Arms control is the regulation of military capabilities and activities through agreed restrictions, transparency measures, and cooperative security arrangements. Its subjects include the development, manufacture, possession, deployment, transfer, testing, and use of weapons. Agreements may limit numbers, prohibit particular weapons, restrict military operations in specified areas, or establish procedures intended to prevent accidents and miscalculation. Although commonly associated with nuclear weapons, arms control also encompasses conventional armaments, chemical and biological weapons, missile technology, and military activities in emerging technological domains.
Arms control overlaps with disarmament but is not identical to it. Disarmament principally concerns reducing or eliminating weapons and armed forces; arms control can also stabilize relationships in which substantial military capabilities remain. Its importance lies in the possibility of reducing danger without first resolving the political disputes that produce military competition. The field combines strategic analysis, diplomacy, international law, scientific measurement, and the practical management of reciprocal obligations.
Purposes, concepts, and strategic logic
Arms control addresses the problem that measures taken by one state to increase its security may make another feel less secure. This interaction, often called the security dilemma, can produce competitive acquisition even when neither side initially intends to attack. Negotiated restraints attempt to interrupt that process by limiting threatening capabilities, clarifying intentions, or making military behavior more predictable. Their effectiveness depends not only on formal obligations but also on geography, force structure, political relations, and the ability to detect significant violations.
A central distinction concerns arms-race stability and crisis stability. Arms-race stability exists when states have relatively weak incentives to expand or transform their forces in response to each other. Crisis stability concerns incentives during an acute confrontation: a stable arrangement should reduce the perceived advantage of striking first. These objectives can diverge. A numerical ceiling may slow procurement while leaving vulnerable forces that leaders fear losing in an attack. Conversely, survivable forces may reduce pressure for immediate use even when their maintenance is expensive.
In nuclear strategy, deterrence and arms control have often developed together rather than as simple alternatives. During the Cold War, analysts argued that a secure retaliatory capability could discourage a first strike. Arms-control measures could support this condition by limiting counterforce capabilities, constraining defenses thought likely to stimulate offensive expansion, or reducing uncertainty about deployments. Other approaches emphasized delegitimizing nuclear weapons and moving toward their elimination. Different agreements therefore embody different assumptions about the relationship between security and retained military power.
Arms control can be quantitative, qualitative, geographical, or behavioral. Quantitative limits cap weapons, launchers, or personnel. Qualitative restrictions prohibit a category or characteristic, such as certain chemical agents or land-based missiles within specified ranges. Geographical measures establish demilitarized areas or nuclear-weapon-free zones. Behavioral rules govern exercises, notifications, encounters, or transfers. A single arrangement may combine several forms, and a limit on deployed systems may coexist with fewer restrictions on stored equipment or production capacity.
Successful design requires attention to substitution and asymmetry. States can compensate for restrictions in one category by expanding another, changing doctrine, improving accuracy, or dispersing forces. Equal numerical limits also need not produce equal military effects: a continental state and a maritime state may depend on different systems. For this reason, negotiations often involve counting rules, conversion procedures, geographic sublimits, and exceptions rather than a simple agreement to possess identical arsenals. The central analytical question is how an arrangement changes incentives and capabilities, not merely how many weapons it removes.
Early agreements and the development of modern arms control
Attempts to regulate armaments long predate nuclear weapons. The 1675 Strasbourg Agreement between France and the Holy Roman Empire prohibited the use of poisoned bullets between the parties. The 1817 Rush–Bagot arrangement between the United States and Britain sharply limited naval armaments on the Great Lakes and Lake Champlain. These measures were geographically and technologically narrow, but they illustrated two enduring approaches: prohibiting a particular means of warfare and reducing forces in an area where reciprocal restraint could improve security.
The Hague Peace Conferences of 1899 and 1907 connected armament questions with the broader development of the laws of war. Their achievements included restrictions on certain weapons and methods of warfare, although efforts to secure general limits on military expenditure did not succeed. The 1899 declaration concerning projectiles whose sole object was the diffusion of asphyxiating or deleterious gases was narrower than a comprehensive chemical-weapons prohibition. Such instruments demonstrated both the possibility of agreement on specific practices and the difficulty of controlling military competition as a whole.
