Nuclear weapons first appeared in the world as the weapons of a democracy. In the eight decades since, they have continued to exist in the arsenals of several democratic states, many of which have credited them with serving to protect their freedoms and systems of government. During the Cold War, nuclear weapons were, as US president Ronald Reagan (Reference Reagan1982) put it, the balance against “totalitarian forces … who seek subversion and conflict around the globe to further their barbarous assault on the human spirit.” Nuclear weapons are then a good thing for democracies, preserving their governments in a world otherwise populated by barbarous destroyers of freedom. At the same time, they safeguard those governments from the pressures of having to fight incessant conventional wars for their survival, which might otherwise lead them down a path to militarist authoritarianism (Friedberg Reference Friedberg1992; Joffe Reference Joffe, Miller and Smith1994).
However, this story misses an important dimension of the relationship between nuclear weapons and democracy. Nuclear weapons do not just affect international relations, but also domestic politics. There, they present a particular object of governance that hinders rather than protects the functioning of democracy. The weapons are uniquely destructive, capable of completely eradicating a country in a way no other type of weapon can. They can achieve this eradication in mere minutes, with the involvement of very few people, and without states being able to defend themselves. These features lead states to have to govern them with particular care.
Their unique destructive features have two types of effect on the possibility of democratic control over them. First, nuclear weapons present inherent limits to the type of democratic power that is possible over their use. Second, they lead nuclear-armed states to develop specific strategies to manage the dramatic vulnerability and extreme power of the nuclear age, which in turn have consequences for democratic control. When states use nuclear threats to deter war, or when they prepare to fight nuclear wars, the specter of nuclear annihilation and the uncertainty of success in avoiding it enable and incentivize undemocratic practices.
This article distinguishes between these two types of effect and shows that although ideas on the requirements of nuclear strategy modify the level of compatibility between nuclear weapons policy and democracy, none allow for sufficient compatibility. Changes in strategic concept can inhibit, but not entirely stop, nuclear weapons from evading democratic control.
The incompatibility of nuclear weapons with democracy does not mean that nuclear-armed states cannot make democratic decisions in other policy fields, nor does it mean that the limitations on democracy are unique to nuclear weapons. It simply means that nuclear-armed democracies are “restricted” in the sense that democratic processes cannot apply themselves to a part of the state’s activities (Fraise Reference Fraise2023, 53–54). This is significant. Nuclear weapons decisions have the potential to be among the most important a democracy takes; they could end millions of lives and destroy the biosphere in minutes. An attachment to democracy for any reason implies an interest in whether these weapons can be democratically controlled.
This article therefore assesses limits to democratic control over nuclear weapons to establish whether it is possible to make nuclear decisions through any democratic process. The analysis relies on a procedural definition of democracy that allows for multiple types of popular control. Doing so serves to base the democratic critique of nuclear weapons on universal principles beyond specific governments’ arrangements. More specifically, this article contributes to the literature on nuclear weapons and democracy by distinguishing the ways in which nuclear weapons are fundamentally incompatible with democracy from the undemocratic choices states have made.
This article identifies three “modes” by which the people of a democracy can hold power, which I call “representative,” “pluralist,” and “participatory,” and briefly outlines the pressures on democracy that are introduced by nuclear weapons. Then, it evaluates the possibility of making democratic decisions under these pressures through prior agreement and post hoc accountability. Structural limits on representative, pluralist, and participatory agreement result from assumptions about the functioning of nuclear strategy. The same is true of far-reaching nuclear secrecy, which hinders accountability. The strategic assumptions driving limits to democracy are logical reactions to the destructivity of nuclear weapons and the uncertainty of nuclear policy making, but they are not inevitable. The article then discusses other limitations to democracy that do necessarily result from the nature of nuclear weapons. The speed of nuclear wars and the involvement of very few people in starting them negate the possibility of reversing launch decisions. Finally, the global reach of nuclear weapons means their governance within individual states cannot be democratic since it fails to include the entire demos of people who would be affected by their use, and should therefore get a say in nuclear decisions.
The Requirements of Democracy
Setting out a definition of democracy requires navigating a diverse set of ideas on what the concept means. But if democracy is when the people rule, as the common etymological definition maintains, it does not matter what ends they pursue, nor does it matter what democratic governance achieves. It matters that voters can control political decisions. As such, a democracy in principle only needs to serve its own preservation. Whether nuclear weapons are compatible with the moral underpinnings of modern democracy relating to human dignity, justice, or the right to live is therefore not in the scope of this discussion. Excellent work on the relationship between nuclear weapons and liberalism exists and is certainly relevant for modern democracies’ attitudes toward their arsenals.Footnote 1 That consideration remains distinct from the possibility of procedural popular control.
This article identifies three broad modes of democratic control: representative, pluralist, and participatory. Democracy requires that at least one of these modes allows citizens meaningful power over policy decisions. They must be able to vote, organize, or participate their way into mattering.
Per Josiah Ober’s (Reference Ober2017, 157) work on “basic democracy,” democracy is a form of “collective and limited self-governance by well-motivated and capable citizens.” Self-government may take the form of direct democracy, but delegation of power can be democratic so long as citizens can remove representatives from power, choose between multiple distinct candidates, and have enough information to judge these alternatives (128–56). Ober identifies four preconditions for this form of democracy: civic equality, political freedom, sufficient public knowledge, and basic civic dignity, which make it possible for citizens to wield power (155–60).
For government to be “collective,” it must treat any individual voter as equal to any other by virtue of their belonging to the electorate. Inherent preference for select groups constitutes benevolent oligarchy at best. Equality is then not a value in itself but merely represents the necessary political balance of power. Political freedom is similarly instrumental. For citizens to be able to express their preferences for governance, freedom of speech must be guaranteed. Freedoms that are not essential for open political debate need not be protected. The requirement for sufficient public knowledge, for its part, gives citizens the ability to usefully discuss their common government and necessitates transparency of government decisions.Footnote 2
Ober’s framework otherwise corresponds well with other conceptualizations of representative democracy. Robert Dahl (Reference Dahl1989, 106–15) sets out a similar list of requirements, including “effective participation,” “voting equality,” “enlightened understanding,” and “control of the agenda.” The first three correspond roughly to political freedom, civic equality, and information access. In these models for democracy, people have power because they choose representatives who hold power, and because these representatives remain responsive to the popular will. The representative mode of control thus requires both elections and accountability.
The use of representatives does not account for the full range of ways people can control their government. Philip Pettit (Reference Pettit2000, 114–23) identifies a “contestatory” dimension in addition to the familiar “electoral” one. It involves a variety of institutions, including procedural constraints on the content and process of decisions (e.g., the separation of powers), consultative resources (e.g., public inquiries), and appellate resources (e.g., ombudsmen or tribunals) (127–33). These offer alternative ways for citizens to influence policy during officials’ terms. Broadly speaking, they do so either through organizations or through direct inclusion. I will refer to these as the pluralist and the participatory modes of control.
