The following are the most substantial, and each carries a cost. Amend the aggression regime. Aligning Article 15 bis with the jurisdictional regime for the other crimes would remove the anomaly that the gravest form of interstate wrongdoing is the hardest for the Court to reach. Wide support existed at the July 2025 special session; a minority secured deferral to 2029. The cost is that some states parties fear exposure of their own use of force. Strengthen cooperation. The single highest-return reform is unglamorous: #implementing_legislation, standing arrest procedures, and voluntary agreements on witness relocation and enforcement of sentences. The Court can do nothing about the enforcement gap alone. Fund the Trust Fund for Victims properly. A reparations order that cannot be paid damages the Court more than no order at all, because it converts a judicial promise into a broken one. Predictable funding, whether through assessed contributions or a dedicated mechanism, would address the pattern documented in the Ongwen literature. Improve prosecutorial performance and internal governance. The #Independent_Expert_Review of 2020 catalogued serious problems in investigation quality, staff management, and case selection, and the Office has since revised its policies. Continued attention is needed. Several acquittals and collapsed cases were caused not by judicial hostility but by weak evidence. Address selectivity honestly. The Court cannot fix the Security Council. It can be transparent about case selection criteria, and it can pursue situations outside Africa, which it now does. The warrants concerning Ukraine and Palestine have not silenced the selectivity critique, but they have made it harder to argue that the Court only pursues the weak. They have also, predictably, provoked the most serious institutional retaliation in its history. Protect the institution against coercion. Sterio (2026) argues that sanctions on judges and prosecutors are a category of attack the Statute never anticipated. The Assembly of States Parties has begun to characterise such conduct as an offence against the administration of justice within the meaning of Article 70. Whether states parties will adopt counter-measures, such as blocking statutes shielding those who cooperate with the Court, remains to be seen. LIMITATIONS OF THIS STUDY This article is doctrinal and interpretive. It does not conduct original empirical research, and its claims about deterrence are reported from the political science literature rather than independently tested. Its account of the Court's docket is current to mid-2026 and will be overtaken by events, particularly in the Palestine, Ukraine, and Philippines situations. It relies on English-language scholarship, which under-represents work published in French, Spanish, Arabic, and Russian, and it does not draw on the perspectives of affected communities except as reported in the fieldwork literature. Finally, the article deliberately avoids taking a position on the merits of any pending case; the individuals named as subject to arrest warrants are suspects, and the presumption of innocence applies to them as it applies to everyone else. CONCLUSION The Rome Statute is best understood as an unfinished settlement rather than a completed achievement. Measured against its own preamble, which speaks of ending #impunity for the most serious crimes of concern to the international community as a whole, the record is modest. The Court has secured a small number of convictions in nearly a quarter of a century. Its warrants against the powerful go unexecuted. Its reparations orders outrun its funds. Its budget is smaller than that of a mid-sized national police force. Measured against the alternative that existed before 1998, the record looks different. Before Rome, there was no permanent institution to which a state could refer a situation, no standing prosecutor, no agreed treaty definition of crimes against humanity, no formal role for victims, and no expectation that a former head of state would ever be arrested and flown to The Hague to answer for the deaths of his own citizens. In March 2025, exactly that happened, and in April 2026 the charges were confirmed. The Court's central weakness is not doctrinal. Its jurisprudence on complementarity, modes of liability, and the crimes themselves is sophisticated and, for the most part, defensible. Its weakness is structural: it depends entirely on the cooperation of states, and the states best able to shield perpetrators are the ones least subject to its reach. That is a defect in the international order, not merely in the treaty. It cannot be cured by better drafting. What the Statute did achieve is worth naming precisely. It made #accountability a default expectation rather than an exception, it created a permanent forum that does not have to be re-invented after each catastrophe, and it gave victims a legal identity. Whether that expectation survives the current period of sanctions, withdrawals, and open hostility will depend on the 125 states that remain parties, and on whether they are willing to pay the political price of the institution they built. The Rome Statute is not a guarantee of justice. It is an argument that justice is owed. Whether the argument prevails is not, in the end, a legal question. HASHTAGS #Rome_Statute #International_Criminal_Court #ICC #international_criminal_law #genocide #crimes_against_humanity #war_crimes #crime_of_aggression #complementarity #international_justice #accountability #human_rights_law #The_Hague #transitional_justice #atrocity_crimes REFERENCES Ambos, K. (Ed.). (2022). Rome Statute of the International Criminal Court: Article-by-Article Commentary (4th ed.). C. H. Beck, Hart, and Nomos. Ambos, K. (2022). Ukraine and the double standards of the West. Journal of International Criminal Justice, 20(4), 875-891. Atingo, J., Allen, T., and Macdonald, A. (2024). Witnessing Ongwen: A betrayal of expectations? Journal of International Criminal Justice, 22(5), 785-802. https://doi.org/10.1093/jicj/mqae029 Ba, O. (2023). Exit from Nuremberg to The Hague: The Malabo Protocol and the pan-African road to Arusha. Global Studies Quarterly, 3(3). Broache, M. P., Cronin-Furman, K., Mendeloff, D., and McAllister, J. R. (2023). The International Criminal Court at 25: Reassessing processes and outcomes. Journal of Human Rights, 22(1), 1-15. Dannenbaum, T. (2022). A special tribunal for the crime of aggression? Journal of International Criminal Justice, 20(4), 859-873. https://doi.org/10.1093/jicj/mqac047 Fisher, K. J. (2024). The Barlonyo massacre, Ongwen's trial, and ICC reparations: Reflections on the dynamics of expectations and disappointment. Journal of International Criminal Justice, 22(5), 803-813. https://doi.org/10.1093/jicj/mqae007 Garcia Iommi, L. (2023). Nothing changed after Rome: Continuity in state support for the International Criminal Court. Journal of Human Rights, 22(1), 16-33. Heller, K. J. (2022). Options for prosecuting Russian aggression against Ukraine: A critical analysis. Journal of Genocide Research. Kress, C. (2025). On recent events concerning the crime of aggression. Journal of International Criminal Justice, 23(3-4), 405-408. https://doi.org/10.1093/jicj/mqaf040 McAllister, J. R. (2023). Casting a shadow over war zones? Hard truths about the ICC's efforts to deter wartime atrocities. Journal of Human Rights, 22(1), 34-52. https://doi.org/10.1080/14754835.2022.2156275 McDougall, C. (2023). The imperative of prosecuting crimes of aggression committed against Ukraine. Journal of Conflict and Security Law, 28(2), 203-230. https://doi.org/10.1093/jcsl/krad004 Miller, A. C. (2023). Without an army: How ICC indictments reduce atrocities. Journal of Peace Research, 60(4), 573-587. https://doi.org/10.1177/00223433221088692 Phooko, M. R. (2023). Existential threats to the International Criminal Court: Making sense of the convergence of disparate geo-political interests and ideological positions. Journal of Law, Society and Development, 8. Robinson, D., Vasiliev, S., van Sliedregt, E., and Oosterveld, V. (2024). An Introduction to International Criminal Law and Procedure (5th ed.). Cambridge University Press. https://doi.org/10.1017/9781009466660 Sabel, R. (2024). Israel, the ICC, and the arrest warrants against Netanyahu and Gallant. Israel Journal of Foreign Affairs, 18(3). https://doi.org/10.1080/23739770.2024.2440838 Sterio, M. (2026). United States sanctions against the International Criminal Court: A threat to the global pursuit of justice. Journal of International Criminal Justice. https://doi.org/10.1093/jicj/mqag028 Wierda, M. (2023). The Local Impact of the International Criminal Court: From Law to Justice. Cambridge University Press. https://doi.org/10.1017/9781009153805

The North Atlantic Treaty, signed in Washington on 4 April 1949, is one of the shortest and most consequential security agreements of the twentieth century. Fourteen articles, running to barely more than a thousand words, created the legal basis for the #North_Atlantic_Treaty_Organization, an alliance that now binds thirty two states in North America and Europe. This article examines the treaty as three things at once: a legal instrument, a political bargain and the seed of an #international_organisation that has outlived the threat it was written to meet. Using a qualitative, document centred approach, the study reads the treaty text against the diplomatic record of 1947 to 1949, and then traces how the meaning of its central promise has been reworked over seventy seven years of practice. The analysis makes four claims. First, the treaty was deliberately written to be vague, because vagueness was the price of American ratification. Second, the commitment in Article 5 has never been self executing, and its force has always come from military presence, planning and habit rather than from the words alone. Third, the alliance survived the disappearance of the Soviet Union because it had become an institution with assets and routines, not merely a pact against a single enemy. Fourth, the pressures visible in 2025 and 2026, from the five per cent spending pledge agreed at #The_Hague_Summit to the American force posture reductions debated before the #Ankara_Summit, are not new departures but sharper versions of arguments that were present in the drafting room in 1949. The article is written for students and closes with directions for further research. Keywords: North Atlantic Treaty; NATO; collective defence; Article 5; deterrence; alliance politics; transatlantic security; burden sharing; treaty interpretation; international organisation 1. Introduction On a spring morning in 1949, foreign ministers from twelve countries gathered in the Departmental Auditorium in Washington and put their names to a document that most of them privately doubted would hold. The United States had never signed a peacetime military alliance with European powers. Its own founding tradition warned against exactly that. Yet within four years of the end of the Second World War, Washington bound itself to the defence of Western Europe, and the #North_Atlantic_Treaty became the legal skeleton of what would grow into the most heavily institutionalised alliance in modern history. This #research_article asks a simple question with a complicated answer. What exactly did the treaty of 1949 create, and why has that creation lasted so long? The question matters for students of international relations, law and history for several reasons. The first is empirical. #NATO is still the central security organisation of the Euro Atlantic area. In July 2026 its leaders met in Ankara and again described their obligation under #Article_5 as ironclad. Understanding what that obligation actually says, and what it does not say, is basic knowledge for anyone studying European security. The second reason is conceptual. The treaty is a working example of a puzzle that runs through alliance theory. States promise to fight for each other, but promises between sovereign states cannot be enforced by any higher authority. So how do such promises acquire credibility? The 1949 text is a good place to study that problem, because its drafters were painfully aware of it and left traces of their anxiety in the wording. The third reason is contemporary. Since 2022, when Russia launched its full scale invasion of Ukraine, the alliance has been through the most demanding period since the early Cold War. Two long neutral states, Finland and Sweden, joined. Defence budgets rose. At the same time, doubts about the reliability of the United States, the largest contributor by far, have grown louder rather than quieter. Anyone who wants to make sense of the arguments now taking place in European capitals needs to know where the original bargain came from. The article proceeds in the following way. Section 2 reviews the scholarly literature and identifies a gap. Section 3 sets out the research design. Section 4 reconstructs the diplomatic path from 1945 to 1949. Section 5 reads the treaty clause by clause. Section 6 explains how a treaty became an organisation between 1950 and 1991. Section 7 covers the post Cold War reinvention. Section 8 examines the return of #collective_defence after 2014. Section 9 brings the story to 2026. Section 10 develops four analytical themes: burden sharing, #credibility, consensus and enlargement. Section 11 discusses what the 1949 text can and cannot explain. Section 12 states the limitations of the study. Section 13 concludes. A note on tone. The article is written in plain English because it is intended for students, including students who are not native English speakers. Plain writing is not the same as simple thinking. The arguments here are meant to be defensible at postgraduate level, but they are expressed without unnecessary jargon. 2. Literature Review Scholarship on the alliance falls into roughly five streams. Each answers a different question, and each leaves something out. 2.1 Diplomatic history The first stream is #diplomatic_history. Recent work by Rynning (2024) provides the fullest single volume account of the alliance from its founding to the war in Ukraine, and argues that NATO has endured because it is more than a mutual security guarantee. It is, in his reading, a political community with a habit of consultation, which is why it has survived quarrels that would have dissolved an ordinary pact. Sayle (2019) makes a related but sharper point. He treats the alliance as an instrument by which the United States managed its allies as much as it deterred its enemies. In this view, the treaty was never only about the #Soviet_Union. It was also about keeping Germany contained, keeping Britain and France from acting alone, and keeping American influence inside European decision making. Hanhimaki (2021) extends the historical argument into the post Cold War decades and asks why the transatlantic relationship did not collapse once its original enemy disappeared. Sarotte (2021) approaches the same period from the enlargement angle and reconstructs, from archival material, the disputed conversations of 1990 about whether the alliance would expand eastwards. Her book matters for students because it shows how contested the historical record itself has become, and how that contest feeds directly into present day political argument. What the historical literature does well is establish the sequence of events. What it does less often is read the treaty text closely as a legal document. 2.2 Alliance theory The second stream is #alliance_theory. Olson and Zeckhauser (1966) supplied the founding statement of the #burden_sharing problem: when security is a collective good, small members have an incentive to let large members pay. Snyder (1984) described the twin fears that haunt every alliance member, #abandonment and #entrapment. States worry both that their partners will not come when called, and that their partners will drag them into wars they did not choose. Walt (1987) argued that states align against threats rather than simply against power, which helps explain why so many European states chose to join the American camp rather than balance against it. Lanoszka (2022) offers the most useful recent synthesis. He surveys formation, entrapment, abandonment, burden sharing, wartime behaviour and termination, and warns against tidy assumptions. Alliances, he argues, behave in ways that are frequently counterintuitive. This literature gives us the vocabulary of the present article, but it works at a high level of abstraction and rarely descends to the specific wording of any particular treaty. 2.3 Institutionalist accounts The third stream, #institutionalism, explains persistence. Wallander (2000) provided the classic answer: NATO survived 1991 because it had accumulated what she called institutional assets, meaning procedures, standards, command structures and habits of working together, which could be repurposed for new tasks. An alliance that is only a promise dies when the threat dies. An alliance that has built machinery can redirect the machinery. Ikenberry (2020) situates the alliance within the broader project of liberal international order, arguing that it functioned as a security foundation for an open economic and political system. Critics reply that this reading is too flattering, and that it downplays the raw balance of power calculations at work. The institutionalist stream explains durability well. It is weaker at explaining strain, because it tends to assume that machinery, once built, keeps running. 2.4 The enlargement debate The fourth stream is the most politically charged. Menon and Ruger (2020) offer a net assessment of enlargement and conclude that its costs to American grand strategy have been underweighted. Goldgeier and Shifrinson (2023) assemble contributors on both sides of the argument. Mearsheimer (2022) makes the strongest realist case that expansion provoked Russia and helped cause the war. Against this, a large body of writing insists that #enlargement was demanded by the new members themselves, that sovereign states are entitled to choose their alignments, and that Russian behaviour in Georgia in 2008 and Ukraine in 2014 and 2022 reflects domestic and imperial drivers rather than a defensive reaction. Students should be careful here. This is a live political dispute, not a settled scholarly finding, and the evidence is used selectively on all sides. The present article does not attempt to resolve it. It does, however, note that the 1949 treaty contains an open door clause, and that whatever one thinks of the policy, expansion was legally provided for from the beginning. 2.5 Capability and burden sharing in the present The fifth stream is contemporary and policy facing. Meijer and Brooks (2021) argue, on the basis of a detailed capability audit, that Europe cannot provide for its own security in any near term if the United States withdraws, because the gaps in enabling capabilities such as strategic lift, intelligence, refuelling and suppression of enemy air defences are too wide to close quickly. Posen (2020) argues the opposite, that Europe has the wealth and the population to defend itself if it chooses. Ringsmose and Webber (2020) propose a middle path built around a European pillar inside the alliance rather than outside it. Deni (2021) is more pessimistic about European political will. Sloan (2020) traces the long arc of American ambivalence. 2.6 The gap Each stream is valuable. What is missing is a study that connects the legal text of 1949 to the political practice of 2026 in a way that is accessible to students. Historians read the archives. Theorists read the abstractions. Lawyers read the clauses. Policy analysts read the budgets. The present article tries to hold the four together, and argues that the strains now visible in the alliance were written into the founding document itself. 3. Research Design and Method This is a qualitative, interpretive study using #document_analysis. It does not test a hypothesis statistically. It builds an argument from documents and from the secondary literature, and it is transparent about how it does so. 3.1 Sources Three categories of material are used. The first is the primary legal text, meaning the treaty of 1949 itself, together with the related instruments of the period: the Brussels Treaty of 1948, the Vandenberg Resolution of the United States Senate, and the relevant provisions of the #United_Nations_Charter, in particular #Article_51, which preserves the right of individual and collective self defence. The second is the public record of alliance decisions, including #summit_declarations and #Strategic_Concept documents. These are official political documents and must be read as such. They are statements of agreed position, not neutral descriptions of reality, and they conceal disagreement as often as they reveal it. The third is the peer reviewed and scholarly secondary literature described in Section 2. 3.2 Method of interpretation The treaty is read using an approach familiar from treaty interpretation in international law: the ordinary meaning of the words, in their context, in light of the object and purpose of the agreement, with reference to the circumstances of its conclusion where the meaning is otherwise unclear. This is not a novel method. It is the standard approach reflected in the Vienna Convention on the Law of Treaties, and although that convention came later than 1949, it codified rules that were already accepted. The article then applies what might be called a #practice_tracing approach. Where the text is ambiguous, the analysis asks what alliance members actually did, since subsequent practice is itself evidence of how the parties understand their obligations. 3.3 Limits of the design Interpretive work of this kind cannot deliver certainty. Different readers, reading the same clause, reach different conclusions, and the archival record is incomplete for the most sensitive discussions. The article therefore aims for a defensible reading rather than a proven one, and it flags disagreement where disagreement exists. 4. Origins: Europe Between 1945 and 1949 4.1 The situation after the war The Second World War left Western Europe physically wrecked and politically frightened. Cities were rubble. Currencies were unstable. Coal was short. In the winter of 1946 to 1947, an exceptionally severe cold season across northern Europe pushed already fragile economies close to collapse. #Communist_parties, which had earned prestige in wartime resistance movements, were strong in France and Italy. Meanwhile, the Soviet Union kept large armies in the centre of the continent and installed friendly governments in the states it occupied. Two fears combined. The first was fear of invasion. The second, and in the short term the more pressing, was fear of collapse from within, that hunger and disorder would deliver governments to Moscow without a shot being fired. Understanding this double fear is essential, because the treaty that emerged was designed to address both. It was a military pact, and it was also a confidence building measure aimed at European populations. 4.2 The American turn American policy moved fast between 1946 and 1949, faster than its own political tradition would have predicted. In 1947 the United States announced support for Greece and Turkey, framed as a general commitment to assist free peoples resisting subjugation. In the same year it launched the European Recovery Program, usually remembered as the #Marshall_Plan, which channelled large scale economic aid into Western Europe. The aid was not charity. It was an investment in political stability, and it was conditioned on European cooperation in administering the funds, which had the useful side effect of teaching European governments to work together. Economic aid, however, did not answer the military question. European governments wanted something harder. They wanted a guarantee. 