After the First World War, the League of Nations made disarmament a central institutional objective. Article 8 of its Covenant linked peace to reducing national armaments to the lowest level consistent with national safety and enforcement of international obligations. The World Disarmament Conference opened in Geneva in 1932, but disputes over security guarantees, equality of armaments, and Germany’s military position prevented a general settlement. Germany withdrew from the conference and the League in 1933.
Naval agreements achieved more concrete, though temporary, results. The Washington Naval Treaty of 1922 limited capital-ship tonnage according to a ratio of 5:5:3:1.75:1.75 for the United States, Britain, Japan, France, and Italy. It also imposed restrictions on individual ships and provided for scrapping vessels. The London Naval Treaty of 1930 extended regulation to additional categories, including cruisers, destroyers, and submarines. These arrangements constrained some forms of competition but left room for technological substitution, disputes over relative status, and changes in strategic circumstances.
The experience of the interwar period exposed a recurring limitation: restrictions cannot be separated from the security order supporting them. Naval agreements depended on continued acceptance by major naval powers, while general disarmament negotiations became entangled with territorial grievances and collective-security failures. After the Second World War, the destructive potential of atomic weapons transformed the scale of the problem. The 1946 Baruch Plan proposed international control of atomic energy, but disagreement over inspection, ownership, sanctions, and the sequence of disarmament prevented its adoption.
Bilateral nuclear arms control during and after the Cold War
The United States and the Soviet Union developed the most extensive bilateral arms-control system because their arsenals and strategic rivalry dominated the nuclear balance. The 1962 Cuban missile crisis sharpened concern about escalation, communication failures, and the short time available for decisions. In 1963, the two governments agreed to establish a direct communications link. That year, the Partial Test Ban Treaty prohibited nuclear explosions in the atmosphere, outer space, and underwater, while leaving underground testing outside its prohibition.
The Strategic Arms Limitation Talks produced two major agreements in 1972. The Anti-Ballistic Missile Treaty limited strategic missile defenses, initially allowing two deployment areas per side; a 1974 protocol reduced that allowance to one. The accompanying Interim Agreement constrained certain strategic offensive launchers. Its limits did not directly cap all warheads, however, and the deployment of multiple independently targetable reentry vehicles allowed the number of warheads to increase without a corresponding increase in launchers. This illustrated the importance of distinguishing delivery systems from the weapons they carried.
SALT II, signed in 1979, established broader limits on strategic delivery vehicles but never entered into force. The 1987 Intermediate-Range Nuclear Forces Treaty took a different approach: it eliminated an entire class of United States and Soviet ground-launched ballistic and cruise missiles with ranges between 500 and 5,500 kilometres. Its prohibition covered both nuclear- and conventionally armed missiles in those categories. By the 1991 elimination deadline, the parties had destroyed 2,692 missiles, supported by an intrusive verification system that included on-site inspections.
The Strategic Arms Reduction Treaty, or START I, was signed in 1991 and entered into force in 1994. It limited accountable strategic delivery vehicles to 1,600 and attributed warheads to 6,000 for each side, with additional sublimits. Following the dissolution of the Soviet Union, Belarus, Kazakhstan, Russia, and Ukraine became involved through the 1992 Lisbon Protocol. The removal of Soviet nuclear weapons from the three non-Russian successor states connected bilateral reductions with wider questions of succession, non-proliferation, and security assurances.
New START, signed in 2010 and effective from 2011, set limits of 1,550 accountable deployed strategic warheads, 700 deployed strategic launchers and heavy bombers, and 800 deployed and non-deployed launchers and heavy bombers combined. Its accounting rules counted each deployed heavy bomber as one warhead, rather than counting every weapon it could carry. A 2021 extension set its expiration date at 5 February 2026. The wider bilateral framework had already weakened through the United States’ withdrawal from the ABM Treaty in 2002 and the termination of the INF Treaty in 2019 following a compliance dispute. Russia announced the suspension of its participation in New START in 2023.
Nuclear non-proliferation, testing, and prohibition
The Treaty on the Non-Proliferation of Nuclear Weapons, opened for signature in 1968 and effective from 1970, established the principal global framework for limiting the spread of nuclear weapons. It distinguishes five nuclear-weapon states—China, France, the Soviet Union, the United Kingdom, and the United States—by reference to the manufacture and explosion of a nuclear weapon or other nuclear explosive device before 1 January 1967. Russia subsequently continued the Soviet Union’s treaty status. This legal category does not encompass every state that possesses nuclear weapons.
