Nuclear Deterrence and Disarmament: What Each Position Is Protecting

On September 26, 1983, a Soviet early-warning satellite reported five incoming American intercontinental ballistic missiles. The officer on duty at the Serpukhov-15 bunker outside Moscow was a lieutenant colonel named Stanislav Petrov. His job, per protocol, was to report an attack confirmed by the automated system up the chain of command. He did not. He judged — correctly, as it turned out — that the satellite system had malfunctioned, that the United States would not launch only five missiles in a first strike, and that the report was a false alarm. He filed it as such. He was reprimanded for failing to follow protocol. The world did not end.

Seven weeks later, NATO began a military exercise called Able Archer 83, simulating a conventional war escalating to nuclear release. The exercise was realistic enough — it included encrypted communications blackouts, practiced nuclear release procedures, and involved senior political figures — that Soviet intelligence concluded it might be cover for an actual first strike. Soviet nuclear forces in Eastern Europe were quietly placed on higher alert. A KGB officer named Oleg Gordievsky, who was secretly feeding information to British intelligence, reported what he observed. The NATO exercise was quietly modified. Tensions subsided. The world did not end.

Deterrence advocates cite both episodes as evidence that deterrence works: the mutual threat of annihilation kept both sides from actually pushing the button even under extreme pressure. Abolitionists cite both episodes as evidence that deterrence is a bet we cannot sustain indefinitely: on at least two occasions within seven weeks, the difference between nuclear exchange and survival came down to the judgment of a single person who had reason not to trust their instruments and chose, against protocol, to trust their gut.

Both readings are accurate. The debate about nuclear weapons is, in part, a debate about what those episodes prove — and what evidence could possibly settle it.

What the classical deterrence position is protecting

The core claim of classical deterrence — developed by theorists including Bernard Brodie, Thomas Schelling, and Robert Jervis from the late 1940s onward — is that nuclear weapons have transformed the logic of international conflict. Because any nuclear exchange would destroy both sides, rational states will not initiate one. Mutual vulnerability to annihilation is not a flaw in the international order; it is what has made the post-1945 period among the least war-prone in recorded history between great powers. The goal is not to eliminate nuclear weapons but to maintain the condition of mutual deterrence — specifically by ensuring that both sides retain a survivable second-strike capability, meaning they can absorb a first strike and still retaliate with unacceptable force.

What this position is protecting is the long peace itself. Its advocates point to a historical anomaly: the great powers of Europe fought catastrophic wars in 1914 and 1939. Since 1945, despite profound ideological conflict and repeated crises — Berlin, Korea, Cuba, Vietnam, dozens of proxy wars — the United States and Soviet Union never exchanged a shot directly. The deterrence position credits nuclear weapons for this absence. As Brodie put it in 1946, in the first serious strategic analysis of atomic weapons, the chief purpose of a military establishment had changed from winning wars to preventing them.

The structural argument goes deeper than the historical record. Jervis's The Meaning of the Nuclear Revolution (Cornell University Press, 1989) argues that mutual vulnerability is not a policy choice but a permanent structural feature of the nuclear era — one that cannot be escaped by building better defenses or acquiring more weapons. Any defense good enough to prevent retaliation would destabilize deterrence by creating a first-strike temptation. The revolution nuclear weapons introduced is that offense is now so cheap and destruction so certain that no defense can restore the pre-nuclear logic of victory through superior force. States that internalize this fact are deterred; states that do not are dangerous.

Schelling's contribution, developed across The Strategy of Conflict (1960) and Arms and Influence (1966), was to analyze the credibility problem: for deterrence to work, the threatened retaliation must be genuinely believable, which means the adversary must believe a state would actually follow through even at catastrophic cost to itself. His answer — deliberately relinquishing some control over escalation, allowing the situation to carry its own momentum, making the threat leave something to chance — is deeply uncomfortable. It means deterrence rests not on a clean rational calculation but on a managed version of the irrational. The deterrence position accepts this discomfort as the price of a peace that has actually held.

What this position costs: it cannot prove the counterfactual. We cannot know whether the great powers would have fought again without nuclear weapons. The peace may have had other causes — economic interdependence, war exhaustion, international institutions, democratic governance. And the position is most compelling when applied to stable great-power dyads with developed command infrastructure. It has much less to say about nuclear weapons in fragile states, or about what happens when deterrence is extended to allies who might drag a patron into a conflict it would not otherwise fight.