4.3 The Brussels Treaty In March 1948, Belgium, France, Luxembourg, the Netherlands and the United Kingdom signed the #Brussels_Treaty. Its Article IV contained an unambiguous #mutual_assistance clause. If one party was attacked in Europe, the others would give all the military and other aid in their power. That is an automatic and unconditional obligation, and the contrast with the language later chosen in Washington is one of the most instructive comparisons available to a student of the subject. The Brussels signatories knew perfectly well that five European states could not stop the Red Army. The treaty was, in part, a demonstration. It was meant to show Washington that Europeans would organise themselves, and thereby to earn an American commitment. 4.4 The Berlin blockade In June 1948 the Soviet Union cut surface access to the western sectors of Berlin. The Western powers responded with an airlift that sustained the city for close to a year. The #Berlin_Blockade did more than any argument to convince American opinion that the Soviet Union would use pressure up to the edge of war, and that Western Europe could not be left to face that pressure alone. 4.5 The constitutional obstacle There remained a serious problem in Washington, and it was not a problem of will but of law and politics. The Constitution of the United States gives Congress the power to declare war. A treaty that automatically committed American forces to combat would, in the eyes of many senators, hand that power to foreign governments. It would also collide with an old and deep aversion to entangling alliances in peacetime. The solution came through Senate Resolution 239 of June 1948, generally called the #Vandenberg_Resolution after the senator who sponsored it. It advised that the United States could associate itself with regional arrangements based on continuous and effective self help and mutual aid, provided such arrangements were consistent with the United Nations Charter. Two conditions are visible in that formula, and both shaped the treaty. Members would have to help themselves, not merely receive help. And the arrangement would have to sit within the framework of the Charter rather than outside it. Note what follows from this. The most quoted line in the treaty, the commitment in Article 5, was drafted in a form that would satisfy the Senate. That constraint, not any abstract theory of deterrence, explains the wording. 4.6 Negotiation and signature Talks in Washington ran through the winter of 1948 and into 1949. The negotiating parties argued about scope, about which countries to include, and above all about the strength of the guarantee. The Europeans wanted the Brussels formula. The Americans could not give it. The compromise text was signed on 4 April 1949 by twelve founding states: Belgium, Canada, Denmark, France, Iceland, Italy, Luxembourg, the Netherlands, Norway, Portugal, the United Kingdom and the United States. Iceland had no armed forces at all. Portugal was an authoritarian state, which sat awkwardly with the treaty preamble's language about democracy, and this tension between declared values and strategic convenience has recurred throughout the alliance's history. The remark most often attached to the alliance's purpose, that it existed to keep the Russians out, the Americans in and the Germans down, is usually attributed to Lord Ismay, the first Secretary General. Its provenance is uncertain and students should treat it as folklore rather than evidence. It is repeated here only because it captures, crudely but usefully, the three way character of the bargain. 5. Reading the Treaty The treaty is short. A careful reading of its main clauses is the fastest route to understanding the alliance. 5.1 The preamble The preamble affirms the parties' faith in the purposes and principles of the United Nations Charter and their desire to live in peace with all peoples and all governments. It states their determination to safeguard the freedom, common heritage and civilisation of their peoples, founded on the principles of democracy, individual liberty and the rule of law. Two points deserve attention. First, the framing is defensive and legalistic. The drafters were careful to present the alliance as an instrument permitted by the Charter, not as a rival to it. Second, the reference to democracy and the rule of law introduced a normative element that has repeatedly caused difficulty, because the alliance has always included members whose domestic practice fell short of that standard. The preamble is not legally binding in the same way as the operative articles, but it is not decorative either. It informs interpretation, and it has been invoked in political argument ever since. 5.2 Article 1 Article 1 commits the parties to settle international disputes by peaceful means and to refrain from the threat or use of force in a manner inconsistent with the purposes of the United Nations. This restates existing Charter obligations. It is a signal to non members that the alliance is not a vehicle for aggression. 5.3 Article 2 Article 2 obliges the parties to strengthen their free institutions, to promote conditions of stability and wellbeing, and to encourage economic collaboration. It is often called the Canadian article, because Canadian negotiators pressed hardest for it. They wanted the alliance to be more than a military instrument, partly out of conviction and partly because a purely military pact would have been harder to sell at home. Article 2 has largely lain dormant in practice, but it is not without consequence. It is the textual hook on which later arguments about the alliance as a community of values have been hung. 5.4 Article 3 Article 3 is the article that the current arguments about #defence_spending run through, and students routinely overlook it. It provides that the parties, separately and jointly, by means of continuous and effective self help and mutual aid, will maintain and develop their individual and collective capacity to resist armed attack. Read that again. The phrase continuous and effective self help and mutual aid is lifted almost directly from the Vandenberg Resolution. It is an obligation of effort. Members must build and keep real military capacity. They do not discharge their duty simply by signing. When alliance leaders agreed at The Hague in 2025 to invest five per cent of gross domestic product annually by 2035, they expressly framed the commitment as being in accordance with Article 3. The spending debate is therefore not a modern invention grafted onto an old treaty. It is a dispute about compliance with an original clause. 5.5 Article 4 Article 4 provides that the parties will consult together whenever, in the opinion of any of them, the territorial integrity, political independence or security of any party is threatened. The threshold is deliberately low. Consultation can be triggered by an opinion, not by proof of attack. In practice #Article_4 has been invoked a number of times, including by Turkey in relation to instability on its southern border and by several eastern members in the days after Russia's full scale invasion of Ukraine in February 2022. It functions as an alarm bell. It creates a duty to talk, not a duty to fight, and this is precisely why it is used far more often than the famous clause that follows it. 5.6 Article 5 Article 5 is the heart of the treaty, and it repays slow reading. It provides that the parties agree that an armed attack against one or more of them in Europe or North America shall be considered an attack against them all. It then states that each of them, in the exercise of the right of individual or collective self defence recognised by Article 51 of the Charter of the United Nations, will assist the party or parties so attacked by taking forthwith, individually and in concert with the other parties, such action as it deems necessary, including the use of armed force, to restore and maintain the security of the North Atlantic area. Measures taken are to be reported to the Security Council and terminated when the Security Council has taken the measures necessary to restore international peace and security. Four features stand out. First, the trigger is an armed attack. Not a threat. Not subversion. Not economic coercion. Not, on the plain wording, a cyber operation, although alliance policy has since developed to treat certain cyber and hybrid actions as potentially rising to the level of an armed attack, a development which shows how much interpretation matters. Second, and most importantly, the response is qualified. Each member takes such action as it deems necessary. Those five words carry enormous weight. They mean that the decision about what to do belongs to each member state, individually. There is no automatic obligation to declare war. A member could, in theory, satisfy the letter of the clause with something less than combat. Third, the phrase including the use of armed force does real work in the other direction. It puts the use of force explicitly on the table. The drafters could have omitted it. Its presence signals seriousness even while the surrounding words preserve discretion. Fourth, the whole clause is nested inside the United Nations framework by the express reference to Article 51. This was not decoration. It was the legal argument that allowed the treaty to be presented as consistent with the Charter. So Article 5 is weaker on paper than the Brussels formula it replaced. Why has it nonetheless worked? The answer, which is central to this article, is that the credibility of the promise never rested on the text. It rested on what was built behind the text: an integrated command structure, permanently stationed American forces, joint plans, shared standards, common exercises and, from the 1950s, nuclear weapons deployed in Europe. American soldiers stationed in Germany were not there because the text required it. They were there because their presence made the promise physical. An adversary contemplating an attack would have had to kill Americans in the first hours, which made the American response near automatic in fact even though it remained discretionary in law. This is the deepest lesson of the treaty. Words create a framework. Deployments create #credibility. 5.7 Article 6 Article 6 defines the geographic scope of an armed attack for the purposes of Article 5. It covers the territory of the parties in Europe and North America, along with certain islands, and it extends to the forces, vessels and aircraft of the parties in specified areas. The definition matters because it limits the alliance. Colonial possessions outside the defined area were excluded, which is why the wars of decolonisation fought by France, Britain, Portugal and the Netherlands were not alliance wars. 5.8 Articles 7 to 9 Article 7 preserves the primacy of the Security Council. Article 8 requires that members not enter into commitments inconsistent with the treaty. Article 9 establishes a council, empowered to set up subsidiary bodies as necessary. Article 9 deserves more attention than it usually gets. It is the clause from which the entire organisation grew. The treaty of 1949 did not create #NATO as we know it. It created a council with the power to create things. Everything else, the Secretary General, the international staff, the integrated military command, the headquarters, the committees, followed from that single permissive sentence and from the political shock of the Korean War in 1950. 5.9 Article 10 Article 10 is the #open_door_policy. It allows the parties, by unanimous agreement, to invite any other European state in a position to further the principles of the treaty and to contribute to the security of the North Atlantic area to accede. Three conditions are embedded here. The candidate must be European. The invitation must be unanimous. And the candidate must be able to contribute, not merely to consume, security. The #unanimity requirement is the reason individual members have been able to delay accessions, as Turkey and Hungary did in the cases of Finland and Sweden. It also explains why membership questions are always as much about internal alliance politics as about the candidate's qualities. 5.10 Articles 11 to 14 Article 11 provides for ratification according to each party's constitutional processes, a clause inserted with the United States Senate very much in mind. Article 12 permits review after ten years. Article 13 allows a party to withdraw, on one year's notice, after the treaty has been in force for twenty years. Article 14 deals with deposit of the instrument. Article 13 is worth pausing on. Membership is not a life sentence. Any member may leave. The fact that no member has ever done so, although France withdrew from the integrated military command structure between 1966 and 2009 while remaining a party to the treaty, is itself evidence about the alliance's value to its members. 6. From Treaty to Organisation, 1950 to 1991 6.1 The Korean shock In June 1950 North Korean forces crossed the thirty eighth parallel. The war was in Asia, but its effect on European security was immediate. Western governments read the invasion as evidence that the Soviet Union was willing to sponsor armed aggression across a divided line, and they looked at the divided line running through Germany. Before Korea, the alliance was a treaty with a council. After Korea, it became an organisation with an army. An #integrated_military_command was established. Dwight Eisenhower was appointed the first Supreme Allied Commander Europe in 1951. American troop numbers in Europe rose sharply. National forces were assigned to allied command in peacetime, which was, and remains, an extraordinary surrender of sovereign control. Students should mark this point carefully. The organisation was not required by the 1949 text. It was a response to fear in 1950, constructed under the permissive authority of Article 9. 6.2 The German question The most difficult problem of the 1950s was the #German_question. The alliance needed German manpower and German territory, but France in particular had obvious reasons to resist German rearmament only five years after occupation. An attempted solution, the European Defence Community, would have created a supranational European army. It was signed in 1952 and then killed in 1954 by the French National Assembly, which refused to ratify it. The failure is instructive, because the arguments used against it, about sovereignty, about command, about whether Europe could act without Britain, are almost word for word the arguments used against European defence integration today. The fallback was to bring West Germany into the alliance directly, which happened in 1955, with the Federal Republic accepting limits on the weapons it would produce. Moscow responded by forming the #Warsaw_Pact in the same year. The two bloc structure of the Cold War was now complete. 6.3 Nuclear weapons and strategic doctrine Nuclear weapons transformed the alliance's internal politics. The doctrine of the 1950s, massive retaliation, promised to answer any significant Soviet attack with overwhelming nuclear force. It was cheap. It was also, once the Soviet Union developed the ability to strike the United States, deeply implausible. Would an American president risk New York to defend Hamburg? That question, sometimes called the credibility problem of extended #nuclear_deterrence, sits at the centre of alliance politics from the late 1950s to the present. France concluded that the answer was no, developed its own nuclear force and, under Charles de Gaulle, withdrew from the integrated command structure in 1966. The alliance headquarters moved from Paris to Brussels. France remained a treaty party throughout, which shows that the treaty and the organisation are separable. The doctrinal answer, adopted in 1967, was #flexible_response. It provided for a graduated set of options rather than an immediate leap to nuclear war. It was politically ambiguous by design, since Europeans wanted the nuclear threshold low enough to deter and Americans wanted it high enough to allow conventional defence. 6.4 The Harmel Report Also in 1967 came the #Harmel_Report, which stated that the alliance had two functions: to maintain adequate military strength and political solidarity, and to pursue a more stable relationship in which underlying political issues could be resolved. Defence and dialogue. Deterrence and detente. Harmel matters because it gave the alliance a second purpose that could survive changes in the threat. Rynning (2024) suggests that something like a new Harmel process may now be needed to define the alliance's relationship with Russia and with its own European pillar. Whether that is realistic is debatable. The point for students is that the alliance has renewed itself before by redefining its purpose in a public document, and that this is a recognisable pattern rather than an improvisation. 6.5 The euromissile crisis The late 1970s and early 1980s produced the last great Cold War crisis inside the alliance. The Soviet deployment of intermediate range missiles led to the dual track decision of 1979: deploy comparable American missiles in Europe while simultaneously negotiating for their removal. Mass protest movements followed in several member states. The deployments proceeded. The negotiation eventually succeeded, in the form of a treaty eliminating an entire class of weapons in 1987. The episode is a good case study of #alliance_cohesion under domestic political pressure, and it demonstrates that unpopular collective decisions can be sustained if members believe the alternative is worse. 7. Reinvention, 1991 to 2014 7.1 Out of a job In 1991 the Warsaw Pact dissolved and the Soviet Union ceased to exist. On any straightforward reading of alliance theory, the alliance should have wound down. Alliances are formed against threats. The threat was gone. It did not wind down. Wallander (2000) provides the standard explanation. The alliance possessed institutional assets that could be redeployed. It had a command structure, interoperability standards, a habit of consultation and a functioning bureaucracy. Those things were expensive to build and cheap to maintain, and they were useful for tasks other than fighting the Red Army. 7.2 New tasks Three new tasks emerged. The first was #crisis_management outside member territory. The wars accompanying the breakup of Yugoslavia drew the alliance into its first sustained combat operations, in Bosnia in 1995 and over #Kosovo in 1999. The Kosovo campaign was conducted without explicit Security Council authorisation, and it opened a legal argument about the use of force that has never been settled. Students should note the irony. An alliance whose founding text bent over backwards to demonstrate consistency with the United Nations Charter conducted a war that many international lawyers considered inconsistent with it. The second was #partnership. Programmes were created to link the alliance with former adversaries, including Russia, and to prepare candidate states for membership. The third, after the attacks of 11 September 2001, was counter terrorism and expeditionary operations. #Article_5 was invoked for the first and only time in its history, not to defend Europe from the east, but to defend the United States after an attack by a non state group. The invocation is legally interesting because it stretched the concept of armed attack beyond an attack by a state, and because the assistance actually rendered was modest relative to the symbolism. The subsequent alliance mission in #Afghanistan lasted two decades and ended in 2021 in a withdrawal that damaged the alliance's confidence in its own judgement. 7.3 Enlargement Between 1999 and 2009 the alliance admitted twelve new members from central and eastern Europe and the Balkans. The formal basis was Article 10, and the political driver was demand from the candidates themselves, who had lived under Soviet domination and wanted an insurance policy. The consequences remain contested, as Section 2 noted. What is not contested is that enlargement changed the alliance's internal balance. It brought in members with intense threat perceptions focused on Russia and comparatively small militaries, and it extended the Article 5 guarantee to territory that would be extremely difficult to defend, most obviously the Baltic states. 7.4 Complacency Between 1991 and 2014, European defence budgets fell steadily. Conscription was abandoned in many states. Tank fleets shrank. Ammunition stocks were run down. Militaries were reorganised around small deployable formations suited to expeditionary operations rather than around territorial defence against a peer adversary. This was not irrational at the time. It was a bet that large scale war in Europe was over. The bet lost. 8. The Return of Collective Defence, 2014 to 2022 8.1 Crimea In 2014 Russia annexed Crimea and supported armed separatism in eastern Ukraine. The response inside the alliance was rapid in rhetoric and slow in substance. At the Wales summit later that year, members agreed the Defence Investment Pledge, committing to move towards spending two per cent of gross domestic product on defence within a decade, with at least twenty per cent of that going to major equipment. The two per cent figure has since become the most quoted number in transatlantic politics, and it deserves scepticism. It measures input, not output. A state can meet it by paying pensions and salaries. A state can miss it and still field capable forces. Spending as a share of national income says nothing on its own about whether the money buys anything useful. 8.2 Forward presence From 2016 the alliance deployed multinational battlegroups to the #eastern_flank, in Estonia, Latvia, Lithuania and Poland. These were small, roughly battalion sized, and they were not intended to defeat an invasion by themselves. They were a #tripwire. Their function was to guarantee that an attacker would immediately be in combat with soldiers from many member states, thereby engaging the whole alliance and removing any hope of a quick, contained land grab. The tripwire logic is a direct descendant of the reasoning behind the American garrison in Germany in the 1950s. It is a mechanism for turning a discretionary legal promise into a practical certainty. This is worth emphasising because it shows continuity across seventy years of practice. 8.3 Political strain The same period saw serious political strain. Successive American administrations complained that European members were not carrying their share. The complaint was not new. Every American president since Eisenhower has made some version of it. What changed was the tone, and the open questioning of whether the United States would honour the guarantee at all in the case of members it judged delinquent. Sperling and Webber (2019) analysed this as a problem of exit and voice, arguing that loud American dissatisfaction can be a bargaining tactic rather than a genuine step towards departure. Whether that reading remains adequate is one of the open questions of the present decade. 