The pluralist mode allows organizations outside the state to represent the interests of a group of citizens and to influence decisions with various modes of power. They shape decisions by competing with one another, either within government (e.g., through political parties) or through outside groups (e.g., labor unions). Organizations must somehow access power, be it through official consultation, lobbying, or protest. Dahl’s (Reference Dahl1982, 10–11) description of “pluralist democracy” requires the same conditions as Ober’s “basic democracy,” as well as freedom of association and the existence of autonomous organizations. Crucially, pluralism does not always represent a democratic form of power. Since advocacy relies on access to nonelectoral resources (e.g., financial, rhetorical, or reputational), interest groups can hinder public participation. Yet they play a crucial role in diversifying the expression of viewpoints and in creating democratic control between elections so long as they wield power that is somewhat proportional to the segment of the population they represent.
Finally, participatory modes of democracy involve citizens directly participating in decision making. Participatory modes of democracy are rooted in consensus or deliberation rather than competition. This is often conceptualized for communities smaller than modern states, since their scale complicates direct input (Menser Reference Menser2018, 11–65). It nevertheless remains possible to combine states with participation through institutions such as referenda, citizen’s assemblies, participatory budgeting, or direct participation in government work (Barber Reference Barber2003, 266–307). If all citizens work toward a common goal, they can have political power without or despite the state through unity of action.
The representative, pluralist, and participatory modes of democratic control all represent ways a state’s citizens can control decisions. They each require certain institutions to render a government responsive to its public, particularly the presence of basic political freedoms, political equality, and sufficient access to information.
Nuclear Weapons and Democracy
In this conceptualization of democracy as a primarily domestic construction regulating decision-making processes, there are no moral imperatives for international state action. Foreign and defense policies do not have to promote democratic principles or respect human rights in other states. For foreign policy to be democratically legitimate, it must fulfill two conditions: it must be controlled by the public through one of the three modes, and it must not undermine the conditions for citizens’ political freedom, equality, and knowledge. This latter condition carries with it the responsibility to ensure the continuation of democracy and thus the survival of the “people” as a collective. While the state is under no democratic obligation to guarantee individual lives, maintain international peace, or abhor the use of nuclear weapons, it is required to ensure that its citizens can continue to govern.
Conditions for popular control and effective policy making do not always overlap. For many policies, there exists a contradiction between their effectiveness and their dependence on consent from electorates prone to short-term thinking, as well as between the equivocating of democratic processes and the decisiveness needed for impactful policy (Diamond Reference Diamond1990, 50–53). Foreign policy in particular requires the state to act decisively, consistently, and at times secretly as a single unit. Centralized rule capable of mobilizing society’s resources and making quick, decisive, and secretive international moves is generally considered more effective than the slow compromises of transparent agreement making in democracies (Gourevitch Reference Gourevitch1978, 896–99).
On a structural level, the balance between democracy and effectiveness is therefore mediated by the international environment: the greater the amount and significance of the threats from outside, the greater the pressure on democratic states to compromise some democratic mechanisms for state survival (Deudney Reference Deudney2008, 56). The particularity of nuclear weapons is then the level of threat they represent, which creates unparalleled pressure on democracies to ensure effectiveness.
Since states cannot defend themselves against nuclear attacks, they may choose to rely on nuclear deterrence to assuage their citizens’ demands for security (Deudney Reference Deudney and Lipschultz1995, 99–103). This is not the only goal of nuclear weapons, which may also serve to facilitate states’ aggression, bolster their prestige, or fight wars. Deterrence is not sufficient to overcome their vulnerability to nuclear violence but it does provide a way to manage it in the short term. The centrality of deterrence in the ability of nuclear-armed states to claim they can provide security relies ultimately on a form of collective bluff. It thus draws democracies into relying on a form of deception.
Together, the pressure on effectiveness in governing nuclear arsenals and the nontransparency that results from practicing deterrence limit how states make nuclear weapons decisions.
This framework relies on some arguable assumptions about nuclear weapons. Primarily, it assumes that deterrence is effective in preventing at least some harms to the state, particularly in the form of nuclear first strikes. It also assumes this benefit outweighs the risks of having an arsenal, and so places the focus of nuclear policy on mitigating threats from outside attack rather than from risks of accident or targeting provided by one’s own nuclear stockpile. Without these assumptions, democracies would be better off not maintaining a nuclear arsenal at all. Any democracy with nuclear weapons thus also implicitly accepts them as true. As a result, this paper will argue within the framework of these assumptions, without prejudice to their accuracy. The next sections assess if limitations on democracy are necessary to allow states to attain security from nuclear strikes, starting with the opportunity for democratic agreement in advance.
Democratic Agreement to Nuclear Policy
Across the nuclear states, the public are largely uninvolved in making nuclear policy. This section discusses the prospects for pluralist and participatory agreement to decisions to use nuclear weapons in one subsection, and for representative agreement in the next. It concludes that these limits are not inherent in nuclear policy but depend on assumptions about what nuclear weapons are good for.
Pluralist and Participatory Agreement in the Nuclear Monarchy
The most direct limit placed on democratic control is the exclusion of the public from the decision to use a nuclear weapon. The next sections examine the necessity of this restriction, which cuts off pluralist and participatory opportunities for control.
The discussion starts from the observation that all nuclear-armed states concentrate power over nuclear use in executives—at the expense of pluralist and participatory modes of democratic control—and lays out the logic of this so-called nuclear monarchy. The monarchy serves to ensure nuclear strikes can be launched quickly, but speed is only an issue for retaliatory strikes. The next sections then introduce and evaluate the necessity of retaliatory strikes, of speed in delivering them, and of the nuclear monarchy in guaranteeing speed. This leads to the conclusion that the nuclear monarchy is not necessary to protect the state in all cases, but the uncertainty and destructivity of nuclear weapons do always create possibilities and incentives for its creation.
The Nuclear Monarchy and Modes of Democratic Control
Nuclear weapons decisions are not made through broad democratic engagement. Born, Gill, and Hänggi’s Reference Born, Gill and Hänggi2010 study found that decision-making power lies with the executive in all nuclear-armed states for which information is accessible. Of course, democratic consent for the maintenance of nuclear arsenals is constantly affirmed through the approval of defense budgets. While US nuclear decisions in the twentieth century were made primarily by the president, they were constrained by popular and congressional opinion (Mlyn Reference Mlyn1995, 155–56). Congress had a say over the nature of the arsenal, which influenced the scope of its possible uses.
However, approval for the maintenance of an arsenal does not constitute consent for any instance of use. Only the executive has the power to make that decision, leaving decision-making power in the hands of an exceedingly small clique of executive officials and military leaders. Launch decisions are generally made by only one person (the president or prime minister of the nuclear state), with checks by a handful of other, largely military, officials at most. France, for instance, requires the execution of a launch order to be verified by two military officials, while Israel needs two people to authorize a launch (Cohen Reference Cohen2019; Pelopidas Reference Pelopidas2019). In neither case is there representation from other government branches, let alone interest groups or citizen collectives. Elaine Scarry (Reference Scarry2014, 37–84) calls this situation the “thermonuclear monarchy” and argues that it cannot be justified on the basis of the American Constitution. Many others agree (DiPippa Reference DiPippa2019; Meyrowitz Reference Meyrowitz1982; Raven-Hansen Reference Raven-Hansen1989; Stone Reference Stone and Raven-Hansen1987).