What the nuclear warfighting position is protecting

A distinct strand of strategic thinking — associated with Herman Kahn, Colin Gray, and much of the actual operational planning done by the U.S. Strategic Command through the Cold War — argues that pure mutual assured destruction is an insufficient and dangerously passive doctrine. Deterrence requires more than the capacity to destroy adversary cities after absorbing a first strike. It requires the capacity to fight, limit, and potentially prevail across a spectrum of nuclear scenarios, from tactical battlefield use to full strategic exchange. An adversary who doubts your willingness to follow through on catastrophic retaliation — because they correctly calculate that using your weapons would destroy you too — is not adequately deterred.

Kahn's On Thermonuclear War (Princeton University Press, 1960) and the subsequent On Escalation (1965) mapped what he called a 44-rung escalation ladder, arguing that nuclear war outcomes exist on a spectrum from bad to catastrophically worse and that planning and capability must address every rung. A state that can only threaten existential retaliation cannot credibly threaten anything less — which means it can be coerced below the nuclear threshold by an adversary who calculates that the patron would never actually risk mutual annihilation over a peripheral interest. This is the extended deterrence problem in its sharpest form: why would the Soviet Union believe that the United States would risk New York to defend Berlin?

The warfighting position is protecting the credibility of alliance commitments specifically and coercive leverage generally. Albert Wohlstetter's 1959 Foreign Affairs essay "The Delicate Balance of Terror" — the most cited article in the history of strategic studies — argued that stable deterrence was not automatic but required constant attention to survivability, accuracy, and the ability to use weapons in controlled and limited ways. The logic of counterforce targeting — aiming at adversary weapons rather than cities — follows from this: if you can destroy enemy missiles before they are launched, you can limit damage to your own side and give an adversary the option to stop the exchange short of apocalypse.

The practical manifestation of this logic runs through every U.S. nuclear posture review from the Cold War through the present: the preference for flexible response options, low-yield tactical weapons, improved accuracy, and — most controversially — the periodic interest in missile defense systems designed to intercept a ragged retaliatory strike after a disarming first strike.

What this position costs: counterforce capability is itself destabilizing in precisely the way it claims to prevent. If your missiles can destroy an adversary's missiles on the ground, the adversary faces a use-it-or-lose-it pressure in a crisis — launch before your forces are destroyed, or lose the ability to retaliate at all. Glenn Snyder called this the stability-instability paradox: nuclear stalemate at the strategic level may enable aggression below the threshold, because each side knows the other won't escalate to existential exchange over a limited provocation. The warfighting position's answer to this problem — develop the capability to fight limited nuclear wars — may create the very instability it attempts to solve.

What the minimum deterrence position is protecting

A third tradition — associated with the later careers of strategists including Robert McNamara, and revived by former Cold War hawks George Shultz, William Perry, Henry Kissinger, and Sam Nunn in their landmark 2007 Wall Street Journal op-ed "A World Free of Nuclear Weapons" — argues that deterrence requires far fewer weapons than any nuclear state currently maintains. States can be deterred by the prospect of even a handful of surviving warheads destroying their major cities. Arsenals of thousands of warheads beyond that floor are expensive, politically corrosive, and militarily redundant.

McNamara, as Defense Secretary under Kennedy and Johnson, developed the concept of "assured destruction" — the capacity to destroy a defined fraction of adversary population and industrial capacity after absorbing a first strike — and argued that a surprisingly small force could achieve it. His later writings, particularly In Retrospect (1995) and the documentary The Fog of War (2003), recanted his own role in nuclear buildup and argued that the United States and Soviet Union had come terrifyingly close to nuclear war on multiple occasions, largely through luck and the good judgment of individuals who chose to deviate from protocol at critical moments.

The minimum deterrence position is protecting deterrence stability itself, but argues that the classical and warfighting traditions have mistaken complexity for credibility. A state that has, say, 300 survivable warheads capable of destroying any adversary's major cities has a credible second-strike deterrent. A state that has 5,000 warheads, various yield options, and a published doctrine for limited nuclear war is not more deterred — but it has spent enormous resources and created strategic postures that give adversaries reason to worry about disarming first strikes, creating crisis instability rather than resolving it.

The position also protects nonproliferation norms. The nuclear Non-Proliferation Treaty's implicit bargain — non-nuclear states forswear weapons in exchange for nuclear states pursuing disarmament under Article VI — is made transparently hollow by arsenals being maintained and modernized rather than reduced. China's long-standing minimum deterrence posture, maintained for decades at roughly 300-400 warheads while the U.S. and Russia held tens of thousands, illustrated that a great power can deter without maximalist arsenals.

What this position costs: it rests on a judgment about adversary rationality that may not hold in all cases. A minimum deterrent deters a rational adversary who believes in the credibility of retaliation. It may not deter a state that doubts the will to use those weapons, or one whose leadership believes it could absorb a retaliatory strike and survive in a politically meaningful sense. The warfighting critique — that pure assured destruction is not credible for alliance defense — has not been answered by the minimum deterrence position, only sidestepped.