9. The Alliance since February 2022 9.1 Invasion and response Russia's full scale invasion of #Ukraine in February 2022 was the largest conventional attack in Europe since 1945. Several eastern members immediately requested consultations under #Article_4. The alliance did not enter the war. This is a point students frequently misunderstand. Ukraine is not a member, so Article 5 did not apply, and members were careful to avoid direct combat with Russian forces. What the alliance did do was reinforce its own eastern territory, doubling the number of multinational battlegroups to eight, while members supported Ukraine bilaterally and through other frameworks with weapons, money, intelligence and training. The distinction between what the alliance does as an alliance and what its members do individually is legally and politically important, and it is one of the reasons the war has not escalated into a direct confrontation. 9.2 Finland and Sweden Two states with long traditions of military non alignment applied to join. #Finland acceded in 2023 and #Sweden in 2024, bringing the membership to thirty two. Finland's accession roughly doubled the length of the alliance's land border with Russia. Sweden's accession made the Baltic Sea, in strategic terms, an area largely surrounded by allied territory. Both accessions were delayed by objections from existing members, illustrating the practical effect of the unanimity requirement in Article 10. Both were also, in a sense, the clearest possible verdict on the invasion. Two countries that had stayed out of alliances for generations concluded that non alignment was no longer safe. 9.3 Strategy and plans The 2022 Strategic Concept identified Russia as the most significant and direct threat to allied security, and for the first time addressed China, describing its stated ambitions and policies as a challenge to allied interests, security and values. In 2023 the alliance adopted new regional defence plans, described as the most comprehensive since the Cold War, which specify what each member is expected to provide. These plans moved the posture away from a tripwire and towards #forward_defence, meaning an intention to hold and defend territory from the first day rather than to liberate it later. 9.4 The Hague and the five per cent pledge At The Hague in June 2025, leaders agreed to invest five per cent of gross domestic product annually by 2035, split into at least three and a half per cent for core defence requirements and up to one and a half per cent for defence and security related spending such as critical infrastructure, cyber defence, civil preparedness, innovation and the defence industrial base. Progress is to be reviewed in 2029. Spain secured an effective exemption, arguing that it could meet its assigned capability targets with a much lower figure. The pledge is remarkable in scale and doubtful in execution. Analysts have raised at least three objections that students should weigh. Defence cost inflation regularly outpaces general inflation, so nominal increases translate into smaller real capability gains than the headline suggests. Many member states carry heavy public debt and face hard choices between defence and welfare. And a spending target remains an input measure, which tells you nothing about whether the money is spent on the right things or wasted on fragmented national programmes. 9.5 American retrenchment and the Ankara summit Through 2025 and into 2026, the United States signalled a reduction in what it will make available to the alliance. Reported measures included the withdrawal of several thousand troops from Germany, the cancellation of a planned armoured brigade rotation to Poland, and cuts to the aircraft and naval assets that Washington would commit to allied plans in a crisis. American officials framed this as a rebalancing consistent with a national defence strategy under which European allies take primary responsibility for conventional defence in Europe, with more limited American support. Senior members of the United States Congress from the president's own party publicly objected that reductions were arriving before European capabilities were ready. At the Ankara summit on 7 and 8 July 2026, leaders again described their commitment to Article 5 as ironclad. They reported that European members and Canada had increased core defence investment by more than one hundred and thirty nine billion dollars in the year since The Hague, announced more than fifty billion dollars in new procurement, launched a major initiative on uncrewed systems, agreed a large investment in fuel storage and distribution infrastructure extending towards the eastern part of the alliance, and pledged seventy billion euros in military equipment, assistance and training for Ukraine in 2026 with a commitment to sustain a comparable level in 2027. The Secretary General described the process as rebalancing and used the phrase a stronger Europe in a stronger NATO. Read carefully, the Ankara declaration is a document about transition. It reaffirms the guarantee in the strongest available language precisely because the material basis of the guarantee is being altered. That combination, stronger words alongside a thinner American presence, is the central tension of the present moment, and it takes us straight back to the drafting problem of 1949. The promise is discretionary, and its force depends on what stands behind it. 10. Four Analytical Themes The narrative sections above supply the facts. This section turns them into analysis by drawing out four themes that connect 1949 to 2026. 10.1 Burden sharing The oldest quarrel in the alliance, #burden_sharing, is about who pays. Olson and Zeckhauser (1966) explained the logic. Security produced by an alliance has the character of a #public_good. If deterrence works, it protects every member, including the member that contributed least. A small state therefore has a rational incentive to under invest and rely on the large state, which will provide deterrence anyway because it wants it for itself. The predicted result is that the largest member carries a disproportionate share, and that is what has been observed for seventy years. Three qualifications should be attached to this classic account, because textbooks often repeat it too flatly. First, the alliance does not only produce a pure public good. Much of what members buy is excludable and specific to national territory, which weakens the free riding prediction. Second, #free_riding is not the only explanation of low European spending. Different threat perceptions matter. A country on the Baltic coast and a country on the Atlantic coast do not face the same risk, and it is not obviously irrational for them to spend differently. Domestic fiscal politics matter. Public opinion matters. Third, and most importantly, the American complaint has never been purely about money. It has been about capability and risk. Spain's position in 2025, that it could reach its assigned capability targets without reaching the headline percentage, is an argument about outputs against a target defined in inputs, and it exposes the weakness of using a spending ratio as the measure of a member's contribution. The five per cent pledge of 2025 restates an argument that Article 3 started in 1949. It does not resolve it. Whether members can convert money into deployable capability, and whether they will keep doing so once the immediate fear recedes, is the open question of the next decade. 10.2 Credibility The second theme is the one this article treats as central. Article 5 does not compel any member to fight. It compels each member to decide, individually, what it considers necessary. So where does #deterrence come from? The answer is that credibility is manufactured, not declared. Over seven decades the alliance has manufactured it through several devices. Forward stationed forces, especially forces from the guarantor power, mean that an attacker cannot avoid killing the guarantor's citizens. Integrated command means that national forces are already plugged into a shared system, which removes the delay in which doubt could grow. Common plans and exercises mean that the response has been rehearsed. #Nuclear_sharing arrangements couple European territory to American nuclear forces. Repetition of the ironclad language at every summit builds a reputational cost for any leader who might later hesitate. Each of these devices is a substitute for the automatic legal obligation that the Senate would not accept in 1949. Which leads to a hard conclusion. If the devices are weakened, for example by a substantial reduction in forward stationed American forces and in the assets committed to allied plans, then the guarantee is weakened, no matter how strong the language of the communique. Words cannot compensate for withdrawn capability, because the words were always the weakest part of the arrangement. This is not a prediction that the alliance will fail. It is an argument about where students should look if they want to measure its health. Look at deployments, plans and stockpiles, not at declarations. 10.3 Consensus The alliance decides by #consensus_rule, and this shapes everything it does. There is no vote and no veto in the formal sense, because there is nothing to veto. A decision exists only when no member objects. This has two consequences that pull in opposite directions. Consensus makes decisions slow, and it hands leverage to any member willing to withhold agreement. The delays to Finnish and Swedish accession are the clearest recent illustration. Members with domestic agendas can extract concessions by holding up business that others want. But consensus also makes decisions strong. Because nobody was outvoted, nobody has an excuse for non compliance later. In an organisation with no enforcement power, that matters more than speed. Critics who propose majority voting rarely explain how a state that was outvoted could be made to send its soldiers to die. The consensus rule is not stated in the treaty. It grew out of practice under Article 9, and it is a good example of how the organisation's most important rules are customary rather than textual. 10.4 Enlargement and the open door Article 10 created a door and left it open. Twenty additional states have walked through it since 1949. The scholarly argument about whether enlargement was wise is unresolved and probably unresolvable, because it turns on a counterfactual that cannot be observed. We cannot know what Russia would have done had the alliance stopped at Germany. Both sides of the debate rely on inference from character and intention, which is weak evidence. What can be said with more confidence is structural. Every accession increases the number of territories the alliance has promised to defend and the number of vetoes on any decision. It thus raises both the cost of the guarantee and the difficulty of agreeing anything. The alliance has grown its liabilities faster than it has grown its capabilities, and the current spending debate is an attempt to close that gap. The case of #Ukraine_membership crystallises the dilemma. Membership would extend the guarantee to a state at war. Refusing membership leaves that state outside the only security arrangement that has reliably worked in Europe. Alliance leaders have repeatedly avoided resolving the question, which is itself a decision. 11. Discussion 11.1 What the treaty explains Reading the 1949 text closely explains several things that would otherwise seem puzzling. It explains why Article 4 is used often and Article 5 almost never. Consultation is cheap, and it is the tool the treaty gives members for expressing alarm without committing to force. It explains why arguments about spending recur endlessly and are never settled. Article 3 imposes an obligation of effort with no metric attached, so the metric has to be renegotiated politically every few years, from the two per cent of 2014 to the five per cent of 2025. It explains why the alliance was able to survive the end of the #Cold_War. The treaty does not name an enemy. Read the text again and notice the absence. The Soviet Union appears nowhere. An alliance defined against a named adversary would have expired with that adversary. This one did not, because it was drafted as an open ended arrangement for the defence of a region. And it explains why the alliance is so sensitive to the credibility of a single member. The discretion embedded in Article 5 means the guarantee has always depended on belief about what the United States would do, and belief is fragile in a way that legal obligation is not. 11.2 What the treaty does not explain The text does not explain the organisation. Nothing in 1949 required an integrated command, a Secretary General, a permanent headquarters or a nuclear planning group. Those were built later, under the permissive language of Article 9, in response to events. If a student wants to understand NATO the organisation, the treaty is only the beginning of the story. The text also does not explain the alliance's out of area operations. Article 6 defines a geographic area, and the campaigns in Kosovo, Afghanistan and Libya sat uneasily with that definition. They were justified politically and by council decision rather than by the treaty's own logic, and they cost the alliance something in coherence. Nor does the text explain the values dimension. The preamble speaks of democracy and the rule of law. The alliance has admitted and retained members whose governments departed from those principles, and it has no mechanism for enforcing them internally. There is no suspension clause, and no mechanism of #democratic_conditionality. This gap is not accidental. Strategic value has consistently outweighed normative purity, from Portugal in 1949 onwards, and honesty about that record is a necessary part of studying the alliance seriously. 11.3 A synthesis Bringing the strands together, the argument of this article is that the alliance is best understood as a legal shell filled with political and military content, and that the content, not the shell, is what deters. The shell has proved extraordinarily durable, because it is vague, because it names no enemy, because it leaves discretion to each member and because it permits the council to invent whatever machinery the moment requires. The content has been rebuilt several times: mass conventional and nuclear confrontation in the 1950s and 1960s, detente and dialogue after 1967, crisis management and enlargement after 1991, expeditionary war after 2001, tripwire deterrence after 2014, forward defence after 2022. The present moment is another rebuild. #Deterrence in Europe is being reconstructed with a larger European share and a smaller American one, at speed, while a war is running on the alliance's border. Whether that rebuild succeeds is not determined by the treaty. The treaty permits it. It also permits failure. 12. Limitations of the Study This study has clear limits and it is better to state them than to hide them. It is document based, and documents produced by governments are self serving. Summit declarations describe unity because they are written to project unity. They are evidence of what members agreed to say, which is not always evidence of what they agreed to do. It draws on events up to mid 2026, and several of those events, in particular the American force posture decisions, were still unfolding at the time of writing. Any assessment of them is provisional. It does not include original interviews or archival research, and it therefore depends on the accuracy of the historians and analysts whose work is cited. Finally, it is written from within a broadly Euro Atlantic frame of reference. A Russian, Chinese or non aligned scholar would ask different questions and would weigh the enlargement debate differently. Readers should seek out those perspectives as well, and should be alert to the fact that this topic attracts strongly committed advocacy on all sides. 13. Conclusion The North Atlantic Treaty is a short document that has carried an enormous weight. It was drafted under a specific #constitutional constraint, the refusal of the United States Senate to accept an automatic commitment to war, and that constraint produced a promise that is conditional in its wording. Each member decides for itself what action it deems necessary. Everything that has made the alliance formidable since 1949 has been an attempt to compensate for that conditionality. Stationed forces, integrated command, common plans, nuclear sharing, repeated public reaffirmation. These are the machinery of credibility, and they are the reason a discretionary promise came to function like an automatic one. That is why the current period matters so much. The language of the guarantee is as strong in 2026 as it has ever been. The machinery behind it is being altered. If the alliance can rebuild that machinery on a more European base, the promise will hold. If it cannot, the words of Article 5 will not save it, because the words were never the source of its strength. For students, the practical lessons are these. Read the primary text before the commentary, because the text is short and most commentary misdescribes it. Distinguish the treaty from the organisation. Distinguish what the alliance does collectively from what its members do individually. Treat spending percentages as political symbols rather than measures of capability. And treat the #enlargement_debate as an open argument in which the evidence is contested, not as a question with a settled answer. Suggestions for further research Four questions seem to this author both important and answerable. First, a comparative study of alliance guarantee wording, setting the 1949 formula against the automatic clause of the Brussels Treaty and against the mutual assistance clause in European Union law, asking whether stronger words produce stronger behaviour. Second, an empirical study of whether spending targets change capability outputs, using the period from 2014 to 2026 as a natural experiment. Third, a study of consultation under Article 4: who invokes it, when, and what actually follows. Fourth, a study of how the alliance's decision making performs under conditions of reduced American participation, using the ministerial and summit record from 2025 onwards. Hashtags #North_Atlantic_Treaty #NATO_1949 #Washington_Treaty #Collective_Defence #Article_5_NATO #Transatlantic_Security #Alliance_Politics #Cold_War_History #European_Security #Deterrence_Theory #Burden_Sharing #NATO_Enlargement #International_Law #Treaty_Interpretation #Security_Studies References Deni, J. R. (2021). Coalition of the unWilling and unAble: European Realignment and the Future of American Geopolitics. Ann Arbor: University of Michigan Press. Goldgeier, J., and Shifrinson, J. R. I. (eds.) (2023). Evaluating NATO Enlargement: From Cold War Victory to the Russia-Ukraine Crisis. Cham: Palgrave Macmillan. Hanhimaki, J. M. (2021). Pax Transatlantica: America and Europe in the Post-Cold War Era. New York: Oxford University Press. Henke, M. E. (2019). Constructing Allied Cooperation: Diplomacy, Payments, and Power in Multilateral Military Coalitions. Ithaca: Cornell University Press. Ikenberry, G. J. (2020). A World Safe for Democracy: Liberal Internationalism and the Crises of Global Order. New Haven: Yale University Press. Kunertova, D. (2023). Drones have boots: Learning from Russia's war in Ukraine. Contemporary Security Policy, 44(4), 576-591. Lanoszka, A. (2022). Military Alliances in the Twenty-First Century. Cambridge: Polity Press. Mearsheimer, J. J. (2022). The causes and consequences of the Ukraine war. Horizons: Journal of International Relations and Sustainable Development, 21, 12-27. Meijer, H., and Brooks, S. G. (2021). Illusions of autonomy: Why Europe cannot provide for its security if the United States pulls back. International Security, 45(4), 7-43. Menon, R., and Ruger, W. (2020). NATO enlargement and US grand strategy: A net assessment. International Politics, 57(3), 371-400. Olson, M., and Zeckhauser, R. (1966). An economic theory of alliances. The Review of Economics and Statistics, 48(3), 266-279. Plokhy, S. (2023). The Russo-Ukrainian War: The Return of History. New York: W. W. Norton. Posen, B. R. (2020). Europe can defend itself. Survival, 62(6), 7-34. Ringsmose, J., and Webber, M. (2020). Hedging their bets? The case for a European pillar in NATO. Defence Studies, 20(4), 295-317. Rynning, S. (2024). NATO: From Cold War to Ukraine, a History of the World's Most Powerful Alliance. New Haven: Yale University Press. Sarotte, M. E. (2021). Not One Inch: America, Russia, and the Making of Post-Cold War Stalemate. New Haven: Yale University Press. Sayle, T. A. (2019). Enduring Alliance: A History of NATO and the Postwar Global Order. Ithaca: Cornell University Press. Schelling, T. C. (1966). Arms and Influence. New Haven: Yale University Press. Sloan, S. R. (2020). Defense of the West: Transatlantic Security from Truman to Trump, 2nd edn. Manchester: Manchester University Press. Snyder, G. H. (1984). The security dilemma in alliance politics. World Politics, 36(4), 461-495. Sperling, J., and Webber, M. (2019). Trump's foreign policy and NATO: Exit and voice. Review of International Studies, 45(3), 511-526. Walt, S. M. (1987). The Origins of Alliances. Ithaca: Cornell University Press. Wallander, C. A. (2000). Institutional assets and adaptability: NATO after the Cold War. International Organization, 54(4), 705-735. Webber, M., Sperling, J., and Smith, M. A. (2021). What's Wrong with NATO and How to Fix It. Cambridge: Polity Press. Zubok, V. M. (2021). Collapse: The Fall of the Soviet Union. New Haven: Yale University Press. Primary documents Charter of the United Nations (1945), in particular Article 51. North Atlantic Treaty (1949), Washington DC, 4 April 1949. Treaty of Economic, Social and Cultural Collaboration and Collective Self-Defence (Brussels Treaty) (1948), Article IV. United States Senate Resolution 239, 80th Congress (the Vandenberg Resolution) (1948). North Atlantic Treaty Organization, Strategic Concept (2022). North Atlantic Treaty Organization, The Hague Summit Declaration (2025). North Atlantic Treaty Organization, The Ankara Summit Declaration (2026).