The treaty rests on interconnected obligations concerning non-proliferation, peaceful nuclear cooperation, and disarmament. Nuclear-weapon states undertake not to transfer nuclear weapons or control over them, while non-nuclear-weapon states undertake not to receive, manufacture, or otherwise acquire them. Article IV addresses peaceful nuclear activities in conformity with the treaty, and Article VI requires good-faith negotiations on measures relating to cessation of the nuclear arms race and nuclear disarmament. In 1995, the parties agreed to extend the treaty indefinitely.
The International Atomic Energy Agency administers safeguards designed to verify that nuclear material subject to safeguards is not diverted to nuclear weapons or other nuclear explosive devices. Comprehensive safeguards agreements cover nuclear material in peaceful activities in non-nuclear-weapon states party to the NPT. The Model Additional Protocol, approved in 1997, provides additional information and access intended to improve the agency’s ability to detect undeclared material and activities. Safeguards do not amount to a general inspection authority over every military installation; their scope depends on the applicable legal instruments.
The Comprehensive Nuclear-Test-Ban Treaty, adopted in 1996, prohibits all nuclear-weapon test explosions and other nuclear explosions. Its entry into force requires ratification by 44 specified states listed in Annex 2. Its preparatory organization has nevertheless developed an International Monitoring System using seismic, hydroacoustic, infrasound, and radionuclide technologies. These methods provide complementary evidence: seismic signals can indicate an underground explosion, while radioactive particles or gases may help establish its nuclear character. National test moratoria remain legally distinct from an operative universal treaty prohibition.
The Treaty on the Prohibition of Nuclear Weapons, adopted in 2017 and effective from 2021, prohibits a broad range of nuclear-weapons activities for its parties. It also contains obligations concerning victim assistance and environmental remediation. Nuclear-armed states did not participate in its adoption and have not joined it. Nuclear-weapon-free-zone treaties, including those covering Latin America and the Caribbean, the South Pacific, Southeast Asia, Africa, and Central Asia, provide a different regional approach. Their protocols seek commitments from nuclear-weapon states, although signature, ratification, and associated reservations vary among states and zones.
Chemical and biological weapons
Chemical and biological arms control developed around weapons whose effects depend on toxic chemicals or disease-causing organisms and toxins. The 1925 Geneva Protocol prohibited the use in war of asphyxiating, poisonous, or other gases and analogous materials, as well as bacteriological methods of warfare. It did not comprehensively prohibit development, production, or stockpiling. Some states also attached reservations concerning retaliation or relations with non-parties, limiting the instrument’s original reach as a universal rule against use.
The Biological Weapons Convention, opened for signature in 1972 and effective from 1975, prohibits biological agents and toxins of types and quantities that have no justification for prophylactic, protective, or other peaceful purposes. It also prohibits weapons, equipment, and means of delivery designed to use them for hostile purposes or in armed conflict. This “general-purpose criterion” avoids dependence on a fixed list of organisms. A newly discovered pathogen is not outside the prohibition merely because it was unknown when the convention was negotiated.
Biological verification presents distinctive difficulties. Facilities producing vaccines, conducting microbiological research, or manufacturing pharmaceutical products may use equipment and expertise relevant to prohibited activities. Small quantities of a biological agent may be significant, while legitimate research can involve dangerous pathogens. The convention has no standing international inspection organization comparable to the chemical-weapons regime. Confidence-building information exchanges began in the 1980s, but negotiations on a legally binding verification protocol failed to produce agreement in 2001. These institutional limitations do not alter the underlying prohibition.
The Chemical Weapons Convention, opened for signature in 1993 and effective from 1997, prohibits the development, production, acquisition, stockpiling, retention, transfer, and use of chemical weapons. It requires destruction of declared chemical-weapons stockpiles and associated production facilities under international verification. The Organisation for the Prohibition of Chemical Weapons, based in The Hague, implements the convention through declarations, inspections, technical assistance, and institutional decision-making. Its schedules organize verification of specified chemicals and precursors without exhausting the convention’s broader prohibition on toxic chemicals used as weapons.