More fundamentally, the concentration of power in the hands of one individual is not compatible with two of the three modes of democracy. It removes the possibility of pluralist and participatory control from the moment of decision. This leaves only the representative mode, which is, of course, minimally present if the monarch is an elected official.Footnote 3
The question is whether this limitation of pluralist and participatory control is necessary to save the public in a nuclear war or prevent one from happening at all. The argument for “nuclear despotism,” by which nuclear weapons inherently produce the monarchy, claims it exists to accommodate the need for speed in launching retaliatory nuclear strikes, given their exceptional destructiveness (Deudney Reference Deudney2008, 255). A decision on using one’s weapons when an enemy has launched theirs (a “second” or “retaliatory” strike decision) has to be taken before the means to conduct a strike or the structures to authorize it are destroyed. Since the invention of high-speed ballistic missiles that cannot reliably be destroyed in flight, there could be just minutes between detection and impact. The window of opportunity for retaliation is short; decisions must be made at a moment’s notice. This can only happen if control is centralized and time-consuming debate between representatives removed. The executive, reduced to one person, is left with the decision. The logic is the same as for exceptional emergency powers to allow states to respond to crises, with the nuance that the emergency powers are constantly necessary because nuclear war can break out at any time (Falk Reference Falk, Cohen and Lee1986, 444–47). The nuclear monarchy is therefore permanent.
To establish whether nuclear weapons are compatible with democracy, it should be established whether the speed of response that the nuclear monarchy makes possible is always necessary to prevent the country’s destruction. A rapid response is relevant only to retaliatory strikes, launched in reaction to an enemy’s nuclear attack. In this scenario, the time between detection of the incoming missiles and their impact is short, and it is in these minutes that monarchs must decide whether to launch. First strikes are not nearly as time sensitive, because there is no incoming missile. The United States estimated in 1985 that it would have two or three days to decide on a nuclear first strike in the event of a Soviet attack on Europe (Raven-Hansen Reference Raven-Hansen1989, 790). Timelines for use may be shorter in other conditions (e.g., when facing imminent defeat in a conventional war, or contemplating a preemptive strike), but there would still be significantly more time than the mere minutes of emergency retaliation. It would be possible to assemble a wider set of representatives for the launch decision. This has regularly been proposed by critics of current government arrangements (Blair Reference Blair2018, 6–13). While fully transparent public debate might not be possible for fear of triggering preemptive strikes on the arsenal, limited subgroups of discreet legislators or members of the judiciary could be included. Even broader participation is possible if the first strike under consideration is aimed at a state not defended by a nuclear arsenal. Without fear of nuclear devastation, the only consequence would be the tactical impracticality of alerting the adversary to one’s intention to resort to nuclear strikes.
If speed is only necessary for second strikes, the nuclear monarchy serves to make them possible. The inevitability of the nuclear monarchy depends on whether retaliatory strikes are necessary to protect the public. If they are not, the nuclear monarchy could be dismantled.
The Necessity of a Retaliatory Strike
For the nuclear monarchy to be a necessary consequence of nuclear weapons, retaliatory strikes should either limit destruction in the event of a nuclear war or prevent one from happening in the first place.
Retaliatory strikes cannot defend against incoming missiles. To be useful in an actual war, they must prevent further harm from subsequent strikes. This can only be done in states whose decision makers can survive the first strike. A small country that is totally destroyed, having lost many of its cities, people, and its social cohesion, has little practical use for a retaliatory strike: it effectively has no citizens, democracy, or interests to protect. This is not an obscure hypothetical involving the decimation of a microstate: it was estimated in 1954 that the United Kingdom could be effectively destroyed by just 10 thermonuclear warheads with yields of 10 megatons (Hughes Reference Hughes2003, 263). Around 15 similar weapons would destroy France (Miclot Reference Miclot2011, 9). Small, densely populated Israel would not fare better. After a large-scale first strike, the nations these countries’ arsenals would seek to protect would be devastated. A second-strike capability is then not necessary from a warfighting perspective because it cannot save the state nor even meaningfully reduce the damage done to it.
By contrast, larger countries with more dispersed populations stand a chance of surviving a first strike. They could feasibly have something to save and a future to fight for. A second strike aimed at the adversary’s arsenal might then be useful to reduce damage from subsequent rounds of strikes and so (perhaps) save not only the state, but millions of its people if it manages to cripple the enemy’s arsenal. Whether this is strategically necessary depends on a state’s assessment of the impact of the first strike on its territory and of the possibility of meaningful damage limitation—which may be limited if the enemy has a large, survivable arsenal.
From a wartime perspective, retaliatory strikes are then only necessary in large states that expect to be able to reduce damage from subsequent enemy strikes. While these conditions for necessity are easy to deduce logically, they are difficult to gauge objectively because they involve predictions for an uncertain future: how large and impactful will an enemy’s first strike be, and how realistic is damage limitation? All strategy is uncertain, and nuclear deterrence is doubly so since no nuclear wars have been fought to provide historical evidence that might reduce that uncertainty. If a state’s answers to these questions lead it to expect that a retaliatory strike might save its people, it has an enormous incentive to adopt it as its nuclear strategy. Even far-fetched scenarios constitute a significant incentive to construct a nuclear monarchy, given the dramatic scale of nuclear destruction.
Moreover, even if second strikes serve no purpose in fighting a nuclear war, they are generally considered important in deterring one. The effectiveness of deterrence does not depend on the rational usefulness of a second strike as a warfighting tool, since irrational factors like revenge may lead states to strike even from the grave (McDermott, Lopez, and Hatemi Reference McDermott, Lopez and Hatemi2017). Preparing a second strike is therefore potentially useful for deterrence purposes even for states that cannot survive first strikes. The possibility of retaliation ensures that an enemy would suffer the devastation of a nuclear attack even if it launched a successful first strike. Without “mutually assured destruction,” it might be tempting for states to destroy their adversaries, and the threat their nuclear arsenals present, in a first strike. A nuclear monarchy ensures the threatened state can strike back.
It is fundamentally unclear if a nuclear monarchy is necessary for effective deterrence, because the power of the threat does not depend on the actual ability to realize it so much as the enemy’s perception of it (Jervis Reference Jervis1982). Establishing requirements for success would only be possible by somehow peering into the minds of the other side’s decision-makers. Even learning from experience is difficult, since the success of nuclear deterrence would have to be judged in large part by the nonoccurrence of deterred events or by the ambiguous role the existence of nuclear weapons has played in specific crises (Bell and Miller Reference Bell, Miller, Narang and Sagan2022, 213–14). For this reason, a broad range of possible deterrence postures based on different interpretations of the historical record have been formulated (Sauer Reference Sauer2009). The uncertainty of success and the magnitude of the risk involved incentivize maximizing certainty, including by ensuring second strikes are realistically achievable in all scenarios. That may not be necessary. Schelling (Reference Schelling1980, 187–94) argues that uncertain threats can still deter effectively, and Waltz (Reference Waltz1981, 5–10) stresses that a nuclear attack might be so devastating that it makes leaders extremely risk averse. So long as the enemy believes you may be able to strike back, it can be deterred. Nuclear weapons could then be effective deterrents by simply existing. Accepting this concept of “existential deterrence” would allow for effective deterrence without a guaranteed second strike and voids the necessity of preparing for one.
In all, actually launching a second strike is only necessary for damage limitation purposes in large states or if a state’s concept of deterrence requires the threat of a second strike to be executable in practice. But because states cannot be certain about the level of necessity and because nuclear weapons are intensively destructive, they have an incentive to conclude retaliatory strikes are necessary.
