The Essence of Cross-Domain Deterrence

Abstract

Both deterrence theory and deterrence practice are evolving to address contemporary strategic challenges. In the military domain, states progressively integrate and synchronise military operations. Outside of it, they exploit grey zone strategies that combine different instruments of influence across multiple domains. These developments are now giving birth to a new wave of thinking about cross domain deterrence (CDD), what it precisely entails, and what favouring conditions are necessary for it to be effective. This chapter situates CDD in the context of today’s challenges, and identifies the prerequisites for these favouring conditions based on a review of a rather diverse body of literature. It finds that one strand of that literature predominantly focuses on practical and technical prerequisites in order for CDD to be effective, leaving the framework of traditional deterrence theory intact. It also finds a second strand that holds that the nature of today’s challenges requires more than mere innovation in application. The ideas about deterrence proposed by this second strand are expanding on common understandings of deterrence to the extent that deterrence is no longer only about fear nor about convincing opponents to refrain from certain behaviour. The conclusion summarises the findings and elaborates their implications for theory and practice.

Tempora mutantur nos et mutamur in illis.

1 Introduction

Deterrence is about convincing adversaries to refrain from certain behaviour through the prospect of costs that outweigh the benefits.¹ As related in the preface to this volume by Osinga and Sweijs, deterrence has been a central tenet of strategic practice throughout history,² even if its logic was only clearly articulated in the aftermath of the Second World War. Deterrence scholarship has since then evolved in four consecutive waves. The first, second and third wave of the deterrence literature, which emerged during the Cold War, tended to almost exclusively focus on deterrence of high-intensity aggression including most importantly the possible use of nuclear weapons alongside large scale conventional invasion.³ Lower-intensity threats which were considered mere nuisances were largely left outside of the scope of investigation.⁴ However, these became more important in the 1990s with the demise of the Soviet Union and the emergence of non-traditional threats such as terrorism.⁵ This gave birth to the fourth wave of deterrence literature that focused on the question whether deterrence would work against such threats that emerged in the 1990s and 2000s.⁶ Over the past decade, a new body of ideas has been emerging concerning the application of deterrence in today’s strategic environment. An important characteristic of our age is the proliferation of ways and means by which hostile activities can be perpetrated. Accordingly, strategists have started to pay more attention to the application of deterrence in new domains and to cross domain deterrence (CDD), across both traditional and new domains. This chapter appraises the contribution of the emerging body of cross domain deterrence literature to deterrence theory and deterrence practice. It explains the context in which theories of cross domain deterrence have emerged and elaborates different conceptualisations of cross domain deterrence distinguishing between two different approaches. The conclusion summarises the findings and elaborates their implications for theory and practice.⁷

2 The Origins of Cross-Domain Deterrence

The shift in attention to CDD can be explained by two principal challenges. The first challenge relates to the progressive integration and synchronization of military operations across different domains (land, air, sea, cyber, and space) and the inherent disharmony between different levels of war (strategic, operational and tactical).⁸ This is because military organizations aspire to better integrate physical, social and communication technologies in their ability to apply violence in the pursuit of political objectives, leading to strategic compression and cross domain warfare. Multi-domain operations concepts are being developed to guide efforts to synchronise actions both horizontally across domains and vertically across levels of war.⁹ In light of the cross-domain nature of the challenge, strategists are envisaging analogous responses, including CDD.

The second challenge relates to the increased salience of “hybrid” or “grey zone” strategies that feature the simultaneous employment of military and non-military instruments, typically below the conventional military threshold, in an ambiguous fashion in order to evade attribution, with the goal to exploit adversary’s vulnerabilities, in the pursuit of political objectives.¹⁰ While the analytical value of the labels as such have caused considerable debate,¹¹ the real-world impact of these strategies poses a serious strategic challenge. Their increased salience stems from the enormous costs associated with interstate wars, which makes major military powers disinclined from waging actual hot wars against each other. These powers therefore try and find alternative ways to achieve their political objectives—in line with the original tenets of the coercive diplomacy literature. Furthermore, the increased salience of grey zone strategies also derives from the opportunities offered by new avenues to hurt opponents due to technological and societal developments because of the global wiring of societies over the past quarter century. Strategically innovative actors have been making frequent use of these avenues over the past decade to considerable effect. These developments have led scholars and strategists to start thinking about the use and utility of cross domain deterrence in dealing with adversaries employing cross domain strategies also outside the traditional military domains.

Authors from both sides of the Atlantic generally concur that cross-domain deterrence involves the use of threats in one domain to deter activities in (an)other domain(s). Some authors define cross domain deterrence exclusively in the military domains land, sea, air, cyber and space albeit at different levels of abstraction. James Scouras, Edward Smyth and Thomas Mahnken assert for example that it is the prospect of retaliation from one domain to another which constitutes the essence of CDD.¹² It is worth noting that the authors seem to focus exclusively on deterrence by punishment rather than denial. James Dawkins emphasizes that CDD involves the use of specific weapons rather than mere threats or retaliation in general. His conceptualization includes both punishment and denial strategies and draws attention to the actual instruments by which deterrent effects are to be achieved.¹³ Despite the differences in abstraction, these authors understand CDD to operate specifically within the military domains.