In July 2023, the OPCW announced that all chemical-weapons stockpiles declared by states parties had been verified as destroyed. This milestone concerned declared stockpiles, not proof that no undeclared chemical weapons existed. Investigations of alleged use, the accuracy and completeness of declarations, abandoned weapons, and old chemical weapons remain separate issues. The convention also distinguishes prohibited warfare uses from purposes not prohibited, including certain protective, medical, industrial, and law-enforcement activities. Riot-control agents may be used for law enforcement, but their use as a method of warfare is prohibited.
Conventional weapons, humanitarian restrictions, and the arms trade
Conventional arms control addresses weapons that are not classified as nuclear, chemical, or biological, but its purposes range from maintaining regional military balances to reducing particular civilian harms. Conventional weapons vary enormously in cost, portability, destructive effect, and military role. Consequently, agreements governing tanks and combat aircraft operate differently from those addressing landmines, ammunition, or small arms. Some arrangements restrict possession and deployment, while others regulate use or international transfer without requiring general reductions.
The Treaty on Conventional Armed Forces in Europe, signed in 1990, limited five categories of equipment: battle tanks, armoured combat vehicles, artillery, combat aircraft, and attack helicopters. Its original structure established ceilings for two groups of states in an area extending from the Atlantic to the Urals. Equipment destruction, information exchanges, and inspections were intended to reduce the capacity for large-scale surprise attack and offensive operations. A 1999 adaptation sought to replace bloc-based limits with national and territorial ceilings, but it did not enter into force. Political and military disputes subsequently undermined implementation of the original regime.
Humanitarian weapons agreements focus on effects that persist beyond combat or create particular risks to civilians. The 1980 Convention on Certain Conventional Weapons provides a framework for protocols addressing specific weapons, including mines and booby traps, incendiary weapons, blinding laser weapons, and explosive remnants of war. Its structure permits states to accept different combinations of protocols. These rules intersect with international humanitarian law, especially the principles of distinction, proportionality, and precautions in attack, but establish more specific obligations for particular weapon types.
The Anti-Personnel Mine Ban Convention, adopted in 1997, prohibits the use, stockpiling, production, and transfer of anti-personnel mines by its parties. It also requires stockpile destruction, clearance of mined areas, and assistance to victims. The Convention on Cluster Munitions, adopted in 2008, similarly combines prohibitions with clearance, stockpile destruction, and victim-assistance obligations. Neither treaty has universal membership. Their operation illustrates how arms control can involve humanitarian organizations and affected communities alongside military establishments, including the International Committee of the Red Cross and civil-society campaigns.
The Arms Trade Treaty, adopted in 2013 and effective from 2014, regulates international transfers of specified conventional arms. It establishes prohibitions and requires assessments before authorizing certain exports, including consideration of risks involving serious violations of humanitarian or human-rights law. It does not prohibit all international arms commerce or generally regulate domestic civilian ownership. Complementary measures include the 2001 UN Programme of Action on small arms and light weapons, export licensing, marking and tracing, stockpile security, and end-use controls. These instruments address diversion as well as formally authorized transfers, although their coverage and legal status differ.
Verification, monitoring, and the management of uncertainty
Verification is the process of assessing whether parties are complying with agreed obligations. It does not require complete knowledge of every military activity. Instead, a verification system seeks sufficient confidence that violations of military significance would be detected in time for an effective response. The appropriate standard depends on the weapon, the prohibition, the consequences of cheating, and the time needed to convert a concealed activity into usable military capability. A small accounting discrepancy and a covert production complex therefore pose different problems.
National technical means include reconnaissance satellites, radar, electronic monitoring, and other capabilities operated by individual states. They can observe fixed installations, missile tests, major deployments, or destruction activities, but cannot reliably reveal every object inside a building or determine every state’s intention. Some agreements explicitly protect the use of national technical means against interference and prohibit deliberate concealment that would impede verification. Cooperative measures can supplement these capabilities with declared data, notifications, exhibitions, and agreed access.
On-site inspections transform verification by allowing designated personnel to examine facilities, equipment, records, or destruction procedures. Their authority is defined by the relevant agreement: inspectors do not possess unrestricted access merely because an arms-control treaty exists. Procedures may specify inspection quotas, advance notice, permitted instruments, escorts, and protected information. “Managed access” allows an inspected state to conceal unrelated sensitive information while still demonstrating compliance. The challenge is to protect legitimate security and commercial interests without making meaningful inspection impossible.