The Treaty of Versailles, signed on 28 June 1919, formally ended the state of war between Germany and the Allied and Associated Powers and created the League of Nations, the first permanent general international organisation in modern history. This article offers a structured academic review of the treaty for students of history, international relations, law, and political economy. It examines the origins of the settlement in the armistice of November 1918, the diplomatic process of the Paris Peace Conference, the substantive terms of the treaty, and the institutional design of the League of Nations. It then analyses the treaty's economic provisions, its reception inside Germany, its reception across Asia, Africa, and the Middle East, and the record of the League in practice between 1920 and 1946. The article argues three main points. First, the treaty was neither uniquely cruel nor unusually generous by the standards of its own time, but it was unusually difficult to enforce, because the coalition that wrote it began to break apart almost as soon as it was signed. Second, the League of Nations was not simply a failed peace project; it was also a machine for organising global power, and its mandate system extended imperial control under a new legal vocabulary. Third, the settlement mattered less because of any single clause than because of the political meanings that different societies attached to it. The article closes with implications for classroom use, an assessment of the treaty's long shadow over the United Nations system, and suggestions for further student research. Keywords: Treaty of Versailles; Paris Peace Conference; League of Nations; reparations; war guilt clause; mandate system; Weimar Republic; collective security; interwar diplomacy; international order 1. Introduction Few documents in modern history have been blamed for as much as the #Treaty_of_Versailles. It has been described as the cause of the Second World War, the trigger of German hyperinflation, the death certificate of European empire, and the birth certificate of modern international organisation. All four claims are partly true and all four are partly misleading. The purpose of this article is to give students a careful, evidence based account of what the treaty actually said, what it actually did, and why historians still argue about it more than a century later. The war that the treaty ended had killed somewhere between nine and eleven million soldiers and several million civilians. It had destroyed four empires: the German, the Austro-Hungarian, the Russian, and the Ottoman. It had bankrupted the victors almost as thoroughly as the defeated. When the delegates gathered in #Paris in January 1919, they were not simply drawing borders. They were trying to answer a question that no previous peace conference had faced at this scale: how do you rebuild an international system after industrial total war? Their answer had two parts. The first part was punishment and containment. #Germany would lose territory, lose its colonies, lose most of its army and navy, and pay for the damage it had caused. The second part was institutional. A permanent organisation, the #League_of_Nations, would manage disputes, supervise disarmament, administer former colonies, and provide what the drafters called #collective_security. The two parts were bound together in a single document. The Covenant of the League forms Part I of the Treaty of Versailles. This was a deliberate choice by Woodrow Wilson, and it had consequences that the drafters did not foresee: it meant that anyone who rejected the punishment of Germany was also, in a legal sense, rejecting the peace organisation, and anyone who rejected the peace organisation was rejecting the settlement. This article proceeds in the manner of a review study. It is organised around four research questions: What were the aims, procedures, and constraints of the peacemakers at Paris, and how did these shape the terms of the treaty? How harsh were the treaty's economic and territorial provisions in comparative and historical perspective? What kind of institution was the League of Nations in design and in practice? How should the treaty's long term legacy be assessed in light of recent scholarship? The article does not claim to present new archival evidence. It presents a synthesis of the published record and of scholarship from the last five years, much of which has quietly revised the older textbook story. Students should note this at the outset. The "harsh treaty caused Hitler" account that still appears in many school syllabuses is not what most working historians now argue, and the reasons for that shift are themselves worth studying. 2. Literature Review Scholarship on the treaty falls into four broad and overlapping traditions. 2.1 The economic critique and its revision The oldest and most influential critique came from within the conference itself. #John_Maynard_Keynes, a young British Treasury official, resigned from the delegation and published a polemic in late 1919 arguing that the #reparations demanded of Germany were beyond its capacity to pay, that the settlement had ignored the economic interdependence of Europe, and that the leaders at Paris had produced a peace of vengeance rather than reconstruction. His book sold enormously, shaped Anglo-American opinion for a generation, and gave German nationalists a respectable foreign authority to quote. Recent work has complicated this picture without dismissing it. The centenary volume edited by Clavin, Corsetti, Obstfeld, and Tooze (2024) treats Keynes's book as a work of political argument as much as economic analysis, and shows that its central theme was the conflict between narrow nationalism and economic integration rather than a simple technical claim about German capacity to pay. Economic historians have argued for some time that the sums finally demanded were within Germany's ability to pay if the political will had existed, and that the real obstacle was not arithmetic but the refusal of successive German governments to be seen paying. Ho (2025), reassessing the Dawes Plan on its centenary, concludes that the fundamental defect of interwar reparations was not the headline figure but the absence of a workable enforcement mechanism and the failure to solve what economists call the transfer problem, that is, the problem of converting German domestic tax revenue into foreign currency without wrecking German trade. 2.2 The diplomatic and legal tradition A second tradition studies the conference as a diplomatic and legal event. The essays collected by Badel, Conze, and Drober (2024) treat Paris 1919 not as a European bargaining session but as the construction of a new global order, with chapters on legal transformation, on the racialisation of the settlement, on Chinese and Brazilian diplomacy, and on feminist participation. Olechowski (2024) reminds students of a point that is constantly confused in popular writing: Versailles was only one of five Paris treaties, and it dealt only with Germany. Austria was covered by Saint-Germain, Hungary by Trianon, Bulgaria by Neuilly, and the Ottoman Empire by Sevres and later Lausanne. Much of what people blame on Versailles was in fact done by one of its sister treaties. Cohrs (2022) makes a stronger revisionist claim. In his account, the real Euro-Atlantic peace order was not founded at Versailles at all. It was founded later, in the London reparations settlement of 1924 and the #Locarno security pact of 1925, when British and American decision makers learned from the failures of 1919 and began integrating Germany rather than isolating it. On this reading, the settlement of 1919 was an unfinished draft, and the tragedy of the interwar period is that the improved version arrived too late and collapsed too quickly. 2.3 The critical and postcolonial tradition A third body of work has moved the study of 1919 out of Europe altogether. Manela's influential concept of a #Wilsonian_moment described how nationalist movements in Egypt, India, China, and Korea took Wilson's language of #self_determination seriously, appealed to Paris, were ignored, and drew radical conclusions from the disappointment. The most forceful recent statement of the critical position comes from Maiolo and Robson (2025), who argue that the League should not be understood as a flawed but well meaning experiment in global governance. In their reading, its overriding purpose was to secure control over the world's resources, populations, and armaments for the victorious Western powers, and it pursued that purpose through force and through the monopolisation of economic and military capacity rather than through persuasion. Its humanitarian and technical work, in this view, provided legitimacy for an essentially imperial project. Kripp (2022) supports part of this argument by examining the racial thinking embedded in the League vision of Jan Smuts, one of the Covenant's principal architects. Not all scholars accept this framing. Kaiga (2021), studying the British pro-League movement, shows that the idea of a league emerged from genuine wartime debate about how to prevent another catastrophe, and that its advocates were neither naive utopians nor simple agents of empire. Seidenfaden (2026), reviewing Maiolo and Robson, notes that the critical reading is powerful but risks flattening the internal conflicts inside the organisation. Students should treat this as a live and unresolved debate rather than a settled question. 2.4 The economic governance and sanctions tradition A fourth and newer tradition studies the settlement as the origin of modern global economic governance. Mulder (2022) traces how the blockade of the First World War was converted into a permanent peacetime instrument, the economic sanction, written into Article 16 of the Covenant. This was, at the time, a genuinely radical idea: an international organisation claiming the right to cut a state off from the world economy without firing a shot. Martin (2022) shows how reparations commissions, stabilisation loans, and financial supervision gave international bodies unprecedented authority over the internal budgets of sovereign states, and how strongly this was resented. Hinrichsen (2023) places German reparations in the long history of war indemnities and sovereign debt. Read together, this literature suggests that 1919 mattered less as a settlement of accounts with Germany than as the moment when a set of coercive economic tools entered permanent international practice. 2.5 The gap this article addresses Most student-facing accounts still rely on the older narrative and rarely integrate these four traditions. This article attempts a synthesis written in plain language, without assuming prior knowledge of international law or economics. 3. Methodology This is a qualitative, interpretive review article. It uses three methods. First, document analysis. The primary object of study is the text of the treaty itself, which is publicly available in full and runs to 440 articles across fifteen parts. Reading the actual text is the single most useful exercise a student can perform, because a surprising amount of popular commentary misdescribes what the clauses say. Article 231, the so-called #war_guilt_clause, is the clearest example. Second, historiographical synthesis. The article compares interpretations across the four traditions identified above, giving weight to work published since 2021 while noting where recent scholarship departs from older consensus. Third, comparative reasoning. The treaty is assessed against other peace settlements, notably the Treaty of Frankfurt of 1871 which Germany imposed on France, the Treaty of Brest-Litovsk of 1918 which Germany imposed on Russia, and the post-1945 settlement with Germany. Comparison is used to test claims about harshness, because "harsh" is a relative word and needs a benchmark. Limitations of the method are acknowledged in Section 15. The article relies on published secondary scholarship and does not conduct archival research. It also privileges Anglophone and European scholarship, which remains a structural bias in the field, although the recent literature has begun to correct this. 4. Historical Background: From Armistice to Conference 4.1 The end of the fighting The military situation in the autumn of 1918 is essential background, and it is often misunderstood. Germany was not invaded. When the guns stopped on 11 November 1918, German troops still stood on French and Belgian soil, and no Allied soldier had entered Germany. But the German army was breaking. The Allied offensives from August 1918 had shattered the Hindenburg Line, Germany's allies were collapsing one by one, the naval blockade had produced severe civilian hardship, and the high command itself told the Kaiser that the war was lost. What happened next shaped everything. Rather than accepting defeat publicly, the German military leadership handed power to civilian politicians and let them sign the #armistice. Within months, right wing propagandists were arguing that the army had never been beaten in the field and had instead been betrayed at home by socialists, liberals, and Jews. This was the #stab_in_the_back myth, and McDonough (2023) treats it as more corrosive to the Weimar Republic than the treaty itself. The myth had no basis in military reality, but it did not need one. It offered a psychologically satisfying explanation for a humiliation that most German civilians had not seen coming, because wartime censorship had fed them a diet of victories almost to the end. 4.2 The Fourteen Points and the question of a promised peace Germany did not surrender unconditionally. It sought an armistice on the basis of Woodrow Wilson's #Fourteen_Points, a programme announced in January 1918 which called for open diplomacy, freedom of the seas, arms reduction, the evacuation of occupied territory, adjustment of colonial claims with regard for the interests of colonised peoples, and a general association of nations. This created a serious problem of legitimacy. Germans believed, with some justification, that they had accepted a specific offer. What arrived at Versailles seven months later looked very different. The Allies had entered reservations to the Fourteen Points in the pre-armistice negotiations, particularly on reparation for damage and on freedom of the seas, and Wilson himself had been forced to trade away points to keep the coalition together. But in German public opinion the gap between the promise of 1918 and the document of 1919 became evidence of a swindle. Historians of the settlement generally agree that this perception, whether or not it was fair, did more political damage than the specific burdens the treaty imposed. 4.3 The world that had to be rebuilt By January 1919 the map of Europe and the Middle East had simply dissolved. The Russian Empire had collapsed into revolution and civil war, and the Bolshevik government was not invited to Paris. Austria-Hungary had disintegrated into successor states that were already fighting each other over borders. The Ottoman Empire was in the process of partition. New states were declaring independence and seizing territory while the conference was still sitting; the peacemakers frequently found themselves ratifying facts created by soldiers on the ground rather than deciding them. Three further pressures shaped the conference. The #influenza_pandemic was killing millions and infected Wilson himself during the negotiations. Revolutionary movements inspired by Bolshevism were active in Berlin, Munich, and Budapest, and fear of #Bolshevism ran through the discussions as an unspoken deadline. And the Allied blockade of Germany was maintained until the treaty was signed, which caused continuing civilian hardship and gave the Germans a further grievance while giving the Allies their principal instrument of leverage. 5. The Paris Peace Conference: Actors, Procedures, and Aims 5.1 Who decided The conference formally opened on 18 January 1919 and involved delegations from around thirty countries. In practice, decisions were made by a very small group. The Council of Ten gave way to the #Council_of_Four: #Woodrow_Wilson for the United States, David Lloyd George for Britain, Georges Clemenceau for France, and Vittorio Orlando for Italy, with Orlando marginal and often absent. Japan sat among the great powers but concentrated on Pacific and Chinese questions. This concentration of authority matters for two reasons. It explains the speed and the incoherence of some decisions, since three exhausted men working through interpreters could not possibly master the technical detail of every border. It also explains the legitimacy problem. The defeated states were not permitted to negotiate. Germany received the draft in May 1919, was allowed to submit written observations, obtained a small number of concessions, and was then told to sign under threat of resumed war and continued blockade. This procedure, known as a #diktat in German political language, converted the treaty into a national humiliation independent of its contents. 5.2 What each power wanted France wanted security above everything. It had been invaded by Germany twice in fifty years, had lost roughly 1.4 million dead, and had fought most of the Western Front on its own territory. Clemenceau wanted the Rhineland detached or permanently occupied, German military power destroyed, and full compensation for physical destruction. French aims were not simply vindictive, as Keynes suggested and as later scholarship has largely rejected. They were the aims of a country that expected Germany to recover its economic and demographic superiority within a generation and wanted structural guarantees before that happened. Britain was ambivalent. Lloyd George had won an election on promises to make Germany pay, and public opinion demanded it. But British strategists did not want France to dominate the continent, did not want Germany to collapse into Bolshevism, and did want German markets restored for British exports. The result was a policy that oscillated between severity and moderation, and the treaty reflects that oscillation. The United States under Wilson wanted a new international system. Wilson's priority was the #League_of_Nations, and he was prepared to concede on other questions to secure it. He believed the League would be able to correct the injustices of the settlement over time, which turned out to be one of the more consequential misjudgements of the twentieth century. Italy wanted the territory it had been promised in the secret Treaty of London of 1915. When it did not receive all of it, the Italian delegation walked out, and the resulting sense of a mutilated victory fed directly into the rise of Fascism. This is worth emphasising to students: the settlement produced dangerous resentment among victors as well as among the defeated. Japan sought confirmation of its seizure of German rights in Shandong, and it proposed a racial equality clause for the Covenant. The clause won a majority of votes in the commission but was ruled to require unanimity and was dropped, largely because of resistance from the United States and the British Dominions, especially Australia. The rejection of that clause is a central piece of evidence for scholars such as Kripp (2022) who read a racial hierarchy into the foundations of the League. 5.3 The problem of self-determination Wilson's principle of self-determination was never precisely defined, and in practice it could not be applied consistently. Central and Eastern Europe was not a mosaic of neatly separated nations; it was a patchwork of mixed villages, mixed cities, and mixed families. Any border drawn there would leave minorities on the wrong side. The peacemakers responded with two devices. Where possible, they used #plebiscites, holding votes in disputed areas such as Upper Silesia, Schleswig, and the Saar basin. Where that was impossible or inconvenient, they imposed #minority_treaties on the new and enlarged states of Central and Eastern Europe, obliging them to guarantee the rights of national and religious minorities under League supervision. Leszczynski (2023) argues that this minorities regime functioned as an instrument of legal and political modernisation for the states bound by it, but he also notes its central weakness: the obligations were imposed only on the weaker states, never on the great powers, which made them look less like universal rights and more like a badge of inferiority. The deeper inconsistency was geographical. Self-determination was applied, imperfectly, to Europeans. It was not applied to Germany's colonies, to the Arab provinces of the Ottoman Empire, to India, to Egypt, to Korea, or to Vietnam. Petitions from those regions arrived in Paris and were set aside. Section 11 returns to the consequences. 6. The Terms of the Treaty The treaty runs to 440 articles. This section summarises the provisions that matter most for analysis. 6.1 Territorial provisions Germany lost roughly thirteen percent of its European territory and about ten percent of its population. The main changes were: Alsace-Lorraine returned to France, reversing the annexation of 1871. Eupen and Malmedy transferred to Belgium after a contested consultation. North Schleswig transferred to Denmark after a plebiscite. The Polish Corridor and most of Posen transferred to the reconstituted state of #Poland, giving it access to the sea and separating East Prussia from the rest of Germany. Danzig, a majority German city, became a Free City under League of Nations administration rather than being given to Poland or kept by Germany. Memel was detached and later seized by Lithuania. Upper Silesia was partitioned after a plebiscite and subsequent fighting, with the industrial districts largely going to Poland. The Saar was placed under League administration for fifteen years, with its coal mines assigned to France as compensation for the deliberate destruction of French mines, after which a plebiscite would decide its future. The 1935 plebiscite returned it to Germany by an overwhelming margin. All German colonies in Africa, the Pacific, and China were surrendered and redistributed as League #mandates. Union with Austria was prohibited, which sat awkwardly beside the principle of self-determination, since many Austrians wished for it. Two points deserve emphasis. First, Germany remained, after all these losses, the largest and potentially most powerful state in Central Europe, with an intact industrial core in the Ruhr, a unified state structure, and no permanent occupation of its heartland. Compare this with what Germany had imposed on Russia at Brest-Litovsk in March 1918, which stripped away roughly a third of the population of the Russian Empire, or with what happened to Germany itself in 1945, which involved total occupation, partition, and the loss of far more territory. Second, the losses were concentrated in areas of mixed population and long historical dispute, which meant that every one of them generated a revisionist grievance that a future German government could exploit. 6.2 Military provisions Part V of the treaty disarmed Germany: The army was capped at 100,000 men, with no conscription, no tanks, no military aircraft, and no heavy artillery. The navy was limited to a small force with no submarines. The interned High Seas Fleet was scuttled by its own crews at #Scapa_Flow in June 1919. The General Staff was abolished. The #Rhineland was permanently #demilitarised and occupied by Allied forces for fifteen years. The stated purpose was to make German #aggression physically impossible and to serve as the first step toward general disarmament, a promise contained in the preamble to Part V and repeated in the Covenant. That general disarmament never came. Germany could therefore argue, with some force, that it had been disarmed under a promise that the victors then broke, and this argument was used to justify German rearmament in the 1930s. An underappreciated point, noted in the economic history literature, is that the restriction on German military spending in the 1920s actually released considerable resources for civilian use. The military clauses were a constraint on German power but not, in any simple sense, a burden on German prosperity. 