The Necessity of Speed
If second strikes are considered necessary, the nuclear monarchy is only required if speed of response is critical. If retaliation can only be done in the minutes between the detection of an enemy attack and its impact because the state’s arsenal or delivery methods will be destroyed, the process for launching missiles must be quick and so involve only a few people.
States’ nuclear arsenals are designed to survive a first strike. All nuclear-armed democracies are largely considered to have survivable arsenals. Their submarines and mobile land-based launchers with sufficiently long-range strike capabilities should survive nuclear attack because they cannot reliably be targeted. They could launch a retaliatory strike days, weeks, or even months after the original attack.
But missiles need to be launched and, in a democracy, their launch must be authorized by its citizens, or at least their representatives. The authorization system must then also survive the first strike. This may be difficult, since a single strike on the capital or on leaders’ known locations could feasibly wipe out the entire apparatus of democracy. “Decapitation” is a likely strategy for nuclear use precisely because it is so simple yet impactful (Steinbruner Reference Steinbruner1981). While states have invested in infrastructure to ensure government survival, the history of such efforts shows their effectiveness is far from guaranteed (Krugler Reference Krugler2007). The time between detection and impact is short, and leaders may be far away from shelters, distracted, or may disbelieve reports of an impending attack. It is not improbable they would die. While “backup” decision makers exist everywhere in case the nuclear monarch is killed, they may not survive either. Since politicians are generally concentrated in a nation’s capital and major cities, these potential deciders are likely to be affected by the same strikes. If it is possible all nuclear decision makers would die before they could launch a retaliation, speed remains a necessity.
This only applies to nuclear retaliation designed for damage limitation. Deterrence may not necessarily require speed even if its functioning does rely on launching a real second strike. A democracy could choose delayed deterrence, whereby a retaliatory strike would follow days or weeks after the first, providing an opportunity to reconstitute its leadership. India, for instance, has selected this strategy (Tellis Reference Tellis2022, 78–83). In principle, this could allow for broader inclusion in launch decisions since concentration of power is less necessary when timelines are longer. Delayed retaliation is, of course, less certain than immediate retaliation, reliant as it is on the reconstitution of a government after the nuclear strike. Even if the state survives a nuclear attack, it is likely to be vulnerable to invasion and occupation—particularly if its adversary is spared a nuclear strike. A rapid response remains the more attractive strategy to pursue, given the uncertainty of nuclear conditions.
The Necessity of the Nuclear Monarchy
Third and finally, if retaliatory strikes are sometimes necessary and sometimes require speed, is the nuclear monarchy the only way to deliver them? Or can other solutions allow democracies to launch their arsenals after an enemy strike? Paradoxically, the desire to launch strikes through a civilian, democratically elected official creates the democratic limitations of the nuclear monarchy.
There is a backstop to decapitation. To ensure a response remains possible in case the civilian leadership is lost or unreachable, authority to launch is delegated to military personnel (Blair Reference Blair1993, 46–52). British nuclear submarines carry “letters of last resort” with orders from the prime minister, to be opened after a nuclear attack. While France’s nuclear arsenal is theoretically impossible to launch without a presidential order (Tertrais Reference Tertrais, Born, Gill and Hänggi2010, 114), there appear to be provisions to devolve power to an unelected inconnu de province (“anonymous provincial”) in case of war (Quilès, Drain, and Collin Reference Quilès, Drain and Collin2018, 63–67). Predelegation guarantees that a nuclear response is possible even after devastating strikes, but it has an obvious democratic deficit: it places the authority over nuclear use with unelected people, often in the military.
Delegation to lower functionaries is standard practice in modern democracies and is not in itself undemocratic. It must, however, remain reflective of the popular will. Democratic delegation, like representation, depends on its responsiveness to the public’s will (Downey Reference Downey2021, 307–11). Viewed as an extension of representative control, democratic delegation requires the ability to hold officials accountable.
Nuclear predelegation does not meet these standards. The identity of France’s inconnu de province is kept secret to ensure they are not targeted. This also ensures they cannot be held accountable for their role in nuclear decisions. British submarine captains are protected from destruction by their survivable submarines and so do not present the same problem. Their names may not be exactly publicized, but their selection criteria and place in government are clear, as are possible punishments for insubordination or treason. It is therefore the secrecy of their orders that makes the letters of last resort undemocratic. This results from a different constraint on democracy arising from deterrence practices, which will be discussed in the next section. So long as effective deterrence requires secrecy to surround its practices, predelegation in this form is insufficiently democratic.
To ensure legitimacy as well as civilian control over the nuclear arsenal, these appointees are therefore supposed to receive orders from democratically established authorities—orders that might arrive only moments before incoming bombs fall. Launch decisions are therefore only urgent if one believes, as a democrat should, that strikes must be authorized by the elected civilian government before its possible destruction. The choice is between nuclear monarchy and disordered nuclear oligarchy—either rapid decisions by a small subset of the government, or later decisions by unelected officials. This urgency excludes the possibility of any broad consultation, particularly because various members of government may be in different locations, occupied, or unreachable. There appears, then, to be a paradox: because the decision has to be made democratically, it has to be made quickly; and because it has to be made quickly, it has to be made undemocratically through the nuclear monarchy.
Overall, the nuclear monarchy is necessary if one’s strategy requires the capacity to actually launch a second strike quickly and with a high degree of certainty. Damage limitation might require a second strike in large or survivable states. In all states, effective nuclear deterrence is widely viewed as conditional on communicating a realistic ability to retaliate, which is easier to do when that ability actually exists. If either of these goals leads states to prepare second strikes, then the survival of the government may require speed, and the democratic need to avoid devolving decisions to local military leaders would then lead to the concentration of power in a nuclear monarch.
The rapid total destruction entailed by the use of nuclear weapons makes possible and incentivizes a degree of power concentration that violates the democratic need to include the public when preparing second strikes. Other uses, particularly first strikes, might involve a greater number of decision makers. The nuclear monarchy is not a strategic necessity in all cases, but results from the assumption that greater certainty in the ability to retaliate increases deterrence credibility or the odds of survival for large states.
Agreement on Nuclear Strategy: Deterrence Credibility and Limits on Information Access
A second set of constraints emerges from states’ attempts to ensure deterrence credibility. These impose a lack of transparency about governments’ and leaders’ intentions. Without transparency, no mode of control is possible. In particular, it takes away the public’s ability to select representatives and so negates the representative mode of control.
Nuclear deterrence between two nuclear-armed states suffers from a “credibility gap”: it is rarely in a state’s rational interest to actually use nuclear weapons when the adversary can retaliate. Since no effective defense exists, the likely outcome of any nuclear first use is the destruction of one’s own state by another state’s weapons. Second strikes, beyond their very specific application in damage limitation as described above, are irrational since they cannot protect states from nuclear devastation. Even a first strike that does not incur a retaliatory nuclear response can have climatic effects that inflict unacceptable damage on the attacker (Robock and Toon Reference Robock and Toon2012). Nonetheless, states may see significant security benefits in making any of these types of nuclear threats in terms of preventing aggression or interference in their national interests. Making a believable nuclear threat is then difficult, but useful.
As a result, leaders of nuclear-armed states have to make sure their threats are believable. By most accounts, this means they must affirm their willingness to launch regardless of their actual intentions of doing so, while governments must maintain deterrence postures that substantiate the threat. Not doing so, even by expressing reservations about the horrors of nuclear war, potentially undermines credibility (Boyd Reference Boyd2019, 115–17). Leaders who want to deter must communicate that they will resort to using nuclear weapons if necessary, regardless of their actual intentions.