Nuclear safeguards often combine material accountancy with containment and surveillance. Operators report inventories, receipts, transfers, and processing, while inspectors assess records and make independent measurements. Measurement uncertainty means that an apparent material imbalance is not automatically proof of diversion. Investigators must examine whether discrepancies arise from instrument error, processing losses, incomplete records, or deliberate removal. Remote cameras, seals, sampling, and complementary access can provide evidence that accounting records alone cannot supply. Preserving continuity of knowledge becomes especially important when monitoring equipment fails or access is interrupted.
Verification of warhead dismantlement poses additional challenges because confirming a genuine nuclear weapon may reveal classified design information. Research has examined information barriers that provide a limited pass-or-fail result without disclosing sensitive measurements. Inspectors may also need confidence that a verified item remains under control through transport and dismantlement, linking verification to chain-of-custody procedures. Open-source imagery, commercial satellite data, and environmental analysis increasingly supplement official monitoring, but their evidentiary value depends on calibration, provenance, and alternative explanations. Monitoring produces observations; a compliance judgment additionally requires interpretation against a precise legal obligation.
Legal forms, institutions, and responses to non-compliance
Arms-control commitments take several legal forms. A binding treaty may establish durable obligations, verification rights, and dispute procedures. Political commitments can create expectations of restraint without the same legal status. Reciprocal unilateral declarations may coordinate reductions without a negotiated treaty text, as occurred with the United States and Soviet presidential nuclear initiatives of 1991. The form selected affects ratification, amendment, withdrawal, domestic implementation, and the ability to adapt commitments when military circumstances change.
The Vienna Convention on the Law of Treaties provides a general framework for treaty interpretation and operation, alongside relevant customary rules. Arms-control treaties commonly contain provisions on entry into force, duration, amendment, reservations, and withdrawal. Some permit withdrawal after notice when extraordinary events related to the treaty’s subject matter jeopardize a party’s supreme interests. Exercising such a clause does not erase questions about compliance before withdrawal, and withdrawal should be distinguished from suspension, termination by agreement, or a treaty’s scheduled expiration.
Domestic implementation determines whether international commitments constrain actual conduct. Legislation may establish export-control systems, criminalize prohibited activities, authorize inspections, regulate dangerous materials, and designate a national authority. The Chemical Weapons Convention, for example, requires states parties to adopt measures implementing its prohibitions within their jurisdictions. Constitutional arrangements also shape the approval process: the role of a legislature, executive authority, or judicial review varies by state. An international promise may therefore depend on extensive administrative and legal action inside each participating country.
Institutional arrangements range from specialized organizations to small bilateral commissions. Review conferences evaluate implementation and may adopt recommendations, while consultative bodies clarify ambiguous provisions and investigate disputes. Within the United Nations, the General Assembly provides a broad forum for disarmament resolutions and negotiations. The Conference on Disarmament in Geneva is a separate multilateral negotiating body that operates by consensus. This procedure protects participants from being bound by an unwanted negotiating outcome but can also prevent agreement on a programme of work.
There is no universal automatic enforcement mechanism for arms-control violations. Responses may include requests for clarification, additional inspections, findings by treaty bodies, diplomatic pressure, suspension of cooperation, or sanctions. The Security Council can act under its Charter powers, but political disagreement among its permanent members can restrict collective action. Resolution 1540, adopted in 2004, imposed obligations on states to prevent non-state actors from obtaining nuclear, chemical, or biological weapons and their means of delivery. Such measures address security risks beyond traditional reciprocal limits between military rivals.
Regional arrangements, confidence-building, and crisis prevention
Regional arms control adapts general principles to particular geographic and political conditions. A restriction that is useful across an ocean may be ineffective between neighboring states whose aircraft or missiles can reach targets within minutes. Borders, disputed territories, alliance commitments, and conventional force imbalances shape what governments consider acceptable. Regional arrangements therefore often combine limits on equipment with notifications, observation, separation of forces, or procedures for handling unexpected military encounters.