What the arms control pragmatism position is protecting

The arms control tradition — founded intellectually by Thomas Schelling and Morton Halperin in Strategy and Arms Control (1961) and institutionalized through the bilateral treaty architecture from SALT I in 1972 to New START in 2010 — holds a position that is neither deterrence advocacy nor disarmament advocacy. Its premise is that the near-term elimination of nuclear weapons is not achievable, and that idealistic abolition campaigns distract from the tractable goal of managing nuclear danger through verifiable, mutual arms reductions that build the habits of cooperation that deeper reductions might eventually require.

Schelling and Halperin's insight was that arms control need not mean disarmament. Mutual measures that reduce the risk of war — by creating transparency, constraining destabilizing weapons, establishing communication channels, and making arsenals more survivable and thus less prone to use-it-or-lose-it pressures — serve the interests of both sides without requiring either to achieve unilateral advantage. The ABM Treaty of 1972 was the purest expression of this logic: by prohibiting missile defenses, both sides locked in mutual vulnerability, precisely because mutual vulnerability makes first strikes irrational.

What the arms control position is protecting is the institutional architecture that three generations of diplomats constructed to reduce nuclear danger. SALT I in 1972, the ABM Treaty in 1972 (the U.S. withdrew in 2002, to the persistent concern of arms controllers), the INF Treaty in 1987 (which collapsed when the U.S. withdrew in 2019 citing Russian violations), START I in 1991, and New START in 2010 each reduced arsenals and — more importantly — created verification regimes, inspection rights, and data-exchange obligations that made both sides' nuclear postures predictable and thus less frightening. The architecture also created diplomatic habits: regular contact between the technical and military communities of both sides, a shared vocabulary, and institutional knowledge of what the other side actually fears.

The Bulletin of the Atomic Scientists moved its Doomsday Clock to 85 seconds to midnight in January 2026, the closest it has ever stood — driven in part by the expiration of New START in February 2026 without a successor agreement. This is the first time since 1972 that the United States and Russia have operated without an arms control framework. The arms control community's concern is not merely symbolic: without data exchanges and inspection rights, both sides must plan against worst-case assumptions about the other's arsenal, recreating exactly the dynamic that arms control was designed to prevent.

What this position costs: the arms control framework has been progressively dismantled by both the United States and Russia over the past two decades, and the pragmatic tradition has not produced a compelling account of why it is collapsing or what would arrest the collapse. China has refused to join bilateral talks. Russia has suspended New START verification while making explicit nuclear threats in Ukraine. The political conditions that produced the original treaties — superpower awareness of shared catastrophic risk, arms race fatigue, détente as a diplomatic framework — do not currently exist. Pragmatism requires two parties who want pragmatic outcomes.

What the humanitarian abolitionism position is protecting

The most fundamental challenge to all deterrence positions does not argue about strategic logic at all. It argues that nuclear weapons are categorically impermissible — that their effects are so indiscriminate, their potential consequences so catastrophic, and their possession so morally corrupting that no military or political purpose can justify their existence. The debate should not be conducted in the language of strategy but in the language of humanitarian law and moral prohibition.

Jonathan Schell's The Fate of the Earth (Knopf, 1982) made the philosophical argument that nuclear weapons occupy a unique moral category because they threaten not just existing people but all possible future people — what Schell called the "second death," the extinction not just of the living but of the entire human lineage. This is not a marginal risk. It is what nuclear deterrence requires: the ongoing credible threat to commit an act that, if carried out, would end human civilization. Schell argued that maintaining such a threat was not a regrettable necessity but a form of ongoing collective moral madness, and that recognizing this required not better strategy but a different kind of politics altogether.

The TTAPS study — published in Science in December 1983, by Turco, Toon, Ackerman, Pollack, and Carl Sagan — added the scientific dimension. It modeled the global climatic effects of nuclear exchange and found that the soot, smoke, and particulate matter injected into the stratosphere by burning cities would block sunlight, collapse global temperatures, and destroy agricultural production worldwide — killing hundreds of millions to billions of people who had nothing to do with the conflict, most of them in the Global South. Subsequent modeling by Robock et al. in 2007 refined the picture: even a "regional" nuclear exchange between India and Pakistan, using roughly 100 Hiroshima-scale weapons, could cause a nuclear famine killing more than a billion people globally within a decade. Nuclear war between the great powers is not a bilateral catastrophe. It is a global one.