This imposes an enormous restriction on citizens’ information access and thus their ability to exert control. In particular, it breaks down the effectiveness of representative control by making the adequate selection of representatives impossible.
Needing to communicate deterrence credibility becomes a problem in public discussions of nuclear policy when the threat and actual intention of nuclear use are misaligned (Cooke and Futter Reference Cooke and Futter2018). If the public agrees it is actually willing to use nuclear weapons in war, or if it decides it is not willing to threaten or carry out a nuclear strike, popular will can be freely implemented. Wars can be fought, or arsenals can be dismantled. But should the consensus land on deterrence but not actual use, the expression of this preference would by itself undermine credible deterrence. It is much more difficult to deter enemies with a threat when one openly admits one does not intend to use that threat. This is a democratic paradox, whereby expressing or even debating the public’s will makes it impossible to implement their intentions.
Deterrence then relies on a bluff that negates government transparency. To retain credibility, leaders cannot reveal when they would act on nuclear threats and when they prefer to bluff. The government must keep its strategic deliberations secret, limiting the public’s ability to know what their current representatives are practicing. Even if voters agree on nuclear use in some cases and on only deterrence in others, the line that justifies use cannot be publicly debated. It cannot be specified without compromising deterrent threats at the lower end of that line. Democratic debate—and with it, consent for any nuclear use—is hindered by the ambiguity needed to achieve the goals of deterrence.
Candidates for political office are caught in the same problem, needing to affirm their dedication to be able to practice deterrence as leaders (Cooke and Futter Reference Cooke and Futter2018). Candidates are, of course, free to take stances in favor of nuclear weapons or disarmament. However, among those who want to practice nuclear deterrence, expressed opinions are homogenized. Those who would not be willing to act on their threats must make themselves indistinguishable from those who would. Affirmations of intent to use become necessarily, even purposefully, ambiguous. Voting for a candidate who expresses a willingness to use nuclear weapons then cannot equate to giving consent to that prospect, since pro-use candidates can be selected by voters who mistakenly interpret their affirmations as an intent to practice nuclear deterrence. Alternatively, if we assume voters take candidates’ affirmations at face value, consent for use is invalidated for a different reason: insufficient diversity of candidates. Since it is impossible to represent voters who want deterrence alone, the selection of representatives will always be inadequate. If people cannot know what candidates truly believe, they cannot give a priori approval for nuclear use.
Taken to the extreme, the desire to uphold deterrence credibility becomes an argument to suppress dissent and further isolate nuclear decisions from the public (Falk Reference Falk, Cohen and Lee1986, 443–45). Since the democratic state’s actions depend on popular consent, popular expressions of dissent are potentially subversive to credibility. Debate would then have to be managed and conducted only when it reinforces popular support (Falk Reference Falk, Andersson and Dahlgren2019, 141; Gnesotto Reference Gnesotto and Rosselet2024; Rosow Reference Rosow1989, 582–83). This in part reflects a long elitist tradition of excluding the public from making foreign policy in the belief that it lacks competence to manage foreign affairs, with the peculiarity that the public’s supposed incompetence might alter the effectiveness of the deterrence policy without altering anything about the policy itself (Dahl Reference Dahl1985, 19–32; Leira Reference Leira2019). Despite this concern, formal limitations on freedom of expression have not been implemented in nuclear democracies to protect deterrence.Footnote 4 Rather, the nuclear monarchy has been the solution. Concentrating power insulates the credibility of deterrence from public opinion and leaves open the possibility that individual monarchs can decide to use nuclear weapons even when the people they lead oppose such use.
As described, deterrence is difficult to reduce to fundamental requirements because it is a psychological game relying on perceptions. The usefulness of deception for security therefore only exists if one believes deterrence dynamics require a credibly communicated intent to resort to nuclear use. States might choose to rely on existential deterrence, and assume the very existence of their nuclear strike capacity poses enough of a risk to adversaries to deter them from attacking (Sauer Reference Sauer2009, 750–51). Deterrence without any communicated intention to use could then be possible. So long as the nuclear arsenal is believed to pose a sufficient risk to potential aggressors, and so long as the public is certain the government is responsive to its preference to never use nuclear weapons, all is well.
Popular acceptance of this strategy may be complicated to bring about. It requires a broadly shared commitment to something approximating “doublethink.” People would have to accept the usefulness of the arsenal as a threat but also believe that it would never be deployed, despite knowing that the credibility of the threat relies on the possibility of its deployment. This is not impossible as a form of logic; the public may simply believe the enemy will be intimidated even by unrealistic threats. Still, it is a hard strategy to agree to with confidence given vulnerability to nuclear harm, particularly if the debate its adoption makes possible leads to a proliferation of voices that advocate for a more demanding deterrence framework. Belief in existential deterrence is then hard to square with the assumption of an aversion to high-risk nuclear strikes on which this strategy relies.
The deceptive politics of deterrence therefore follows from the perceived need to express a credible intention to use nuclear weapons. Such politics removes access to necessary information and blocks all modes of control, including by hindering the selection of elected representatives. Just like the nuclear monarchy, it results from the desire to maximize the effectiveness of one’s arsenal in the face of uncertain conditions for success.
Democratic Accountability for Nuclear Policy
The problems with a priori democratic control over nuclear decisions described in the preceding sections leave accountability after the fact as the only remaining mechanism. For the representative mode, this is necessarily imperfect without adequate prior elections. Still, accountability could make leaders responsive to the popular will, while pluralist or participatory modes could find ways to undo bad decisions. Here, nuclear weapons create secrecy, which hinders access to necessary information. This challenge can be overcome, but the fundamental features of nuclear weapons create shortcomings in establishing accountability for nuclear war that cannot be avoided.
Accountability for Nuclear Decisions: Nuclear Secrecy and Access to Information
Nuclear weapons are shrouded in secrecy and ambiguity across the nuclear states. As an extreme example, the US’s attempts to prevent other states, particularly the Soviet Union, from building their own nuclear weapons led it to declare that all information on nuclear science was “born secret,” meaning that any information, even from unclassified sources, pertaining to building nuclear weapons was secret by nature (DeVolpi et al. Reference DeVolpi, Marsh, Postol and Stanford1981, 11). Consequently, it was impossible to know what information fell under the label. This approach, while ultimately unsuccessful in controlling the spread of nuclear science, contributed to allowing “necessary” secrets to snowball to excessive proportions and ultimately impeded democratic control. Over several decades, it morphed into a complex regime, affecting policy areas not only beyond the essential secrets of proliferation and deterrence, but well beyond the subject of nuclear weapons in general (Wellerstein Reference Wellerstein2021; Wills Reference Wills2010). In the UK, France, and Sweden, similarly far-reaching secrecy regimes formed as a result of the inherent security implications of nuclear programs, as well as US diplomatic pressure and domestic political choices (Fraise Reference Fraise2023). These again did not apply to nuclear science alone: voters in these countries were kept in the dark not only about the decision to pursue a nuclear weapon, but about costs, risks, and achievements. This secrecy fundamentally undermines the public’s access to information necessary to judge government policy, and so hinders any mode of democracy. Citizens cannot know what policies their governments have pursued, and they lack the information necessary to judge the policies’ effects.