In Europe, confidence- and security-building measures developed through the Helsinki process and subsequent agreements. The 1975 Helsinki Final Act included advance notification of certain major military maneuvers. The 1986 Stockholm Document strengthened notification and observation provisions and introduced inspection rights. The Vienna Document, whose major updated version was adopted in 2011, provides for information exchanges, notifications of specified military activities, observation, and mechanisms for consultation about unusual military activity. Such measures do not eliminate military forces; they seek to reduce uncertainty about their organization and movement.
The Open Skies Treaty, signed in 1992 and effective from 2002, permitted unarmed observation flights over participating states under agreed rules. Its purpose was cooperative transparency rather than direct force reduction. The United States withdrew in 2020, and Russia withdrew in 2021, reducing the treaty’s geographical coverage and strategic significance. The treaty illustrates both the value and vulnerability of access-based arrangements: shared observation can create a common evidentiary basis, but its operation depends on continued political acceptance and compliance with flight procedures.
Crisis-prevention agreements address the possibility that an incident will escalate before political leaders can clarify what happened. The 1972 United States–Soviet Agreement on the Prevention of Incidents On and Over the High Seas established rules for safer interactions between naval vessels and aircraft. India and Pakistan concluded an agreement on reducing the risk of accidents relating to nuclear weapons in 2007 and an agreement on pre-notification of ballistic missile flight tests in 2005. They also exchange annual lists of nuclear installations under an agreement signed in 1988 and effective from 1991.
These measures differ from security guarantees and peace settlements. A hotline provides a channel of communication but does not ensure that leaders will use it, believe the information conveyed, or agree on a response. Exercise notification may reduce surprise while still leaving disputes over an activity’s purpose. Nuclear-weapon-free zones similarly depend on regional participation and external commitments, rather than geography alone. Their practical contribution is to establish procedures and expectations that can remain useful even when political relations are poor, provided the parties continue to honor them.
Technological change and enduring design problems
Technological change repeatedly challenges the categories used in arms-control agreements. A treaty may distinguish ballistic from cruise missiles, deployed from stored systems, or offensive from defensive capabilities, while new designs blur those boundaries. Hypersonic glide vehicles, long-range conventional precision weapons, and maneuverable delivery systems raise questions about warning time, mission ambiguity, and coverage under existing definitions. The relevant issue is not speed alone: trajectories, basing arrangements, payloads, and command systems influence how another state interprets a launch or deployment.
Missile defense illustrates the relationship between offense and defense. Interceptors may protect against limited attacks, yet an adversary may respond by adding warheads, decoys, maneuvering vehicles, or additional launchers. The resulting interaction depends on scale, effectiveness, geography, and the forces being defended. Agreements that regulate only offensive launchers may therefore leave important drivers of competition untouched. Conversely, including every related technology in a single negotiation can make agreement much more difficult, especially when states assess the same system differently.
Outer space and cyberspace create further verification and attribution problems. The 1967 Outer Space Treaty prohibits placing nuclear weapons or other weapons of mass destruction in orbit or stationing them in outer space, but it does not prohibit all military space activity. Dual-use satellites can support civilian services and military operations simultaneously. In cyberspace, software can be copied, modified, and concealed more readily than a large physical weapon. Agreements focused on behavior, protected targets, or communication procedures may therefore be more practicable in some circumstances than inventories of prohibited tools.
Artificial intelligence also affects military decision-making, targeting, intelligence analysis, and autonomous systems. Discussions within the Convention on Certain Conventional Weapons have examined lethal autonomous weapons systems, but states differ over definitions and regulatory approaches. Questions include the predictability of system behavior, the allocation of responsibility, and the character of human control over the use of force. These issues intersect with existing humanitarian-law obligations, while extending beyond the numerical limits characteristic of earlier strategic treaties.
Several broader design problems persist across technologies. Bilateral limits may become less comprehensive as additional states acquire relevant capabilities, yet multilateral negotiations must reconcile more varied security interests. Transparency can reassure an adversary but also expose vulnerabilities. Restrictions can reduce deployed weapons while leaving production capacity, stored components, or rapid rearmament options intact. Irreversibility therefore has degrees: destroying a launcher, dismantling a warhead, converting fissile material, and eliminating specialized infrastructure each remove different capabilities. Assessing an agreement requires attention to its specific coverage, verification, durability, and consequences for military incentives, rather than treating formal participation or weapon counts as sufficient measures of effectiveness.