The abolitionist position draws on this evidence to argue what the International Campaign to Abolish Nuclear Weapons (ICAN) — awarded the Nobel Peace Prize in 2017 — has built into a legal strategy: the Treaty on the Prohibition of Nuclear Weapons, adopted by 122 UN member states in July 2017 and entered into force in January 2021. The TPNW is modeled on the precedent of landmines and cluster munitions — the theory being that legal prohibition stigmatizes a weapon, delegitimizes its possession, raises the political cost of retention, and over time creates the normative environment in which abolition becomes politically conceivable.

The philosophical backbone was provided by Randall Forsberg, the disarmament theorist who created the nuclear freeze movement, and later by natural law philosophers including Finnis, Boyle, and Grisez in Nuclear Deterrence, Morality and Realism (Oxford University Press, 1987), who argued through careful Catholic moral theology that deterrence cannot be justified as bluff, as counterforce, as lesser evil, or as a transitional measure — that the doctrine requires the genuine intention to commit mass murder, and that no end justifies that intention. The philosopher Jeff McMahan made the parallel analytic argument in "Is Nuclear Deterrence Paradoxical?" (Ethics, January 1989): the very thing that makes deterrence work is the genuine willingness to do something that would be catastrophically wrong.

What this position costs: the TPNW has been signed by no nuclear weapons state and no NATO ally. The legal prohibition strategy has created a powerful normative statement — and left the weapons intact. The abolitionist position has not resolved the transition problem: how a world with nuclear weapons moves to a world without them in a way that doesn't create worse instability at every intermediate step. A state that disarms while its adversaries retain weapons has not made the world safer — it has made itself vulnerable. The abolitionist movement has been better at naming the destination than at mapping the route.

Where the debate actually is

The nuclear debate has a distinctive structure: it is among the most technically sophisticated in international relations, and it reaches an impasse that no amount of sophistication resolves. All five positions agree that nuclear war must be prevented. They differ on whether nuclear weapons are the instrument of prevention or the threat to be prevented — and this is not a technical disagreement but a philosophical one about the nature of peace, the reliability of institutions, and the time horizon over which we are making judgments.

The credibility paradox sits at the center and has never been resolved. Deterrence requires that the threat of retaliation be genuinely credible — meaning the adversary must believe the state would actually follow through, even knowing that doing so would trigger mutual annihilation. If retaliation would be irrational, the threat to retaliate is not credible. If retaliation would be rational, we must explain what rationality means when both sides are destroyed. Schelling's answer — that credibility comes from deliberately leaving something to chance, binding yourself in ways that make backing down impossible — is intellectually elegant and politically terrifying. It means deterrence rests on controlled irrationality, which means its failure could come from exactly the same place as its success.

The luck-versus-design problem is what Sagan's The Limits of Safety (Princeton University Press, 1993) systematized from the historical record. Drawing on declassified archives, Sagan documented what he called a "disturbing history of near-catastrophes" — military accidents, misread warnings, confused communications, and command breakdowns that did not produce nuclear exchange not because the system worked as designed but because someone deviated from protocol at the right moment or chance intervened. Petrov in 1983. Vasili Arkhipov on a Soviet submarine during the Cuban Missile Crisis, who refused to authorize a nuclear torpedo launch when the other two required officers had already consented, while the submarine was under depth charge attack and communications with Moscow were severed. The deterrence community acknowledges these episodes and argues they demonstrate resilience. Sagan argues they demonstrate that we have been lucky, and that luck is not a strategy.

The Ukraine precedent has injected a new urgency into the disarmament side of the argument — and a new difficulty. Ukraine inherited Soviet nuclear weapons when the Soviet Union collapsed. Under the 1994 Budapest Memorandum, it surrendered them in exchange for security assurances from Russia, the United States, and the United Kingdom. Russia's full-scale invasion in 2022 demonstrated that security assurances are not deterrence. The clear lesson for the proliferation debate is that nuclear weapons provide a form of security guarantee that no treaty can replicate — and the equally clear lesson is that Russia's nuclear threats during the invasion shaped Western decision-making on weapons supply throughout the conflict, demonstrating deterrence's coercive utility in a war that was already underway. Both lessons are real. They point in opposite directions.

What is genuinely unresolvable — and important to name — is that the debate about nuclear weapons is also a debate about how much we trust institutions under extreme pressure. Deterrence advocates trust that rational actors will not initiate annihilation. Arms controllers trust that diplomatic frameworks can manage the relationship between adversaries who each maintain the capacity to destroy the other. Abolitionists trust that legal and normative prohibition, once achieved, would hold even in crisis conditions. None of these trusts has a fully adequate empirical foundation. Each represents a bet on human nature and institutional reliability under conditions that have, so far, been tested only at the edge.