There are multiple rationales for nuclear secrecy: preventing nuclear proliferation, ensuring arsenal survivability, or maintaining deterrence credibility are all justifiably important goals. They also all have an unclear standard of success, and so are prone to overapplication (Fraise Reference Fraise2024, 6–7). To protect their deterrence credibility, states might hide the record of nuclear deployments, rationales for decisions, evaluations of acceptable losses, or the existence of disagreement among officials. If nuclear deterrence requires ambiguity to be credible, any hint of the state’s priorities and its assessment of risks is a potentially dangerous subversion. This is in part why, even after leaving office, leaders in existing democratic states often face criticism if they admit they may not have been prepared to use nuclear weapons (Cooke and Futter Reference Cooke and Futter2018, 507; Pelopidas Reference Pelopidas2022, 179). The same is true of arsenal survivability. India’s nuclear secrecy, for instance, resulted from attempts to withstand US scrutiny of its nuclear program, but now protects its arsenal’s survivability with such zeal that even the costs of nuclear energy programs are hidden (Mishra Reference Mishra, Behera and Mishra2022, 48–49).
The lack of clarity on how much secrecy is necessary allows for the implementation of further secrecy intended to limit political consequences. Relevant data on nuclear issues are at times purposefully kept secret by the military, scientific, and intelligence establishments to artificially create public approval (Elworthy Reference Elworthy, Marsh and Fraser1989). In the case of the US—the most transparent of the nuclear states—we now know that accidents were not only denied but their histories were also distorted to maintain a sense of infallibility (Lewis et al. Reference Lewis, Williams, Pelopidas and Aghlani2014; Schlosser Reference Schlosser2013). Costs of arsenals have been greatly understated (Schwartz Reference Schwartz1995). Since it is executives who control decisions on secrecy, the perceived importance of keeping nuclear knowledge secret makes it possible to hide information so as to escape scrutiny from voters or legislatures (Elworthy Reference Elworthy, Marsh and Fraser1989, 170–73). These secrets are kept for an exceptionally long time: US nuclear documents take an average of 57 years to be declassified (Connelly Reference Connelly2023, 60). This goes far beyond the timelines for electoral accountability or timely pluralist and participatory contestation.
Under these conditions, citizens cannot make judgments on how nuclear arsenal design, crisis management, or deterrence are conducted because they cannot know basic facts about them. This violates the condition of sufficient public information and prevents accountability for nuclear decisions through any of the three modes of democratic power.
Secrecy is not unique to nuclear policy and may be justifiable, even necessary, for the protection of democracies. There is a strong case for the necessity of nuclear secrecy. But its acceptability depends on the eventual publication of hidden information and on the ability to discuss what is being kept secret (Thompson Reference Thompson1999). For secrecy itself to be democratically controllable, it should establish the goals and scope of the secrets it wants to keep as well as some form of accountability to prevent abuse (Bok Reference Bok1980, 177; Mokrosinska Reference Mokrosinska2019, 14–15). This is not possible for secrets related to deterrence credibility. Revealing that one is lying about being willing to use nuclear weapons undermines credibility about as much admitting that one is unwilling to do so (Edyvane Reference Edyvane2015, 313–15). Secrecy that protects the arsenal’s vulnerability or proliferation might be limited to only more specific technical details of the design, production, and vulnerabilities of nuclear devices, though it is not clear which parts of existing secrecy regimes serve those goals (Schaper Reference Schaper2004). Controlling their scope would in all cases depend on accepting less certainty about protecting all the arsenal’s vulnerabilities.
Secrecy for deterrence credibility or arsenal survivability then results in large part from the same assumptions as the limits to democratic agreement discussed above. It assumes the actual capability to respond and the credible willingness to use nuclear weapons are essential for nuclear strategy. Rejecting these assumptions to rely on “existential” deterrence requiring neither deception nor a guaranteed second strike could make transparency possible, while a specification of exact proliferation concerns would remove far-reaching limits on everything but technical information. Crucial information regarding costs, risks, and histories of nuclear weapons would not have to be limited under these assumptions. Accountability would largely become possible.
Accountability after Nuclear War: Destructivity, Speed, and Independence from the Public
The final limit on accountability is a direct result of nuclear weapons rather than strategic choices about them. Nuclear weapons introduce the possibility for a type of decision that should have no place in a democratic state. Their basic features give decision makers the power to destroy democracy without an opportunity for voters to reestablish their control. Speed and independence from popular participation prevent reversals of consent by the electorate. The destructivity of nuclear strikes means that any nuclear use could potentially end the state and, with it, any institutions for representative or pluralist accountability.
Participatory Accountability for Nuclear Use: Reversing Nuclear Exchanges
The nature of nuclear strikes excludes even a minimalist form of participatory power. Their speed and the fact that the weapons require only a few hands to be used render the process of deploying them independent of the public’s participation in war. Since nuclear war is alienated from collective action, participatory modes of accountability are impossible.
The waging of conventional war relies on the participation of the public on the front lines, in factories, and in support roles. The public participates directly, though not always willingly, in the decision to fight. Popular consensus not to fight can therefore be a form of participatory control even after the war starts. Citizens may refuse to enlist, dodge drafts, or disrupt the war effort. More subtly, successful military operations require keeping up the morale of troops, which depends in part on their support for the cause and the government for which they are fighting (Connable et al. Reference Connable, McNerney, Marcellino, Frank, Hargrove, Posard and Zimmerman2018).
This stands in contrast to nuclear wars, whose irreversible and quickly escalating nature renders the withdrawal of consent unfeasible (Scarry Reference Scarry2014, 33–144). Since the participation of the public at large is not necessary to fire nuclear weapons, the refusal of significant parts of society to participate in war is not problematic. The concentration of decision-making power, together with the immediacy and intensity of nuclear impact, means there is no avenue for the public to contest nuclear wars once they have started. The initial decision requires no public participation, and its impact is so fast—mere minutes, depending on target distance and weapon speed—that there is too little time to halt or reverse it. The impact then results in exceptional, “out-of-ratio” damage to the target relative to the small group authorizing it (Scarry Reference Scarry1987, 154). Any escalation through retaliatory strikes could similarly happen at exceptional speed. Consent cannot be withdrawn in between rounds of fire. Nuclear weapons use cannot therefore be subject to adequate participatory control.
Democratic Accountability after Nuclear War: The Destructivity of Strikes
Beyond the speed of nuclear weapons, their destructivity presents further limits to democratic accountability after use against another nuclear-armed state. Using a nuclear weapon is likely to lead to the devastation of one’s own state and, with it, the institutions for representative or pluralist accountability. This is the expected outcome of nuclear use against a nuclear-armed opponent and a death blow to democratic functioning.
The destructivity of nuclear weapons ravages the institutions of accountability. This is mainly true because the most commonly imagined end point of nuclear war is the devastation of the state. A situation where one state uses a nuclear weapon against another state defended by a nuclear arsenal, whether as a first or second strike, is likely to end in mass death or profound societal collapse. Large-scale nuclear war could leave a “republic of insects and grass” (Schell Reference Schell2000, 65) with no humans left alive to organize a recall election. Nuclear war may not be capable of bringing about total human extinction (Kattan Reference Kattan2022), but it can kill enough of the demoi of democratic countries to destroy their institutions of state.