6.3 Article 231 and the reparations settlement #Article_231 is the most misquoted sentence in modern diplomatic history. It states that Germany and its allies accept responsibility for causing all the loss and damage suffered by the Allied governments and their nationals as a consequence of a war imposed on them by the aggression of Germany and its allies. Three things about this clause are essential. First, the words "guilt" and "war guilt" do not appear. The clause was drafted by American lawyers, including John Foster Dulles, as a legal preamble establishing theoretical liability. Article 232 immediately follows and limits actual liability, acknowledging that Germany's resources are not adequate to make complete reparation and restricting the claim to specified categories of civilian damage. Article 231 was, in the Allies' own understanding, a preamble clause that made the limitation in Article 232 legally coherent. Second, an identical clause with the name changed appeared in the treaties with Austria, Hungary, and Bulgaria. It was standard drafting, not a special insult aimed at Germany. Third, none of this mattered politically. German commentators read Article 231 as a moral verdict, translated it as the war guilt clause, and built an entire propaganda industry around refuting it. The German Foreign Ministry funded a well resourced campaign of historical revisionism, the so-called war guilt section, which produced document collections designed to demonstrate German innocence. The clause became the emotional centre of German objections to the treaty, which is a striking demonstration that in international politics the perceived meaning of a text can matter more than its legal content. The financial settlement itself was not fixed at Versailles. The treaty established a #Reparation_Commission, which reported in 1921 with the London Schedule of Payments, setting the total at 132 billion gold marks. Even this headline number was partly presentational: it was divided into A, B, and C bonds, and the C bonds, which made up the bulk of the sum, were widely understood to be unlikely ever to be collected. The realistic figure was closer to 50 billion gold marks. 6.4 Other provisions The treaty also required Germany to hand over war criminals for trial, including the former Kaiser, who had taken refuge in the Netherlands and was never surrendered. A small number of trials were eventually held before the German Supreme Court at #Leipzig, produced few convictions, and are usually treated as a failure, although recent legal scholarship reads them as an early and instructive episode in the history of international criminal justice. Article 238 required the restitution of objects removed from occupied territory. Lincoln (2024) has shown that this comparatively obscure clause helped consolidate norms protecting civilian property in wartime and gave private individuals a voice in international claims, even though the restitution process itself was slow, politically fraught, and shaped by social class. Articles 297 and 298 legalised the wartime seizure of German property abroad, a set of provisions that Mulder (2022) situates within the emergence of economic warfare as a permanent instrument of statecraft. 7. The League of Nations: Design and Institutions 7.1 Origins of the idea The League was not invented at Paris. Kaiga (2021) traces its intellectual origins to British wartime debate, where lawyers, journalists, and campaigners argued from 1914 onward that the war had exposed the failure of the old balance of power system and that some permanent machinery was needed to prevent a repetition. Similar movements existed in the United States, France, and elsewhere. What Paris did was to select among competing designs. The version that emerged owed a great deal to Wilson, to the British official Robert Cecil, and to the South African statesman Jan Smuts, whose memorandum shaped the mandate system. Kripp (2022) argues that Smuts's vision cannot be separated from his racial thinking, and that the Covenant's civilisational language was not incidental decoration but part of the design. 7.2 Structure The #Covenant, which forms Articles 1 to 26 of the Treaty of Versailles, established: An Assembly, in which every member state had one vote. A Council, with permanent seats for the principal Allied powers and rotating seats for others. A permanent Secretariat based in #Geneva, headed by a Secretary-General. This was a genuine innovation: an international civil service, staffed by officials who were supposed to serve the organisation rather than their own governments. The #Permanent_Court_of_International_Justice, established under Article 14, which became the direct ancestor of the current International Court of Justice. The #International_Labour_Organization, created by Part XIII of the treaty, which survives today as a specialised agency of the United Nations and is arguably the settlement's most durable institutional product. The critical decision rule was #unanimity. Both the Council and the Assembly generally required unanimous decisions on substantive matters, excluding the parties to a dispute. This was intended to protect sovereignty. In practice it meant that any determined state could block collective action. 7.3 The security machinery #Article_10 obliged members to respect and preserve the territorial integrity and political independence of all members against external aggression. #Article_16 was the enforcement clause: a member resorting to war in disregard of its obligations would be deemed to have committed an act of war against all other members, who would immediately sever trade and financial relations with it. Mulder (2022) shows how radical this was. Article 16 converted the wartime blockade, an instrument that had caused enormous civilian suffering, into a standing peacetime threat. The theory was deterrence: no state would risk being cut off from the world economy. The practice was more complicated. #Sanctions worked reasonably well against small states and not at all against great powers, they were slow, they hurt the states applying them, and governments proved extremely reluctant to use them in earnest. 7.4 The absent members The League began with a structural wound. The United States Senate refused to ratify the treaty, and the #United_States never joined the organisation its own president had designed. The reasons were partly constitutional, centring on whether Article 10 would commit American forces to war without congressional approval, and partly political, involving a bitter personal conflict between Wilson and Senator Henry Cabot Lodge. Wilson's refusal to compromise on reservations contributed directly to the defeat. Germany was excluded until 1926 and left in 1933. Soviet Russia was excluded until 1934 and expelled in 1939. Japan left in 1933, Italy in 1937. For most of its life, the League was therefore an organisation that lacked at least two and often four of the world's most significant military powers. Any assessment of its record must begin from that fact. 8. The Mandate System and the Colonial Question #Article_22 of the Covenant created the mandate system. Former German colonies and former Ottoman Arab provinces were not returned, not made independent, and not formally annexed. They were assigned to Allied powers as mandates, to be administered on behalf of the League as a "sacred trust of civilisation", with the mandatory power reporting to a #Permanent_Mandates_Commission in Geneva. Mandates were divided into three classes. Class A, covering the Arab territories of the former Ottoman Empire, was described as close to independence; these included the British mandates in Palestine and Mesopotamia and the French mandates in Syria and Lebanon. Class B, covering most of the former German territories in tropical Africa, involved direct administration. Class C, covering South West Africa and the Pacific islands, was administered as if the territory were an integral part of the mandatory power. The official justification was tutelage: peoples said not yet able to stand by themselves under the strenuous conditions of the modern world would be guided toward self-government by advanced nations. The reality, as Maiolo and Robson (2025) argue with some force, was that the mandate system distributed the spoils of the war while dressing the transfer in a new legal vocabulary. Class C mandates in particular were annexation in all but name. The petitions system did give colonised subjects a channel to address an international body, and that channel was used, sometimes skilfully. But the Commission had no enforcement power, and no Class B or C mandate achieved independence through the League. For students, the analytical point is this: the mandate system is where the tension inside the 1919 settlement becomes visible. The same document that proclaimed self-determination created a legal mechanism for administering millions of people who had not consented to be administered. Whether one treats this as hypocrisy, as a necessary compromise, or as the true purpose of the exercise is the central interpretive question in the current literature. 9. Economic Consequences and the Reparations Debate 9.1 What Germany actually paid The gap between the reparations demanded and the reparations received is one of the most important and least known facts about the settlement. Germany paid substantially less than the headline figures suggest. Payments were suspended, rescheduled, reduced, and eventually abandoned. The sequence is worth memorising: 1921: London Schedule of Payments sets the nominal total at 132 billion gold marks. 1923: Germany defaults on deliveries. France and Belgium occupy the #Ruhr. The German government responds with passive resistance, funds it by printing money, and #hyperinflation destroys the currency. 1924: The #Dawes_Plan reorganises payments, stabilises the currency, and channels large American loans into Germany. 1929: The #Young_Plan reduces the total and extends payments over fifty-eight years. 1932: The #Lausanne_Conference effectively cancels reparations. 1933: Hitler repudiates what remains. 9.2 The scholarly assessment Ho (2025) argues that the Dawes Plan solved the immediate monetary crisis but created a new problem: Germany financed its reparations largely by borrowing from the United States, so that when American capital was withdrawn after 1929, the whole structure collapsed and the German economy fell further than it otherwise would have. The plan, on this reading, was never incentive compatible; it lacked serious enforcement and never addressed the transfer problem. Most economic historians now hold that reparations were within Germany's capacity to pay in a narrow accounting sense, and that the decisive obstacle was political. No German government could survive being seen to pay willingly. Fulfilment was therefore always a policy of fulfilling in order to demonstrate impossibility. The claim that reparations directly caused the hyperinflation of 1923 does not survive scrutiny either: the inflation was driven by wartime deficit financing, by the political decision to fund passive resistance in the Ruhr with the printing press, and by the collapse of confidence in the currency. This does not make the reparations regime wise. Martin (2022) shows that the intrusive financial supervision which accompanied stabilisation, including foreign control over German railways and the central bank, was experienced as a humiliating loss of sovereignty and generated exactly the nationalist backlash it was meant to prevent. Hinrichsen (2023) situates the episode in a longer history and draws the same practical lesson: reparations that a defeated state cannot politically accept will not be paid, whatever the arithmetic says. 9.3 The Keynes question Was Keynes right? The centenary assessment in Clavin, Corsetti, Obstfeld, and Tooze (2024) suggests a nuanced answer. He was wrong about the arithmetic, or at least he exaggerated it. He was wrong to portray French aims as purely vindictive. He was right that Europe's economies were interdependent and that impoverishing Germany would impoverish its neighbours. He was right that the settlement contained no mechanism for European economic reconstruction. And he was right, in a broader sense, that the contest between narrow nationalism and economic integration would define the century. He was also, unintentionally, enormously damaging. By giving the German revisionist case an eloquent English voice, he helped persuade British and American opinion that the treaty was unjust, which in turn made those governments reluctant to enforce it. A treaty that the enforcers no longer believe in is not a treaty at all. 10. Germany and the Politics of Resentment The #Weimar_Republic was born carrying a corpse. It had to sign the treaty, and it was blamed for signing it. The German delegation's leader resigned rather than accept the terms; the National Assembly voted to sign only under the threat of invasion and the continued blockade. McDonough (2023) argues that the stab in the back myth was more destructive than the treaty's material burdens, and that the myth attached itself to the treaty rather than the other way round. The people who signed were labelled the #November_criminals. Matthias Erzberger, who signed the armistice, was assassinated in 1921. Walther Rathenau, the foreign minister who pursued accommodation with the Allies, was assassinated in 1922. The politics of the Republic were poisoned from the start by the accusation that democracy itself was a foreign imposition and a national betrayal. #Adolf_Hitler built his early career on this. #Versailles appears constantly in his speeches, not as an economic argument but as a symbol of national shame requiring national revenge. It is important to be precise here. The treaty did not cause Nazism in any mechanical sense. Germany was not economically crushed; by 1928 its industrial output exceeded 1913 levels. What destroyed the Republic was the #Great_Depression, which arrived in 1929 and doubled unemployment within three years. Before 1930 the Nazi Party was a fringe movement. What the treaty supplied was not the crisis but the story: a ready made explanation of who was to blame and what national salvation would look like. The distinction matters for students. It is the difference between saying that Versailles caused the Second World War, which is too simple to be useful, and saying that Versailles supplied the grievance that a movement later used, which is defensible and much more precise. 11. Global Reactions Beyond Europe One of the most productive shifts in recent scholarship has been the recognition that the settlement was a global event with global consequences. China. The transfer of German rights in #Shandong to Japan rather than their return to China provoked the #May_Fourth_Movement of 1919: mass student protests, boycotts, and an intellectual awakening that reshaped modern Chinese politics. China refused to sign the treaty. Both the Chinese Nationalist and Communist traditions trace part of their origins to that betrayal, and the episode is a reminder that Wilsonian rhetoric had consequences its author never intended. Korea. The #March_First_Movement of 1919 appealed to self-determination against Japanese colonial rule. It was suppressed, and Korea received no hearing at Paris. Egypt. The 1919 revolution followed the refusal to allow an Egyptian delegation to travel to Paris to argue for independence from Britain. India. Indian nationalists who had supported the war effort expected political concessions and received the repressive Rowlatt Acts and the #Amritsar_massacre instead. The Arab world. The wartime promises made to Arab leaders were incompatible with the #Sykes_Picot agreement between Britain and France and with the #Balfour_Declaration. The mandate system in Syria, Lebanon, Palestine, and Mesopotamia resolved that incompatibility in favour of the imperial powers. The consequences are still being lived with. Japan. The rejection of the racial equality proposal, combined with a sense that Japan was treated as a junior partner, fed a lasting resentment among Japanese elites toward the Anglo-American order. Manela's argument that this collective disappointment produced a generation of anticolonial radicals is now widely accepted, and the essays in Badel, Conze, and Drober (2024) extend it, showing how delegations from China, Brazil, and elsewhere used the conference as a stage even when they could not shape its outcome. Mona Siegel's work on feminist diplomacy at Paris, included in that volume, further widens the cast of actors beyond the Council of Four. 12. The League in Practice, 1920 to 1946 12.1 What worked The League is usually taught as a failure, and in its central security mission it was one. But the record is more mixed than the caricature allows, and students who write only about failure will miss half the evidence. The organisation settled a number of minor #disputes: between Sweden and Finland over the Aaland Islands in 1921, between Greece and Bulgaria in 1925, and in the Upper Silesian partition. It administered the Saar and #Danzig with reasonable competence. It ran a #refugee_regime under Fridtjof #Nansen that produced the Nansen passport, the first international travel document for stateless persons. It supported the financial reconstruction of Austria and Hungary in the 1920s. Its health organisation and its work against trafficking, slavery, and epidemic disease built technical networks that outlived the organisation itself, and much of that machinery was transferred directly to the United Nations and its agencies. Maiolo and Robson (2025) are sceptical of celebrating these achievements, arguing that the technical and humanitarian work provided a respectable cover for an essentially imperial project. Students should engage with that argument seriously, but they should also note that the machinery in question is still in use, which suggests that at least some of it answered real needs. 12.2 What failed The security system failed in three stages. Manchuria, 1931. Japan invaded Chinese #Manchuria. The League commissioned an inquiry, the #Lytton_Commission, which took over a year and concluded that Japan had been the aggressor. The Assembly adopted the report. Japan withdrew from the League and kept Manchuria. No sanctions were applied. Abyssinia, 1935. Italy invaded #Ethiopia, a member state. The League declared Italy an aggressor and imposed economic sanctions, but the sanctions excluded oil, and the Suez Canal was not closed to Italian shipping. Britain and France, meanwhile, were secretly negotiating to hand Mussolini most of what he wanted in the #Hoare_Laval_pact, which leaked and destroyed what remained of the League's credibility. Italy conquered Ethiopia. Sanctions were lifted. Rhineland, 1936 and after. Germany remilitarised the Rhineland in violation of both Versailles and Locarno. There was no effective response. From that point the settlement of 1919 existed only on paper, and Hitler dismantled the rest of it in stages. 12.3 Why it failed Several explanations are usually offered, and they are complementary rather than competing. The membership problem: the absence of the United States, and the intermittent membership of Germany, the Soviet Union, Japan, and Italy, meant the League never represented the actual distribution of world power. The decision rule problem: unanimity gave every state a veto, and the Covenant provided no reliable procedure for rapid action. The enforcement problem: the League had no armed force. It relied on member states, and member states acted on national interest. Sanctions were the only real weapon, and, as Mulder (2022) demonstrates, governments feared that using them seriously against a great power might cause the very war they were meant to prevent. The great power problem, which is the deepest of the three: Britain and France were not willing to fight in 1931 or 1935. They were financially exhausted, politically divided, facing severe domestic economic crisis, and in Britain's case governed by an opinion that had come to regard the settlement of 1919 as unjust and therefore not worth defending. Collective security requires collective willingness. It was never there. 13. Discussion: Was the Treaty Too Harsh, Too Soft, or Simply Unenforceable? The oldest debate about the treaty asks whether it was too severe. The current scholarly answer, broadly, is that this is the wrong question. The case for severity. Germany lost territory, population, colonies, its armed forces, and its financial autonomy, and was made to accept formal responsibility for the war. The manner of imposition, without negotiation and under blockade, was humiliating. The occupation of the Ruhr in 1923 and the intrusive financial supervision that followed were experienced as national degradation. The case for leniency. Compare the terms Germany itself imposed on Russia at #Brest_Litovsk a year earlier, which were far more extreme. Compare the #Treaty_of_Frankfurt of 1871, in which a victorious Germany annexed Alsace-Lorraine and extracted a large indemnity from France, an indemnity France paid quickly and without the country collapsing. Compare 1945, when Germany was occupied, divided, and stripped of far more territory. Against all three benchmarks, Versailles was moderate. Germany remained united, industrially intact, unoccupied in its core, and surrounded by weak new states rather than strong old empires. In purely strategic terms, Germany's position in Central Europe was arguably better in 1919 than in 1914, because Russia and Austria-Hungary had ceased to exist as great powers. The synthesis. The treaty was harsh enough to enrage Germany and not harsh enough to disable it. It was severe in symbol and moderate in structure. This combination, sometimes called the fatal middle path, is the core of the modern critique. And crucially, it depended on continuous enforcement by a coalition that dissolved almost immediately: the United States withdrew, Britain lost conviction, and France was left to enforce alone a settlement it could not enforce alone. Cohrs (2022) pushes this further, arguing that a workable order was in fact constructed after 1919, through the London agreement of 1924 and Locarno in 1925, and that its failure was a product of the Depression rather than of the treaty's clauses. On this reading, Versailles was not a doom machine. It was a flawed draft that was subsequently improved and then destroyed by an economic catastrophe nobody at Paris could have anticipated. A final consideration. Was any better settlement politically available in 1919? Public opinion in Britain and France demanded punishment. The soldiers who had survived and the families who had not expected compensation. A conciliatory peace of the kind Keynes wanted would have required leaders willing to lose elections. This is the constraint that students tend to overlook, and it is the reason the counterfactual "they should simply have been more generous" is weaker than it looks. 14. Legacy: From Geneva to New York The settlement's most durable legacy is institutional. The idea that permanent international organisation is normal, that states meet in a standing assembly, that a secretariat serves the international community, that a world court adjudicates disputes, that labour standards and health and refugees are proper subjects of international administration: all of this begins with the Covenant. The architects of the United Nations in 1944 and 1945 studied the League's failures closely and designed against them. The Security Council replaced unanimity with a system in which five permanent members hold vetoes but decisions can otherwise be taken by qualified majority, an attempt to reconcile power with procedure. Chapter VII provides for enforcement action, including military force, rather than sanctions alone. The Charter was separated from the peace treaties, so that membership of the organisation was not a referendum on any particular settlement, an explicit lesson from the American rejection of 1919. The great powers were brought in from the beginning, and the United States hosted the organisation. The trusteeship system replaced the mandates and was explicitly directed toward independence, which in practice was achieved. Whether these changes solved the underlying problem is a fair question. The #Security_Council veto has paralysed the #United_Nations repeatedly. Enforcement still depends on the will of powerful states. Maiolo and Robson (2025) argue that the UN inherited the League's function as a vehicle for the projection of dominant power, and that it continues to serve it. Students should notice that the debates of 1919 are not archived; they are current. 