The combined damage of explosions and radiation sickness, economic and infrastructural breakdown, and the simple horror of nuclear bombardment will leave a society in which democracy, or any other political organization, is likely to be unmanageable (Borrie and Caughley Reference Borrie and Caughley2014; Cochrane and Mileti Reference Cochrane, Mileti, Solomon and Marston1986; Sanders-Zakre, de Verdier, and Lind Reference Sanders-Zakre, de Verdier and Lind2022). In such cases, democratic accountability obviously does not apply as the state in its preapocalyptic form will have ceased to meaningfully exist. Even if pockets of civilization reestablish democracy, they could not deliver accountability on behalf of a state that is no more.
The fundamental limit of nuclear weapons in a democracy is that they, by their nature, provide the opportunity to permanently destroy it. Other policies may require suspending democracy, but nuclear war makes such suspensions singularly irreversible. Other war-making powers could allow governments to start wars they cannot win, causing the state to be invaded or destroyed. However, this is both avoidable and reversible. Virtually all democracies require military operations to be approved by legislatures; while the reversal of such a discussion is politically difficult, it is at least physically possible because the democratic state and its demos can expect to survive (Peters and Wagner Reference Peters and Wagner2011). Many democracies have been occupied by authoritarian powers, but survived to become democracies once more after the occupiers were expelled. Mass death in nuclear war, on the other hand, eliminates all prospect of democracy returning (Axinn Reference Axinn, Cohen and Lee1983).
Crucially, first use against a nuclear-armed enemy always risks this outcome, as it is impossible to estimate with certainty the amount of damage that will result from retaliation. Even if a state somehow knew its adversary’s war plans, the fact that decisions are made by individual humans who might be far down the chain of command, with possible errors in communication or technical challenges to boot, means that the shape of the state’s retaliation is fundamentally unpredictable. Striking a nuclear-armed enemy risks escalation and can thus be expected to create these challenges for democratic accountability.
Even if parts of the state survive, nuclear war is liable to destroy modes of accountability, in the first instance because leaders are likely to be among the dead. As discussed, decapitation strikes are plausible, while civil protection is unreliable. Other organs of state will likely be crippled by the aftermath of the war and may not function adequately for elections to be held, disciplinary proceedings started, or criminal investigations undertaken. Legal accountability, arguably a form of pluralist accountability from a rival branch of government, would be reduced. The courts, which would likely be decimated after a nuclear war, have so far played little role in nuclear matters (Born, Gill, and Hänggi Reference Born, Gill and Hänggi2010, 226). Even international law, whose institutions might be left somewhat more intact following a limited nuclear war, offers little prospect of prosecution (Perkovich Reference Perkovich2020, 116).
It is the destructivity of nuclear wars that makes nuclear weapons fundamentally incompatible with democratic accountability. Their speed provides a less fundamental additional obstacle: it makes it difficult to establish who triggered the exchange of hellfire. This matters for accountability because judgments on the war are likely to be significantly based on the reasons for use and particularly on whether it was retaliatory. Such information may not survive a nuclear exchange. Overall, nuclear crises are chaotic: a lot is happening at the same time, with multiple simultaneous datastreams, deliberations, and commands (Blair Reference Blair2020). Both sides may believe the other launched first, or the exchange of fire may happen so quickly that no decision makers are left alive. It is difficult to hold leaders accountable for mistaken first strikes if voters cannot know that it was a first strike. Technical failures or misinterpretation could lead decision makers to “respond” to a first strike that is not in fact incoming: false alarms have nearly triggered nuclear war on numerous occasions (Lewis et al. Reference Lewis, Williams, Pelopidas and Aghlani2014, 7, 12–13). These challenges can be compounded by newer problems relating to the hackability of arsenals (Futter Reference Futter2018; Lin Reference Lin2021) and the risks of information warfare (Ajir and Vailliant Reference Ajir and Vailliant2018; Cimbala Reference Cimbala1999; Lin, Loehrke, and Trinkunas Reference Lin, Loehrke and Trinkunas2020). All told, it becomes exceptionally difficult to establish who started nuclear wars. Although other military conflicts may also arise through accident, there is a better chance of reconstructing the path to war when more people are involved over longer periods of time and with a greater need for situation-specific planning.
These limitations can be compounded by the existing culture of secrecy around nuclear weapons, which obscures possible causes. The incentives to engage in distortions of the record would further muddle post hoc accountability. Incentives to more favorably recast one’s role in a nuclear disaster would be all the greater in the aftermath of a nuclear war, which conveniently also provides an occasion to justify a suspension of democracy (Martin Reference Martin1990). The US did precisely this during its postwar military occupation of Japan, forbidding the spread of information about the bombings of Hiroshima and Nagasaki for several years (Brau Reference Brau1991).
There is, then, a fundamental limitation on democracy resulting from the capacity of nuclear weapons to permanently destroy the democratic state, with no prospects for reversal or accountability. The speed and unpredictability of nuclear exchanges could additionally make information more difficult to come by, while secrecy might further exacerbate this challenge. While these limits to democratic accountability may not happen in all cases of nuclear use, they can happen with any nuclear arsenal. Since the use of nuclear weapons can be expected to destroy the institutions that provide accountability for their deployment, nuclear weapons cannot be fully democratically controlled.
The Political Equality of the Demos and the Global Nuclear Threat
A final restriction on democracy stems not from nuclear weapons’ controllability within the state, but from their global reach. All modes of democratic control are only democratic on condition of the political equality of citizens. The scale of the impact of nuclear weapons, and the reach of the threat they entail, present a limit to political equality. Nuclear weapons constitute a coercive threat of force posed by a handful of states toward everyone on earth, yet—so long as there is no world government to override these states—they can only be controlled by a small subset of people within the nuclear powers.
For democratic control via any mode to be valid, it must be accessible to the public. A fundamental problem for theorizing democracy lies in defining adequate limits of the “public” for each governmental unit (Simmons Reference Simmons2013). It is not self-evident who has a right to be involved in decisions. Common solutions to this problem define the public based on who must live with a state’s decisions. Two major approaches exist in this vein: the all affected principle and the all subjected principle. As the name suggests, the first posits that all those who are affected by a state’s decisions should be allowed to participate in those decisions (Arrhenius Reference Arrhenius2023; Goodin Reference Goodin2007). The all subjected principle is a somewhat more limited approach, whereby only those who are subject to coercively imposed laws should have a say in shaping the idea of justice that authorizes such coercion (Goodin Reference Goodin2016, 370–73). When it comes to democracy and nuclear weapons, both principles lead to the same conclusions.
The all subjected principle defines the demos based on who can be coerced by the state. Coercion refers to the use or threat of violence to condition or punish behavior (Macleod Reference Macleod, Reidy and Riker2008). It requires the deployment of force as an impediment to the freedom of others. A state acts coercively when it uses its capacity to impose violence on citizens in a way that restricts citizens’ autonomy. This need not involve the actual use of violence. The threat of violence is in itself coercive because it engenders fear in order to constrain freedom (Abizadeh Reference Abizadeh2010, 123–25). The extensive use of incarceration is a coercive constraint even on nonincarcerated citizens because it threatens a negative consequence to misbehavior.