15. Limitations and Directions for Future Research This article is a synthesis of published scholarship and does not present original archival research. It relies mainly on English-language work, which under-represents German, French, Polish, Turkish, Chinese, Japanese, and Arabic scholarship on the same events. Its treatment of the four sister treaties is brief. Four directions look promising for student research projects. First, the non-European archives of 1919. What did delegations from China, Brazil, Liberia, or the Hejaz actually say and do at Paris, and how were their arguments recorded and dismissed? Second, the economic governance question. Following Martin (2022) and Mulder (2022), how did the tools of financial supervision and economic sanction developed after 1919 shape the practice of international institutions down to the present? Third, the minorities regime. Leszczynski (2023) opens a comparative question: did the League's minority protection system actually protect anyone, and did it modernise the legal systems of the states bound by it, or did it mainly stigmatise them? Fourth, the memory politics of the treaty. How did Versailles become, in German, Italian, Japanese, Chinese, and Arab political culture, a shorthand for injustice, and how does that shorthand still operate? 16. Conclusion The Treaty of Versailles ended a war and began an argument that has not stopped. It redrew the map of Europe, disarmed Germany, transferred colonies under a new legal form, and established the first permanent general international organisation. Each of these acts was contested at the time and remains contested now. Three conclusions follow from the evidence reviewed here. First, the treaty was less severe than its reputation and less effective than its designers hoped. Its material terms were survivable; Germany survived them and by the late 1920s had largely recovered. Its symbolic terms were intolerable to German opinion, and its enforcement depended on a coalition that had already begun to dissolve when the ink dried. A settlement that is resented but not enforced is the worst of both worlds. Second, the #League_of_Nations should be judged on two registers rather than one. As a security organisation it failed, for reasons that were built into its membership, its decision rules, and above all the unwillingness of its leading members to fight for it. As an administrative and technical body it did more than it is usually credited with, and its institutional descendants are still working. Whether that technical work redeems the organisation or merely legitimated an imperial order is the sharpest question in the current literature, and students should form their own view rather than accept either answer ready made. Third, the deepest lesson of 1919 is not about clauses. It is about the relationship between #legitimacy and power. A settlement that the losers regard as fraudulent, and that the winners are unwilling to defend, will not hold, however carefully it is drafted. That lesson was expensive. It is also, unfortunately, the kind of lesson that each generation seems obliged to learn again. Hashtags #Treaty_of_Versailles #Versailles_1919 #Paris_Peace_Conference #League_of_Nations #World_War_One #Interwar_Europe #Article_231 #War_Reparations #Woodrow_Wilson #Weimar_Republic #Collective_Security #Mandate_System #International_Relations_History #Peace_Treaties #Twentieth_Century_History References Badel, L., Conze, E., and Drober, A. (Eds.). (2024). The Paris Peace Conference of 1919: The Challenge of a New World Order. New York and Oxford: Berghahn Books. https://doi.org/10.1515/9781805398097 Clavin, P., Corsetti, G., Obstfeld, M., and Tooze, A. (Eds.). (2024). Keynes's Economic Consequences of the Peace after 100 Years: Polemics and Policy. Cambridge: Cambridge University Press. Cohrs, P. O. (2022). The New Atlantic Order: The Transformation of International Politics, 1860-1933. Cambridge: Cambridge University Press. Hinrichsen, S. (2023). When Nations Can't Default: A History of War Reparations and Sovereign Debt. Cambridge: Cambridge University Press. Ho, T. (2025). The Dawes Plan: A Centennial Retrospective and Re-Evaluation. Kyklos, 78(3), 977-990. https://doi.org/10.1111/kykl.12452 Kaiga, S. (2021). Britain and the Intellectual Origins of the League of Nations, 1914-1919. Cambridge: Cambridge University Press. Kripp, J. (2022). The Creative Advance Must Be Defended: Miscegenation, Metaphysics, and Race War in Jan Smuts's Vision of the League of Nations. American Political Science Review, 116(3), 940-953. https://doi.org/10.1017/S0003055421001362 Leszczynski, P. A. (2023). The system of protection of national and religious minorities in post-Versailles Central Europe as an instrument of legal and political modernisation. Rocznik Instytutu Europy Srodkowo-Wschodniej, 21(3), 215-234. https://doi.org/10.36874/riesw.2023.3.11 Lincoln, V. (2024). Restitution and the Treaty of Versailles: Restitution from Germany to France, 1918-1928. Contemporary European History, 34(2), 365-379. Mahmood, S. A. N. (2024). The Treaty of Versailles: Catalyst for Change and Conflict in the 20th Century. Sprin Journal of Arts, Humanities and Social Sciences, 3(6), 28-34. https://doi.org/10.55559/sjahss.v3i6.364 Maiolo, J., and Robson, L. (2025). The League of Nations. Elements in Modern Wars. Cambridge: Cambridge University Press. Martin, J. (2022). The Meddlers: Sovereignty, Empire, and the Birth of Global Economic Governance. Cambridge, MA: Harvard University Press. McDonough, F. (2023). The Weimar Years: Rise and Fall 1918-1933. London: Head of Zeus. Mulder, N. (2022). The Economic Weapon: The Rise of Sanctions as a Tool of Modern War. New Haven: Yale University Press. Olechowski, T. (2024). The Paris Peace Treaties. Iris Online Journal of Arts and Social Sciences, 2(2). Seidenfaden, E. E. (2026). Review of The League of Nations by J. Maiolo and L. Robson. Peace and Change. https://doi.org/10.1111/pech.70064

The General Treaty for the Renunciation of War as an Instrument of National Policy, signed in Paris on 27 August 1928 and known to history as the Kellogg-Briand Pact or the Pact of Paris, is one of the most misunderstood documents in the history of international relations. For most of the twentieth century it was dismissed as a naive gesture that failed the only test that mattered, because the Second World War began eleven years after its signature. This article argues that such a verdict, while emotionally understandable, is analytically incomplete. Drawing on recent legal scholarship on the prohibition of the use of force, on the prohibition of annexation, and on the revival of the crime of aggression, the study reconstructs the origins, text, negotiation, reservations, institutional weaknesses, and long term normative effects of the #Kellogg_Briand_Pact. It shows that the Pact did not stop the wars of the 1930s because it lacked definitions, institutions, sanctions, and a coherent theory of enforcement. It also shows, however, that the Pact performed a different and more durable function. It reversed the legal presumption in favour of war, supplied the doctrinal foundation for the charge of crimes against peace at Nuremberg and Tokyo, and provided the raw material later refined into Article 2(4) of the Charter of the United Nations. The article closes by using the Pact as a diagnostic tool for present day debates about the erosion of the prohibition on force, the legality of annexation, and the design of new accountability mechanisms such as the Special Tribunal for the Crime of Aggression against Ukraine. The central claim is that the Pact should be studied not as a failed peace treaty but as an unfinished constitutional moment whose defects were corrected slowly, unevenly, and at enormous human cost. Keywords: Kellogg-Briand Pact; Pact of Paris; outlawry of war; prohibition of the use of force; crime of aggression; interwar international law; Article 2(4); non-recognition; annexation; collective security 1. Introduction 1.1 The problem Very few international agreements have been mocked as thoroughly as the treaty signed in the Salle de l'Horloge at the French Foreign Ministry on 27 August 1928. Fifteen states, including the great powers of the day, solemnly declared that they condemned recourse to war and renounced it as an instrument of national policy in their relations with one another. Within four years Japanese forces had seized Manchuria. Within seven years Italy had invaded Ethiopia. Within eleven years Germany had invaded Poland and the deadliest conflict in human history had begun. The simplest reading of these facts is that the Pact accomplished nothing at all. That simple reading has become a kind of intellectual reflex. Textbooks describe the Pact as a monument to #interwar_idealism. Diplomatic historians treat it as a diplomatic evasion, a way for the United States to appear engaged in European security without accepting any obligation whatsoever. Political realists cite it as the classic demonstration that words on paper cannot restrain states that want to fight. This article does not deny the obvious. The Pact did not prevent the Second World War. But a treaty can fail in its immediate purpose and still transform the legal and moral vocabulary that later generations use to argue about war. The claim advanced here is that the Pact of Paris is best understood as a #normative_turning_point whose consequences were delayed, indirect, and only partly intended by its own authors. 1.2 Why the topic still matters The question is not merely antiquarian. The prohibition on the use of force between states is under visible strain. Recent scholarship on the prohibition of annexations describes a rule that is foundational to the modern legal order but poorly understood and now at genuine risk of erosion (Brunk and Hakimi, 2024). Analysts of the international system warn that both great and middle powers are challenging the prohibition on force in ways that would have been unthinkable a generation ago. Legal scholars continue to argue about what counts as a prohibited use of force at all, particularly in the cyber domain and in the grey zone of coercive operations below the threshold of armed attack (Pobjie, 2024; Buchan and Tsagourias, 2021). Every one of these debates has a genealogy that runs back through the Charter of the United Nations to the Pact of Paris. When a modern lawyer says that #aggressive_war is illegal, that statement is not a self evident truth about the world. It is the product of a specific historical decision taken in 1928 and then repaired, extended, and institutionalised after 1945. To understand how fragile or how durable that decision is, students must understand where it came from. 1.3 Research questions This article addresses five questions. First, what political and intellectual forces produced the Pact, and what did its drafters actually believe they were doing? Second, what does the text of the treaty say, and what did the accompanying diplomatic correspondence add to or subtract from that text? Third, why did the Pact fail to prevent the wars of the 1930s, and were those failures failures of design, of enforcement, or of political will? Fourth, what did the Pact contribute to the legal architecture that emerged after 1945, especially the criminalisation of aggression and the general prohibition of force? Fifth, what can students and policymakers learn from the Pact about the design of international rules that are meant to constrain the strongest actors in the system? 1.4 Structure of the article Section 2 reviews the scholarly literature and identifies the main interpretive camps. Section 3 explains the method and the sources. Section 4 reconstructs the legal world before 1928, in which war was a recognised right of sovereign states. Section 5 traces the diplomatic road to Paris. Section 6 provides a close reading of the text. Section 7 examines the reservations and understandings that qualified it. Section 8 analyses the institutional design failures. Section 9 tests the Pact against the crises of the 1930s. Section 10 examines Nuremberg and Tokyo. Section 11 traces the path from Paris to San Francisco. Section 12 assesses the Pact's living legacy in contemporary law. Section 13 sets out the theoretical debate. Section 14 offers an overall assessment and a set of design lessons. Section 15 states the limitations of the study. Section 16 concludes. 2. Literature Review 2.1 The dismissive tradition The oldest and still most widespread scholarly attitude is dismissal. In this tradition the Pact is treated as a symbol of the gap between legal aspiration and political reality. Writers in this camp emphasise that the treaty contained no definition of war, no definition of self defence, no machinery of enforcement, no court, no sanctions, and no procedure for determining who had violated it. They point out that states which signed the Pact went on to fight without ever formally declaring war, thereby dodging the letter of the obligation while destroying its spirit. On this reading the Pact did not outlaw war; it merely outlawed the declaration of war, and the effect was to drive hostilities into the convenient legal fog of the incident, the intervention, and the police action. This tradition draws strength from the writings of interwar critics who saw the flaw immediately. The realist critique associated with the diplomatic history of the period stresses that the Pact reassured public opinion without altering the balance of power, and that this false reassurance may have made #appeasement easier by encouraging the belief that a legal formula could substitute for military preparedness and credible alliances. 2.2 The revisionist tradition Against this stands a revisionist tradition which insists that the Pact mattered enormously, but not in the way its authors expected. The best known statement of this position holds that the outlawry of war in 1928 marked the decisive break between an old world order, in which war was a lawful means of enforcing rights, and a new world order, in which the resort to force became a wrong that could generate criminal responsibility, invalidate territorial gains, and justify collective response (Hathaway and Shapiro, 2017; Hathaway and Shapiro, 2019). The revisionist argument runs roughly as follows. Under classical international law, war was legal. Because war was legal, conquest transferred good title to territory, neutrality imposed strict duties of impartiality, and economic pressure short of war was suspect. Once war was declared unlawful, every one of these connected doctrines had to change. #Conquest could no longer produce valid title. Neutrality could give way to discrimination between aggressor and victim. Sanctions became conceptually available. The Pact, in this reading, is the hinge on which the entire structure turned. 2.3 The critical tradition A third body of work resists both the dismissal and the celebration. Critical and Third World approaches point out that the outlawry project was written by and for a small circle of mostly Western powers, and that the reservations attached to the Pact carefully protected colonial policing, imperial spheres of influence, and the freedom of powerful states to use force against peoples they did not recognise as sovereign. On this view the Pact did not abolish organised violence; it redistributed the legal labels attached to it, privileging the violence of empire and delegitimising only the violence of interstate rivalry among recognised states. Critical scholarship also questions the causal story told by the revisionists. If interstate conquest declined after 1945, that decline may owe more to decolonisation, nuclear deterrence, the economics of modern production, and the sheer cost of holding hostile populations than to a treaty signed in 1928. Empirical work on territorial conquest supports a nuanced picture in which the norm against conquest constrains large scale annexation of whole states far more effectively than it constrains small seizures of contested land (Altman, 2020; Altman and Lee, 2022). 2.4 The legal doctrinal tradition Finally, there is the technical legal literature that treats the Pact as a stage in the development of the modern law on force. Standard works on the prohibition of the use of force locate the Pact as the moment when international law began the passage from #jus_ad_bellum, the law regulating the right to go to war, to jus contra bellum, the law against war (Corten, 2021). Contemporary studies of what counts as a prohibited use of force still cite Article I of the Pact when interpreting the scope of the Charter rule, because the Charter language on international relations continues a tradition set in Paris (Pobjie, 2024). 2.5 The gap this article addresses Each tradition captures something real. The dismissive tradition is right about institutional design. The revisionist tradition is right about the long arc of legal doctrine. The critical tradition is right about who was protected and who was not. The doctrinal tradition is right about continuity of text. What is missing is a synthesis written for students that holds all four in view at once, and that treats the Pact not as a verdict to be delivered but as a case study in how international rules are made, broken, repaired, and reused. That is the gap this article addresses. 3. Method and Sources This is a qualitative, doctrinal, and historical study. It uses three kinds of material. The first is the primary legal text: the treaty itself, its preamble and three articles, together with the exchange of diplomatic notes that accompanied it and the reservations and understandings recorded by the negotiating states. These documents are treated as the authoritative statement of what was agreed. The second is the record of state practice between 1928 and 1945, in particular the Manchurian crisis, the doctrine of non-recognition, the Italian invasion of Ethiopia, and the charges of crimes against peace brought at Nuremberg and Tokyo. This practice is used to test the operational meaning of the Pact rather than its declared meaning. The third is the modern secondary literature on the prohibition of force, the prohibition of annexation, and the crime of aggression, drawn mainly from work published in the last five years. This literature is used to trace continuity and change between 1928 and the present. The method is interpretive rather than statistical. The article does not attempt to measure the Pact's causal effect on the frequency of war with quantitative tools. Instead it reconstructs the legal meaning of the instrument, follows the way that meaning was contested, and evaluates the arguments made by each interpretive camp against the documentary record. Where empirical claims about conquest and territorial change are relevant, the article reports the findings of quantitative researchers rather than generating its own data. This is a limitation, and it is acknowledged as such in Section 15. 4. The Legal World Before 1928: War as a Right 4.1 The classical position To appreciate what the Pact attempted, students must first accept a proposition that sounds shocking to modern ears. Before 1928, going to war was, in general, legally permissible. The classical law of nations regarded #war as a recognised procedure by which sovereign states enforced their claims against one another when negotiation failed. It was regrettable, expensive, and often unjust in the moral sense, but it was not unlawful. The great treatises of the nineteenth and early twentieth centuries expressed this with a certain calm. War was a condition regulated by international law, not a condition prohibited by it. The law had a great deal to say about how war should be fought, who counted as a lawful combatant, how prisoners should be treated, and what duties fell on neutrals. It had almost nothing to say about whether a state was entitled to begin fighting in the first place. 4.2 The consequences of legal war The legality of war was not an isolated rule. It was the keystone of an entire architecture. Because war was legal, #conquest was a valid mode of acquiring territory. A state that defeated another and annexed its provinces obtained good legal title, and other states were expected to recognise that title in due course. The map of the world, including many borders that survive today, was drawn on this assumption. Because war was legal, neutrality demanded strict impartiality. A neutral state could not favour the victim of an attack over the attacker without abandoning its neutral status and exposing itself to lawful reprisal. To help the wronged party was, paradoxically, to commit a wrong. Because war was legal, forcible measures short of war, such as reprisals, pacific blockades, and armed interventions to collect debts, occupied a permissive grey zone. If the greater act was allowed, the lesser could hardly be forbidden. 4.3 The first cracks The nineteenth and early twentieth centuries produced a growing movement of jurists, peace societies, and reformers who found this architecture intolerable. The Hague Peace Conferences of 1899 and 1907 produced conventions on the pacific settlement of disputes and created the Permanent Court of Arbitration. They regulated the conduct of hostilities in unprecedented detail. What they did not do was prohibit resort to war. Latin American states, whose experience of European and North American intervention was direct and painful, pressed early and persistently for rules against forcible territorial acquisition and against intervention. Recent historical scholarship emphasises that the duty of non-recognition of forcible territorial change, usually attributed to a United States Secretary of State in the 1930s, had important Latin American antecedents decades earlier (Brunk and Hakimi, 2024). This is a useful corrective to the habit of treating the outlawry of war as a purely Anglo-American invention. 4.4 The First World War and the Covenant The catastrophe of 1914 to 1918 destroyed the intellectual respectability of the classical position. Roughly ten million soldiers died. Empires collapsed. The idea that war was a rational instrument of policy, an extension of diplomacy by other means, became difficult to defend in public. The Covenant of the League of Nations, agreed in 1919, responded to this crisis but did not resolve it. The Covenant did not prohibit war. It imposed procedural delays. Members undertook to submit disputes likely to lead to rupture to arbitration, judicial settlement, or inquiry by the Council, and agreed not to resort to war until three months after the award or report. War after that cooling off period, and war against a state that refused to comply with a unanimous Council report, remained lawful. This is the crucial point for understanding what followed. The Covenant created a #procedural_gap, sometimes called the gap in the Covenant, through which a patient aggressor could walk. It regulated the timing of war rather than the right to wage it. Legal experts of the period said so plainly. The right of war survived Versailles, wounded but intact. 