Thus, nuclear deterrence inherently creates a situation of coercion because it imposes the threat of force on populations as punishment for political action. For democratic target states, it is perhaps more obvious that state coercion limits autonomy by bounding their citizens’ political action. Citizens in authoritarian states do not make decisions on policy and thus do not lose any freedom to nuclear coercion. However, just because political decisions would have been limited in any case does not negate deterrence’s coercive limit on possible behaviors. The freedom of a teen caught shoplifting is limited by the threat of consequences from both the police and the teen’s parents. Both authorities are coercive, regardless of the existence of the other. Similarly, nuclear threats restrict people’s political freedoms by limiting the decisions the states representing them can make. Such threats have been central in facilitating Russia’s attempted coercion of the Ukrainian people. Recently announced changes in its nuclear doctrine expand the scope of these threats to third states, as Moscow could consider any attack on its territory by a non-nuclear state backed by a nuclear-armed one as an attack by the latter. It could consider nuclear retaliation even to attacks using small airborne weapons like drones (Putin Reference Putin2024). As another example, take the American position that nuclear weapons can be used against a state that is in violation of its obligations under the Non-Proliferation Treaty (Woolf Reference Woolf2022), which gives Washington coercive power over the nuclear energy choices of non-Americans. Although such threats are aimed at states, coercion occurs at a personal level, aimed not just at the representatives of state power, but at all citizens. Nuclear weapons fundamentally cannot discriminate between individuals and the state because they work by holding the civilian population hostage (Lee Reference Lee1985). The very basis of nuclear strategy is then collective punishment. These coercive nuclear threats affect all individual citizens, and the weapons’ global reach mean all of humanity is a coerced population.
Nuclear decisions made within individual states are thus incompatible with the all subjected principle of democratic inclusion. The indiscriminate nature of nuclear weapons and the scale of their impact also mean they cannot stand the test of the all affected principle. Nuclear weapons policy affects all those on earth. The production and testing of nuclear weapons so far has already had global environmental and health impacts (Higuchi Reference Higuchi2020; Jacobs Reference Jacobs2022). Most dramatically, of course, a future nuclear war would have global repercussions, not least through massive climate disruptions and subsequent famines (Toon, Robock, and Turco Reference Toon, Robock and Turco2008; Xia et al. Reference Xia, Robock, Scherrer, Harrison, Bodirsky, Weindl and Jägermeyr2022). Yet the regulation of the rules of justice governing nuclear weapons lies with a small subset of citizens in nine nuclear-armed states. In effect, the global population lives in a regime of oligarchy governed by the nuclear states. Even if these states could be perfectly democratic, only their voters would decide how many nuclear weapons exist, where those weapons are, and what constitutes a reason for their use.
This gap in democratic representation is most easily demonstrated for states that have explicitly placed themselves under the protection of a nuclear umbrella. Nuclear weapons decisions affect their citizens directly through taxes, exposure to targeting by other states, and implied support for the possibility of nuclear use (Egeland and Pelopidas Reference Egeland and Pelopidas2021, 245). But democratic control over decisions is taken away by the guardianship of the state supplying the umbrella. Even formal allies are not adequately or consistently included in decisions on nuclear use, for strategic reasons (Michaels Reference Michaels2022). While such “umbrella states” may have democratically reached the decision to outsource their nuclear decision-making capacity, they have alienated control and cannot now reassert it. They cannot, as it were, remove their elected representative from office. As an illustration, consider the secret replacement of nuclear weapons on Dutch territory in the 1980s, despite overwhelming popular condemnation leading to the country’s biggest-ever protests (Eichenberg Reference Eichenberg1983, 144; NOS 2020). In states under nuclear umbrellas, governments tend to support and enable the decisions of the state supplying the umbrella even if popular support for nuclear weapons is exceedingly low (Baron, Gibbons, and Herzog Reference Baron, Gibbons and Herzog2020; Onderco and Smetana Reference Onderco and Smetana2021).
At a global level, this leads nuclear-armed states (and their allies) to refuse to effectively pursue disarmament or even engage seriously with its possibility (Pelopidas and Verschuren Reference Pelopidas and Verschuren2023, 5; Ritchie and Kmentt Reference Ritchie and Kmentt2021, 85–87). They have rejected attempts to codify it into law through the Treaty on the Prohibition of Nuclear Weapons (UK Mission to the UN in New York 2018). This has happened in spite of generally high levels of support for the treaty (ICAN 2021; IFOP 2022; Nanos 2021; Shoben Reference Shoben2021). Since nuclear weapons thus constrain the freedoms of all the world’s citizens, their possession by national governments is fundamentally at odds with democracy.
Conclusion
The conditions for legitimate democratic foreign policy are fundamentally undermined when applied to nuclear weapons, and further limited as a result of attempts to cope with their security implications. Democracy requires that voters should be able to control their state’s policy through one of three modes of control, which each require political freedom, equality, and access to information.
Even if no other limitations exist, the speed and destructivity of nuclear weapons and their independence from control by all but the smallest groups of decision makers make accountability for nuclear war impossible. At the same time, the global reach of their threat means that no one state’s population represents a legitimate demos for their use. Nuclear weapons are therefore always incompatible with democracy.
Their speed and destructivity also make possible and incentivize a further web of limitations on democratic control as states attempt to avert their utter destruction. If nuclear strategy requires damage limitation, or if nuclear deterrence depends on constructing an assured second strike capability, the centralization of power in the nuclear monarchy prevents pluralist or participatory modes of control. If nuclear deterrence also requires making threats credible by communicating an intent to use even where there is none, citizens do not have the information necessary for representative control. If nuclear deterrence, damage limitation, or nonproliferation require far-reaching secrecy, relevant information cannot become available after the fact to allow for adequate reevaluation.
Abandoning or loosening some of these assumptions reintroduces partial modes of democratic control. States could choose to rely on delayed retaliation, or on a less certain immediate response, allowing for at least some pluralist processes to be applied to decisions about their use. The information limits imposed by deterrence ambiguity and secrecy would still mean this broader group of decision makers would lack a legitimate mandate from the public, but broader inclusion would be a step forward, nonetheless. Alternatively, if informational limits are loosened but rapid retaliation is still considered necessary, democracies could transparently delegate decisions to functionaries. Inherent restrictions on accountability would still hinder actual democratic control. If neither rapid retaliation nor deterrence ambiguity are necessary, the process for a priori democratic control lies wide open, albeit with fundamentally the wrong demos and inadequate accountability in the worst case.
All these changes are possible even when assuming that nuclear weapons are a net security benefit, useful in deterring outside threats or fighting nuclear wars. In existing democracies, this may be arguable. If no major outside threats exist, or if nuclear weapons are unhelpful in countering them, nuclear disarmament would completely remove their limitation on democracy.
The sobering conclusion that nuclear states will never be fully democratic should therefore not deter citizens from attempting to increase control. Partial fixes that deliver real, important gains are possible. They would require a fundamental change in the relationship of states with the vulnerability and uncertainty of the nuclear age. The point of this article is not to argue that such changes would be simple, nor does it argue for or against their desirability. Instead, it shows that some limits on democracy need not be accepted as necessary outcomes of nuclear weapons, but should be defended on the basis of the specific threats they protect against and the ways in which they do so.
Acknowledgments
The author would like to thank Benoît Pelopidas and Thomas Fraise, as well as the four anonymous reviewers, for their invaluable help in improving this article. This work was supported by the European Research Council under the European Union’s Horizon 2020 research and innovation program via grant no. 759707, NUCLEAR. The author reports there are no competing interests to declare.
Open access