4.5 Locarno and the search for security The 1920s produced a series of attempts to close that gap. The Draft Treaty of Mutual Assistance of 1923 and the Geneva Protocol of 1924 both tried to link disarmament, arbitration, and security, and both failed, principally because Britain and its dominions refused open ended commitments to defend borders they did not care about. The Locarno treaties of 1925 succeeded where the broader schemes failed, but they succeeded by narrowing the ambition. Germany, France, and Belgium accepted the inviolability of their mutual western frontiers, with Britain and Italy as guarantors. Germany's eastern frontiers received no such guarantee, a silence that was noticed in Warsaw and Prague and remembered in Berlin. By 1927 the situation was as follows. The League existed but had not closed the gap. Locarno protected some borders but not others. The United States, the world's leading creditor and industrial power, stood outside the League entirely. Into this situation stepped a French foreign minister with a proposal, and an American peace movement with an idea. 5. The Road to Paris 5.1 The outlawry movement The intellectual origin of the Pact lies not in a chancellery but in the private campaign of an American lawyer, Salmon O. Levinson, who from 1918 onwards argued that the fundamental error of international law was that it treated war as a legal institution. Duelling had once been a legal institution too, he pointed out, and society had not abolished duelling by regulating it. Society had abolished duelling by declaring it a crime and stripping it of honour. Levinson therefore proposed that war be #outlawed: removed from the category of lawful acts, declared a crime, and replaced by compulsory adjudication. His campaign, conducted through the American Committee for the Outlawry of War, attracted an unusual coalition. It included pacifists and internationalists, but also senators such as William Borah of Idaho, an isolationist who distrusted the League intensely but liked the idea of abolishing war by declaration precisely because it required no entangling alliance and no commitment of American troops. That combination is essential to understanding both the achievement and the weakness of the Pact. #Outlawry appealed to Americans who wanted to end war and to Americans who wanted to avoid being drawn into other people's wars. The formula was popular because it was ambiguous. 5.2 The Briand initiative On 6 April 1927, the tenth anniversary of the American entry into the First World War, the French foreign minister Aristide Briand addressed a message to the American people. He proposed that France and the United States conclude an agreement renouncing war between themselves. Briand's motives were not primarily philosophical. France's central strategic problem was German revival. France wanted the United States entangled, however loosely, in the European security system. A bilateral pact renouncing war with the United States would, in effect, produce a moral obligation on Washington, and would make American neutrality in a future Franco-German conflict awkward. It was a diplomatic manoeuvre dressed in the language of peace, and it was a skilful one. 5.3 The American counter-move The United States government understood the manoeuvre perfectly and did not want the entanglement. Secretary of State Frank B. Kellogg, under sustained pressure from the outlawry movement and from a public opinion that had embraced the idea enthusiastically, found an elegant escape. Instead of rejecting the French proposal, Kellogg proposed to expand it. Let the renunciation of war be not a bilateral arrangement between Paris and Washington but a #multilateral_treaty open to all the states of the world. A universal renunciation of war would commit the United States to nothing in particular. It would create no alliance, no obligation of assistance, and no guarantee of any frontier. It would satisfy the peace movement, embarrass any state that refused to sign, and leave American freedom of action untouched. Briand was trapped by his own rhetoric. Having proposed the renunciation of war as a matter of high principle, he could hardly object when the principle was made general. France therefore accepted, while working hard through the negotiations to build in qualifications that would protect the League system, the Locarno guarantees, and France's network of alliances in Eastern Europe. 5.4 Negotiation and signature The negotiations occupied 1927 and 1928 and were conducted largely through exchanges of diplomatic notes rather than through a conference. This procedural choice mattered. It meant that the meaning of the treaty was constructed in correspondence, in statements of interpretation that were not incorporated into the text, and in speeches to national legislatures. On 27 August 1928, the General Treaty for the Renunciation of War as an Instrument of National Policy was signed in Paris by fifteen states: the United States, France, Germany, Belgium, the United Kingdom, Canada, Australia, New Zealand, South Africa, the Irish Free State, British India, Italy, Japan, Poland, and Czechoslovakia. The treaty was then opened for adherence, and most states of the world joined within a few years. Ratifications were exchanged and the Pact entered into force on 24 July 1929. The United States Senate approved it by 85 votes to 1, a margin that tells us a great deal about how little the Senate thought it was giving away. The Soviet Union, initially excluded from the negotiations, adhered and then went further, concluding in 1929 the protocol associated with Maxim Litvinov which brought the renunciation of war into force early among the Soviet Union and its neighbours. It is one of the ironies of the period that the state most feared by the Western signatories was among the most eager to be seen implementing the Pact. 6. The Text: A Close Reading 6.1 The preamble The preamble declares that the signatories are persuaded that the time has come for a frank renunciation of war as an instrument of national policy, that all changes in relations between them should be sought only by pacific means, and that any signatory which hereafter seeks to promote its national interests by resort to war should be denied the benefits furnished by the treaty. That final clause is the closest thing the Pact contains to a sanction. It is a #reciprocity_clause: a violator forfeits the protection of the treaty as against the other parties. In practical terms this means that if state A attacks state B in breach of the Pact, the other parties are released from their obligation not to make war on state A. The Pact thus quietly authorises what later law would call collective self defence, but it does so in the preamble, without machinery, and without any body competent to determine that a breach has occurred. 6.2 Article I Article I contains the central undertaking. The parties declare that they condemn recourse to war for the solution of international controversies, and renounce it as an instrument of national policy in their relations with one another. Three features of this sentence deserve close attention. First, the object of the renunciation is #war, not force. This is a narrower category. The Charter of the United Nations would later prohibit the threat or use of force, a formula deliberately chosen to capture hostilities that states decline to call war. In 1928 the drafters used the older word, and states promptly exploited the gap by fighting without declaring war. Second, the renunciation covers war as an instrument of national policy, and war for the solution of international controversies. It does not, by its terms, cover war undertaken for other purposes, such as war in self defence, war to enforce a decision of the League, or war against a state that has itself broken the Pact. Third, the phrase in their relations with one another limits the obligation to relations between the parties. Force used against non-parties, or against entities not recognised as states, sits outside the express prohibition. In the imperial world of 1928 this was not a small qualification. 6.3 Article II Article II provides that the settlement or solution of all disputes or conflicts, of whatever nature or origin, which may arise among the parties, shall never be sought except by pacific means. This is in some ways the more radical provision, and it is routinely overlooked. Article I renounces a means. Article II imposes a positive obligation of #pacific_settlement covering every dispute of every kind. Read together, the two articles do not merely say that war is forbidden. They say that the only permissible route to the resolution of a dispute is a peaceful one. The weakness is equally obvious. Article II specifies no method. It does not require arbitration, does not require judicial settlement, does not require conciliation, and does not create any tribunal with compulsory jurisdiction. A state can comply with Article II by proposing negotiations, refusing every compromise, and waiting. The obligation is real, but it is empty of procedure. 6.4 Article III Article III deals with ratification, deposit of instruments in Washington, adherence by other states, and entry into force. It is administrative rather than substantive, but two points are worth noting. The Pact was designed for universality: any state could adhere. And it contained no denunciation clause and no time limit, which is why it is still technically in force today, a treaty that has never been terminated and never been formally repudiated by its parties. 6.5 What the text does not contain A useful exercise for students is to list what a modern lawyer would expect to find in a treaty of this ambition, and then to observe its absence. There is no definition of war. There is no definition of aggression. There is no definition of self defence, and indeed no mention of self defence anywhere in the text. There is no organ empowered to determine that a violation has occurred. There is no obligation to assist a victim. There is no economic sanction. There is no reference to the League of Nations or to any existing institution. There is no dispute settlement clause. There is no compulsory jurisdiction of the Permanent Court of International Justice. There is no provision on the legal consequences of a breach, apart from the reciprocity idea in the preamble. The treaty is, in short, a declaration of principle wearing the clothes of a legal instrument. That was not an accident. It was the price of getting fifteen states, and eventually most of the world, to sign. 7. The Reservations and Understandings 7.1 Why they matter The text of the Pact is short and absolute. Its meaning, however, was fixed largely outside the text, in the diplomatic correspondence exchanged during the negotiations and in the statements made by governments when they signed and ratified. These qualifications are where the real bargain lies, and any serious analysis must engage with them. 7.2 Self defence The most important qualification concerns #self_defence. The Pact says nothing about it. Kellogg, however, explained repeatedly that the right of self defence was inherent in every sovereign state, that it required no express reservation because it was implicit in the very nature of sovereignty, and, critically, that every state remained free to decide for itself whether circumstances requiring self defence had arisen. That last element is the fatal one. An obligation not to wage war, subject to an exception for self defence, whose existence is determined solely by the state invoking it, is an obligation that a determined government can always escape. Contemporary critics said exactly this at the time, observing that the treaty allowed each nation to interpret for itself when self defence was involved, and that in consequence almost any military engagement could be justified. History proved them right. Japan described the seizure of Manchuria as a defensive response to Chinese provocation and to the protection of Japanese lives and railway interests. Italy justified the invasion of Ethiopia partly as a response to a border incident. Germany presented the invasion of Poland as a reply to Polish aggression, complete with a staged provocation. Every one of these states was a party to the Pact. Not one of them admitted to waging aggressive war. 7.3 The British reservation The United Kingdom attached what became known informally as the British Monroe Doctrine. Britain declared that there were certain regions of the world whose welfare and integrity constituted a special and vital interest for its peace and safety, that Britain had been at pains to make clear that interference with those regions could not be suffered, and that their protection against attack was a measure of self defence. Britain did not name the regions. The reservation was, in effect, a reserved sphere of imperial action of undefined geographical extent. Egypt and the approaches to the Suez Canal were widely understood to be among them. 7.4 The French position France insisted that the Pact must not conflict with its obligations under the Covenant of the League, under the Locarno treaties, or under its treaties of alliance and neutrality with other states. In practice this meant that war undertaken pursuant to League obligations, or pursuant to a guarantee of a frontier, was not renounced. 7.5 The American position The United States insisted that the Pact created no obligation to act against a violator, no commitment of any kind to the League system, and no impairment of the Monroe Doctrine. American ratification rested on the understanding that the treaty imposed no duty of enforcement whatsoever. 7.6 The combined effect Put the reservations together and the picture changes considerably. War is renounced, except in self defence, as defined by the state claiming it, except in defence of unspecified vital regions, except in performance of League obligations, except under regional guarantee treaties, except against a state that has already broken the Pact, and with no obligation on anyone to do anything about a violation. Some contemporaries concluded that after these qualifications there was no renunciation of war left at all. That judgment is too harsh, but it identifies the structural problem correctly. The Pact established a #rule_without_exceptions in its text and then permitted exceptions to be created unilaterally through interpretation. A rule whose exceptions are self judged is a rule that transfers legal authority from the collective to the individual state. This is the core design flaw, and it is worth stating as such because it recurs in international law to this day, most obviously in disputes about pre-emptive self defence. 8. Institutional Design: The Missing Machinery 8.1 No definitions Legal rules require categories. The Pact prohibited war but never said what war was. If two states send armies across a frontier, fight for months, and kill tens of thousands of people, but neither issues a declaration and both call the fighting an incident, has war occurred? In the 1930s this was not a hypothetical question. It was the standard practice. The drafters of the Charter of the United Nations learned this lesson and chose the broader phrase, the threat or use of force, precisely to defeat the semantic escape route. Modern scholarship devotes entire monographs to the meaning of a prohibited use of force, and traces the interpretive lineage directly back to the language of Article I of the Pact (Pobjie, 2024). The problem was inherited, but the Charter at least attacked it. 8.2 No institution The Pact created nothing. It established no council, no commission, no court, no secretariat, no reporting requirement, and no meeting of the parties. There was no forum in which a victim could formally accuse an aggressor, no body empowered to make a finding of violation, and no procedure for coordinating a response. This absence was deliberate. Any institution would have implied obligations, and obligations were precisely what the United States Senate would not accept. The Pact's universality and its emptiness were two sides of the same coin. It could be signed by everyone because it asked nothing of anyone. 8.3 No enforcement The Covenant of the League, whatever its weaknesses, contained an enforcement mechanism in Article 16, providing for economic and potentially military sanctions against a member that resorted to war in disregard of its obligations. The Pact contained nothing comparable. The most it offered was the preambular idea that a violator forfeits the benefits of the treaty. Enforcement therefore depended entirely on the willingness of individual states to act, and states in the 1930s were unwilling. The Great Depression arrived in 1929, one year after signature. Governments facing mass unemployment, banking collapse, and political extremism at home had neither the resources nor the domestic support for costly action abroad. The Pact was born into the worst possible economic and political conditions for #collective_security. 8.4 The paradox of universality There is a general lesson here that goes beyond 1928. There is often a trade-off in the design of international rules between the breadth of participation and the depth of obligation. The Pact chose maximum breadth and minimum depth. Almost every state in the world eventually adhered, and almost nothing was required of any of them. The Charter of the United Nations attempted a different bargain: a strong prohibition, a central organ with binding authority, and enforcement powers concentrated in a Security Council in which the great powers each hold a veto. That design purchased depth at the cost of paralysis whenever a permanent member is itself the accused. Neither design has solved the underlying problem, which is that the states with the greatest capacity to break the rule are also the states whose consent is needed to enforce it. 9. The Pact on Trial: Practice from 1929 to 1939 9.1 Manchuria, 1931 The first serious test came quickly. In September 1931, following an explosion on a railway line near Mukden staged by Japanese officers, the Japanese Kwantung Army occupied Manchuria and established the client state of Manchukuo in 1932. Japan was a party to the Pact and a member of the League. China appealed to both. The League appointed a commission of inquiry, chaired by Lord Lytton, whose report in 1932 concluded that the Japanese action was not a legitimate exercise of self defence and that Manchukuo was not the product of a genuine independence movement. The League Assembly adopted the report. Japan withdrew from the League in 1933. Manchukuo remained. 9.2 The doctrine of non-recognition The most creative legal response came from the United States. On 7 January 1932, Secretary of State Henry Stimson informed Japan and China that the United States would not recognise any situation, treaty, or agreement brought about by means contrary to the Pact of Paris. This #non_recognition doctrine is one of the Pact's most important offspring. Its logic is elegant. If war is unlawful, then the fruits of unlawful war cannot be lawful either. A conqueror may hold territory by force, but cannot acquire legal title to it. The principle is captured by the maxim that law does not arise from injustice. Two qualifications are essential. First, as recent scholarship stresses, the idea was not invented in Washington in 1932. Latin American jurists and governments had advanced the duty of non-recognition of forcible territorial change for decades before Stimson gave it his name, and the failure to acknowledge those origins distorts the history of the norm (Brunk and Hakimi, 2024). Second, non-recognition did not liberate Manchuria. Japan held it until 1945, and it took the total defeat of Japan in a world war to reverse the conquest. Non-recognition is thus a doctrine of #delegitimisation rather than of restitution. It denies the aggressor the legal validation of the gain. It does not take the gain away. That is precisely its character in modern practice, where the international response to forcible annexation still consists largely of declarations of non-recognition, and where the continuing occupation is not thereby ended. 9.3 The Chaco and the Latin American response The Chaco War between Bolivia and Paraguay, fought from 1932 to 1935 over disputed territory, killed roughly one hundred thousand people. Both belligerents were bound by the Pact. Neither was stopped by it. The regional response, however, was legally significant. The Anti-War Treaty of Non-Aggression and Conciliation of 1933, associated with the Argentine foreign minister Carlos Saavedra Lamas, condemned wars of aggression, provided for conciliation procedures, and declared that territorial questions must not be settled by force and that territorial acquisitions obtained by force would not be recognised. It reinforced the same principles the Pact had declared, with more procedural content, in a regional setting where the participating states had a shared interest in restraining the strongest actor among them. 9.4 Ethiopia, 1935 Italy invaded Ethiopia in October 1935. Both states were parties to the Pact and members of the League. The League this time did impose economic sanctions under Article 16 of the Covenant. The sanctions excluded oil, the one commodity whose denial might have crippled the Italian campaign, and the Suez Canal remained open to Italian shipping. Italy completed the conquest and proclaimed an empire in 1936. The Ethiopian case is the more damaging failure, because unlike in Manchuria, the machinery was actually used. Sanctions were imposed and they were insufficient, half hearted, and abandoned. The lesson learned by every revisionist government in Europe and Asia was not that aggression carried a price, but that the price was low and that the guardians of the system were unwilling to pay the cost of collecting it. 9.5 The road to 1939 By the late 1930s the Pact had been reduced to a rhetorical formality. Germany reoccupied the Rhineland in 1936, absorbed Austria in 1938, dismembered Czechoslovakia through the Munich Agreement and its aftermath, and invaded Poland in September 1939. The Soviet Union invaded Finland in November 1939 and was expelled from the League for it, a gesture whose practical effect was nil. Yet even at this lowest point, something curious was happening. Aggressor states did not repudiate the Pact. They did not announce that war was once again a lawful instrument of national policy. They lied instead. They manufactured provocations, invented defensive justifications, staged incidents, and produced elaborate legal arguments to show that they were victims rather than attackers. This is worth pausing on, because it is the first evidence of the Pact's real effect. The #hypocrisy of the aggressors is itself a form of testimony. The need to lie shows the existence of a standard. Before 1928 a state could annex a province and say frankly that it had conquered it. After 1928 that frankness disappeared from the diplomatic record, and it has never returned. 10. Nuremberg, Tokyo, and the Birth of the Crime of Aggression 10.1 The problem of legality When the Allied powers decided to try the leaders of Nazi Germany, they faced an obstacle that was not merely technical. They wished to charge those leaders not only with atrocities against civilians and violations of the laws of war, but with the act of starting the war itself. The Charter of the International Military Tribunal, agreed in London in 1945, accordingly listed crimes against peace, defined as the planning, preparation, initiation, or waging of a war of aggression or a war in violation of international treaties, and participation in a common plan to accomplish any of these. The defence raised an obvious objection. There was, it argued, no rule of international law in force before 1939 that made the starting of a war a crime for which an individual could be punished. To convict on such a charge was to apply a law created after the fact, in violation of the principle that there can be no crime without a pre-existing law. 10.2 The answer given The tribunal's answer relied heavily on the Pact of Paris. The reasoning proceeded in steps. Dozens of states, including Germany, had solemnly renounced war as an instrument of national policy. A treaty that renounces war necessarily makes resort to war unlawful. Those who plan and wage such a war are therefore engaged in an unlawful act. Since international law already recognised individual responsibility for other breaches, there was nothing entirely novel in attaching individual responsibility to the gravest breach of all. The principle against retroactive punishment, the tribunal added, is a principle of justice, and it would be unjust to allow those who had defied treaties and attacked their neighbours without warning to escape on the ground that no one had previously written down the penalty. Whatever one thinks of that reasoning, and it has been debated ever since, one point is undeniable. The Pact was doing real legal work. Without it, the argument that #aggressive_war was already unlawful in 1939 would have been far weaker. The treaty that had failed to prevent the war became the legal foundation for punishing those who started it. 10.3 Tokyo and the critics The International Military Tribunal for the Far East reached similar conclusions in relation to Japan and encountered similar objections. The dissent of Justice Radhabinod Pal of India remains the most sustained critique, arguing among other things that the law of aggression was being applied selectively, that the victorious powers had themselves used force in ways they now condemned, and that a criminal charge invented by the victors and applied only to the vanquished was an act of power dressed as an act of justice. That critique should not be brushed aside. It is the origin of the persistent charge of #victors_justice, and it explains much of the resistance among post-colonial states to accountability mechanisms that appear designed for use only against the weak. Any student who wishes to understand contemporary arguments about selective enforcement of the law on aggression should begin with Pal. 10.4 What the trials established Despite these criticisms, the trials established three propositions that survive. First, that the resort to aggressive war is unlawful, and that this unlawfulness predated 1945. Second, that individuals, and not merely abstract states, may bear #criminal_responsibility for that unlawful resort. Third, that the leadership of a state does not enjoy immunity from that responsibility by reason of official position. These propositions were affirmed as principles of international law recognised in the Charter and judgment of the Nuremberg tribunal, and they form the doctrinal ancestry of the crime of aggression as it exists today. 11. From Paris to San Francisco: Repairing the Design 11.1 The Charter rule The drafters of the Charter of the United Nations in 1945 knew what had gone wrong in 1928 and set out to fix it. Article 2(4) provides that all members shall refrain in their international relations from the threat or use of force against the territorial integrity or political independence of any state, or in any other manner inconsistent with the purposes of the United Nations. Compare this with Article I of the Pact and the repairs become visible. The prohibited conduct is force, not war. This closes the semantic escape route by which states fought while insisting they were not at war. The threat of force is prohibited alongside its use, which extends the rule to coercive diplomacy backed by mobilised armies. The rule protects territorial integrity and political independence expressly, which links the prohibition of force directly to the prohibition of #annexation. The phrase in their international relations preserves the limitation to interstate conduct, and here the continuity with the Pact is explicit. Commentators observe that the Charter formula continues the tradition established in Article I of the Pact, which confined the prohibition to the solution of international controversies between the parties (Pobjie, 2024). 11.2 The exceptions, defined at last The Charter did what the Pact had refused to do. It wrote the exceptions into the text. Article 51 preserves the inherent right of individual or collective self defence if an armed attack occurs, but subjects it to conditions. Measures taken must be reported immediately to the Security Council, and the right endures only until the Council has taken the measures necessary to maintain international peace and security. Self defence is thus recognised, but bounded, conditioned, and made subject to collective review. The purely self judging character that destroyed the Pact was, at least on paper, removed. Chapter VII gives the Security Council authority to determine the existence of a threat to the peace, a breach of the peace, or an act of aggression, and to decide on measures ranging from economic sanctions to the use of armed force. The determination of breach, which the Pact left to each state acting alone, is entrusted to an organ. 11.3 The unfinished repair None of this made the system reliable. The veto power of the five permanent members means that when a permanent member or a close ally is the accused, the Council is paralysed by design rather than by accident. Scholars of the law against war have documented in detail how the prohibition has been stretched by claims of humanitarian intervention, pre-emption, protection of nationals abroad, and consent given by contested governments (Corten, 2021; Buchan and Tsagourias, 2021). The debate over the exact boundary of a prohibited use of force, particularly in cyberspace, remains open (Pobjie, 2024). Yet the difference in kind between 1928 and 1945 is real. The Charter has a rule, a set of defined exceptions, an organ competent to determine breach, an enforcement chapter, a court with a role in adjudicating disputes about force, and a permanent institutional life. The Pact had a rule and a hope. 12. The Living Legacy: The Pact in Contemporary International Law 12.1 The prohibition of force as fundamental law The prohibition of the use of force is today regarded as one of the most fundamental rules of the international legal order, binding not only as treaty law under the Charter but as customary law, and widely treated as a peremptory norm from which no derogation is permitted (Corten, 2021). Its lineage runs through Paris in 1928, London and Nuremberg in 1945, San Francisco in 1945, and onwards through the General Assembly instruments of 1970 and 1974 that elaborated the principle and defined aggression. 12.2 The prohibition of annexation The doctrine that force cannot generate title, first given operational form in the non-recognition response to Manchuria, has become one of the load bearing walls of the modern order. Recent scholarship argues that the prohibition of annexations is foundational to contemporary international law and ties together three of its central projects: settling title to territory, regulating force across settled borders, and allowing the people within those borders to determine their own future (Brunk and Hakimi, 2024). The same scholarship warns that the rule is now at risk. The annexation of Crimea in 2014 and the purported annexation of four Ukrainian regions in 2022 are open violations of it. Legal analysis of the 2022 invasion concluded that it was a prima facie breach of the prohibition on the use of force, and that the justifications advanced, including collective self defence on behalf of purported republics and intervention to prevent an alleged genocide, did not withstand examination (Green, Henderson and Ruys, 2022). Recent monographs have turned to the practical question of how the rule of law can be strengthened against attempts to acquire #territory_by_force (Mikanagi, 2025). For students the connection to 1928 is direct. Every argument that annexation cannot produce lawful title rests on the premise that the force used to achieve it was unlawful. That premise entered general multilateral treaty law with the Pact of Paris. 12.3 The crime of aggression revived The crime that Nuremberg constructed partly out of the Pact lay largely dormant during the Cold War. The Rome Statute of the International Criminal Court included aggression among the crimes within the Court's jurisdiction but left the definition and the conditions of jurisdiction to be settled later. These were agreed at Kampala in 2010, and jurisdiction was activated in 2018 under restrictive conditions that leave substantial gaps, particularly where the aggressor state is not a party to the Statute. Those gaps became concrete after 2022. The response has been the creation of the Special Tribunal for the Crime of Aggression against Ukraine, established by an agreement between Ukraine and the Council of Europe signed in June 2025. If it becomes operational it will be the first international court to prosecute the crime of aggression since Nuremberg and Tokyo, and it raises difficult questions concerning jurisdiction, the immunities of senior state officials, and proceedings conducted in the absence of the accused (Nuridzhanian, 2025). The line from Paris to this tribunal is not metaphorical. It is doctrinal. The charge that a leader has committed the crime of #aggression is intelligible only because war ceased to be a lawful instrument of national policy, and the first general treaty to say so was signed in 1928. 12.4 The unresolved question of scope Modern controversies about what counts as force at all, including cyber operations, economic strangulation, support to armed groups, and the supply of weapons to a state under attack, are in a sense the descendants of the Pact's refusal to define war (Green, 2023; Pobjie, 2024). Every generation of lawyers inherits the boundary problem that the drafters in Paris chose not to solve. 13. Theoretical Perspectives 13.1 Realism The realist reading is straightforward and powerful. States pursue their interests, constrained by the distribution of power. Treaties that lack enforcement do not alter that distribution and therefore do not alter behaviour. The Pact provided moral satisfaction to publics and political cover to governments while leaving every material factor untouched. Worse, it may have encouraged the illusion that security could be purchased cheaply, weakening the case for rearmament and for firm alliances during the years when firmness might still have worked. The realist argument is at its strongest when it confines itself to the 1930s. It is at its weakest when asked to explain why, since 1945, the outright conquest and absorption of an entire recognised state has become a rare and scandalous event rather than an ordinary feature of international politics. 13.2 Liberal institutionalism and legalism The legalist reading holds that rules matter even when enforcement is weak. Rules create expectations, allocate the burden of justification, give victims a language in which to complain, and give third parties a reason to respond. On this account the Pact reversed the presumption in favour of belligerence, and everything that followed, including non-recognition, criminal responsibility for aggression, and Article 2(4), was constructed on that reversal (Hathaway and Shapiro, 2019). 13.3 Constructivism Constructivist scholarship focuses on how a norm becomes taken for granted. The key evidence is not compliance but justification. A norm exists, on this view, when violators feel compelled to deny that they have violated it. Judged by that standard the Pact succeeded to a remarkable degree. Since 1928 no major state has publicly claimed a general right to make war as an instrument of national policy. They claim self defence, they claim invitation, they claim humanitarian necessity, they claim pre-emption. The claims are frequently false, but the fact that they must be made at all is evidence. The constructivist reading also explains the mechanism of #non_recognition. It works not by taking territory back but by denying the aggressor the one thing it cannot seize by force, which is the acknowledgment of others that the gain is legitimate. 13.4 Critical and Third World approaches Critical scholarship insists on asking whom the rule protected. The Pact prohibited war between its parties, and its parties were recognised states in a world where large parts of humanity lived under colonial administration and were not recognised as states at all. The British reservation set aside unnamed regions for forcible protection. Colonial policing continued. Wars of national liberation lay outside a prohibition designed for relations between sovereigns. Critical scholars therefore argue that the outlawry of war was in part the legalisation of the existing distribution of territorial spoils. It froze the map at a moment convenient to those who had already conquered. The related empirical literature adds that the decline of conquest after 1945 is real but bounded. The seizure of small contested areas continues, and the norm constrains the total absorption of states more effectively than it constrains opportunistic land grabs (Altman, 2020; Altman and Lee, 2022). Recent work on the origins of the prohibition of annexation offers a partial answer to the critique from within the discipline. It shows that formerly colonised and non-aligned states, from nineteenth century Latin America through the era of decolonisation, were active agents in constructing the very norm often presented as a gift from the great powers (Brunk and Hakimi, 2024). The rule is not simply an imperial artefact. It has several parents, and some of them were the objects of empire rather than its authors. 13.5 Reconciling the perspectives These perspectives are less contradictory than they appear once the question is stated with care. If the question is whether the Pact prevented the Second World War, realism wins outright. If the question is whether the Pact changed the legal categories through which war is now argued about, judged, and punished, the legalist and constructivist accounts win. If the question is who benefited from the change and who was left outside it, the critical account is indispensable. A careful student holds all three, and asks which question is being answered before choosing a verdict. 14. Assessment and Design Lessons 14.1 Judging by the right standard Treaties should be judged against what they could plausibly have achieved. The Pact was signed by governments that would not accept enforcement obligations, in an international system with no reliable machinery for collective action, on the eve of the deepest economic depression of the modern era. Under those conditions the choice was not between a strong pact and a weak one. It was between a weak pact and nothing at all. Judged as a #peace_enforcement_mechanism, the Pact was a failure of the first order. Judged as an act of #norm_creation, it was among the most consequential legal instruments of the twentieth century. 14.2 Six design lessons The Pact yields a compact set of lessons about how to build international rules intended to bind the powerful. Each is stated in general form, because each remains live. First, a rule must define its own subject matter. A prohibition on war that does not define war invites evasion by relabelling. Second, exceptions must be written into the instrument. An exception whose scope is determined unilaterally by the state invoking it is not an exception. It is a licence. Third, some body must be competent to determine that a breach has occurred. Without an authoritative determination, every violation dissolves into a contest of assertions. Fourth, a rule without a remedy will be tested. If a violator suffers no cost, and observers see that no cost is imposed, the demonstration effect is immediate. Manchuria taught Rome. Ethiopia taught Berlin. Fifth, sanctions must be capable of hurting. Excluding oil from the sanctions imposed on Italy converted a demonstration of resolve into a demonstration of its absence. Sixth, breadth of membership and depth of obligation trade off against one another, and the trade-off should be made consciously. The Pact purchased near universal membership by asking for nothing. That was a real choice with real consequences, and it deserves to be analysed as a choice rather than dismissed as an oversight. 14.3 The pessimistic reading A darker reading of the Pact's legacy deserves a fair hearing. It is that the prohibition of aggressive war has not abolished war but changed its grammar. Conflicts are now fought through proxies, through irregular forces, through operations calibrated to remain below the threshold of an armed attack, and through the seizure of small pieces of territory in ways designed not to trigger a decisive response. Force has adapted to the rule rather than being suppressed by it. There is evidence for this. The empirical record shows that outright annexation of whole states is rare while contested seizures of smaller areas persist (Altman and Lee, 2022). Serious analysts of the present system argue that the prohibition on force is being pushed toward the point of collapse by states that have learned to exploit the ambiguity between force and coercion. If this reading is correct, then the achievement of 1928 is not that violence ended but that violence must now be disguised. That is a smaller achievement than the outlawry movement hoped for. It is not nothing. The requirement of disguise imposes costs, limits scale, and preserves a standard against which conduct can eventually be judged, as the current effort to prosecute the crime of aggression demonstrates. 15. Limitations of This Study Several limitations should be stated openly. The study is doctrinal and interpretive. It does not generate original quantitative evidence about the effect of the Pact on the incidence of war, and it relies on the published findings of quantitative researchers where such evidence is discussed. Those findings are themselves contested, and readers should treat causal claims about the decline of conquest with appropriate caution. The study draws principally on sources in English and reflects the state of the anglophone literature. Archival work in French, German, Japanese, Russian, and Spanish would enrich, and might complicate, the account offered here, particularly regarding the motives of the non-American parties. The treatment of the interwar crises is necessarily compressed. Manchuria, the Chaco, and Ethiopia each carry a specialised literature that this article can only gesture toward. Finally, the article writes about a legal order that is presently in motion. Assessments of the erosion of the prohibition on force, and of the prospects of the tribunal now being constructed for the crime of aggression, are provisional by nature and may be overtaken by events. 16. Conclusion The Kellogg-Briand Pact of 1928 promised more than it could deliver, and delivered something other than it promised. It promised the end of war as an instrument of national policy. It did not deliver that. It was signed by states that went on to invade Manchuria, Ethiopia, Poland, and Finland, and it contained no machinery capable of stopping any of them. The reservations attached to it, above all the self judging right of self defence, left an opening wide enough for every aggressor of the 1930s to walk through while claiming to be a victim. Judged by the standard its own authors set, the Pact failed, and the cost of that failure was counted in tens of millions of lives. What it delivered instead was a #legal_revolution that required two decades and a second world war to complete. It reversed the ancient presumption that war was a right of sovereigns. It made the resort to armed force a matter requiring justification rather than a matter of sovereign discretion. It supplied the legal premise for the doctrine that conquest cannot create title, for the charge of crimes against peace at Nuremberg and Tokyo, and for the prohibition of the threat or use of force in Article 2(4) of the Charter of the United Nations. Every one of those rules is alive and contested today, and every one of them can trace its ancestry to a treaty of three articles signed in Paris in the summer of 1928. The lesson for students is therefore neither cynicism nor celebration. It is a more demanding proposition: that a legal norm can be at once ineffective and transformative, and that the interval between the two conditions may be measured in generations rather than in years. The Pact of Paris was not a peace treaty that failed. It was a constitutional draft left unfinished, and the task of finishing it, of defining the rule, bounding its exceptions, giving it institutions, and enforcing it against the strong as well as the weak, is one that the present generation has inherited and has not yet completed. Hashtags #Kellogg_Briand_Pact_1928 #Pact_of_Paris #General_Treaty_for_Renunciation_of_War #outlawry_of_war_movement #jus_contra_bellum #prohibition_of_the_use_of_force #crime_of_aggression #interwar_diplomacy #League_of_Nations_Covenant #Stimson_Doctrine #non_recognition_of_conquest #Article_2_4_UN_Charter #Nuremberg_crimes_against_peace #international_legal_history #collective_security References Altman, D. (2020). The evolution of territorial conquest after 1945 and the limits of the territorial integrity norm. International Organization, 74(3), 490 to 522. https://doi.org/10.1017/S0020818320000119 Altman, D., and Lee, M. M. (2022). Why territorial disputes escalate: The causes of conquest attempts since 1945. International Studies Quarterly, 66(4), sqac076. https://doi.org/10.1093/isq/sqac076 Brunk, I., and Hakimi, M. (2024). The prohibition of annexations and the foundations of modern international law. American Journal of International Law, 118(3), 417 to 467. https://doi.org/10.1017/ajil.2024.26 Buchan, R., and Tsagourias, N. (2021). Regulating the use of force in international law: Stability and change. Edward Elgar Publishing. Corten, O. (2021). The law against war: The prohibition on the use of force in contemporary international law (2nd ed.). Hart Publishing. https://doi.org/10.5040/9781509949021 Green, J. A. (2023). The provision of weapons and logistical support to Ukraine and the jus ad bellum. Journal on the Use of Force and International Law, 10(1), 3 to 16. https://doi.org/10.1080/20531702.2023.2200321 Green, J. A., Henderson, C., and Ruys, T. (2022). Russia's attack on Ukraine and the jus ad bellum. Journal on the Use of Force and International Law, 9(1), 4 to 30. https://doi.org/10.1080/20531702.2022.2056803 Hathaway, O. A., and Shapiro, S. J. (2019). International law and its transformation through the outlawry of war. International Affairs, 95(1), 45 to 62. https://doi.org/10.1093/ia/iiy240 Mikanagi, T. (2025). Toward stronger rule of law against attempts to acquire territory by force. Springer. https://doi.org/10.1007/978-981-95-2814-1 Nuridzhanian, G. (2025). The Special Tribunal for the Crime of Aggression against Ukraine: Jurisdiction, immunities and due process. Nordic Journal of International Law. Advance online publication. https://doi.org/10.1163/15718107-bja10115 O'Mahoney, J. (2018). Denying the spoils of war: The politics of invasion and non-recognition. Edinburgh University Press. Pobjie, E. (2024). Prohibited force: The meaning of use of force in international law. Cambridge University Press. https://doi.org/10.1017/9781009022897