Other authors also consider non-military domains and instruments. Accordingly, Manzo Vince understands CDD to refer to deterrent efforts on land, at sea, in the air, in space, in cyberspace and through economic sanctions as well as other non-violent instruments.¹⁴ King Mallory, too, includes both non-military instruments and non-military domains, arguing that CDD is about preventing escalation in any domain and across them.¹⁵ Sean Monaghan, Patrick Cullen and Njord Wegge assert that contemporary deterrence strategies should include an array of non-military means to detect, deter and respond in a tailored way.¹⁶ More generically, Erik Gartzke and Jon R. Lindsay conceive of CDD as “the use of threats of one type, or some combination of different types, to dissuade a target from taking actions of another type to attempt to change the status quo”.¹⁷

3 The CDD Literature: Practical Innovation Versus Theoretical Reconceptualisation

3.1 Innovation in Practical Application

Over the past decade or so, two approaches to CDD have emerged. The first approach emphasises that CDD requires merely the extension and refinement of the practical application of general deterrence theory. Authors within this approach argue that deterrence has always been cross domain in nature, albeit only in the context of traditional military domains.¹⁸ Despite the emergence of new domains, deterrence in today’s world is as such not different, so they argue.¹⁹ Accordingly, Christopher Buckley asserts that cross domain deterrence has been practiced in the West for a very long time simply because “deterrence policy and strategy are concepts too big to be constrained in a single domain.”²⁰ Still, it is acknowledged that particular aspects of deterrence in practice are in need of refinement. Gartzke and Lindsay, for example, observe that the “increasing complexity in the entire portfolio of means now available now appears to necessitate the refinement of deterrence as both a military and political process.”²¹ But what does refinement precisely entail for CDD to be effective? The authors in the refinement camp tend to focus on practical problems associated with the necessary conditions for effective CDD. Important requirements of deterrence in general that they focus on include attribution, threat credibility and proportionality, signalling and escalation management.²² Attribution depends on the ability and the willingness to ascribe responsibility for a particular act to an actor. Without the possibility of attribution, transgressors can act undetected and therefore escape allocation of blame. Credibility is rooted in the perceived capability and willingness to act. It is crucial for deterrence to work because adversaries have to believe they will suffer negative consequences for their wrongdoings. Threats that are not credible are irrelevant for deterrence purposes. In general, threats which are proportional to their triggers are likely to be perceived more credible than disproportionate ones. Signalling refers to the process of communicating one’s willingness and capabilities to act to instil that belief in the adversary. Attribution, credibility, threat proportionality and signalling together are prerequisites for escalation management, which is the regulation of the intensity and scope of the conflict.²³ These four themes are given elaborate treatment in the CDD literature in the context of today’s challenges.²⁴

3.2 Attribution

CDD authors point out that the emergence of new domains and the proliferation of hostile actors complicates attribution in the cross-domain context. Both state and non-state actors can dispose of a range of military and non-military instruments to cause damage from afar. Geographic proximity is no longer required. Proxy wars have become increasingly salient, in the context of a steep increase in internationalized intrastate conflicts.²⁵ The democratization of the means of violence in combination with the foggy nature of new domains, especially cyber space, are singled out as formidable challenges to attribution in the cross domain context.²⁶ Special Forces and irregular combatants without uniforms, both of which are hard to identify, constitute key actors of choice to carry out contemporary military operations.²⁷ Low cost unmanned aerial vehicles enable conflict actors, including non-state actors such as ISIS in Iraq and Syria and the Houthis in Yemen, to target objects of value from a safe distance.²⁸ Individual grey zone events “are difficult to distinguish from one-off actions, statecraft, or diplomacy”.²⁹ In the virtual realm, offenders can avoid attribution by hiding behind the anonymity provided by cyber space.³⁰ Though cyber attribution is possible in general, it is seldom certain in particular cases.³¹ Perpetrators can exploit the complexity of cyberspace to pretend they act on behalf of a third party.³² Furthermore, collecting sufficient evidence about the origins of cyber-attacks may take months.³³ By that time, too much time has passed for an effective response to effectuate deterrence.³⁴ Attribution in space brings its own set of challenges. The devices that scan the environment, those which keep track of space systems’ health as well as those which identify the origins of the hostile activities, have many blind spots.³⁵ Additionally, actors in possession of space assets will likely only know they have been attacked because of the effects of the attack. Also, space weather can produce damage akin to the adversarial action.³⁶ The origins of the attack and the identity of the perpetrator are therefore hard to pin down. The widespread use of non-military measures adds another layer of complexity to the attribution challenge. The different actors taking part in election meddling, disinformation campaigns, espionage, intelligence theft, critical infrastructure infiltration, political corruption or market stock manipulation may be hard to identify in acceptable time frames, or at all.³⁷ Overall, recent technological progress combined with the proliferation of actors and domains complicates attribution in both military and non-military domains and across them.

Yet, CDD scholars come up with various solutions to these obstacles to attribution which are first and foremost practical and technical rather than theoretical in nature. In general, the scholars acknowledge that the solutions to the attribution challenge across domains require international and inter-organizational cooperation, information sharing, technical expertise, analytical skills as well as political will. To deal with the hard-to-identify non-state actors and the wide spectrum of instruments at their disposal, it is suggested to attribute and threaten those upon whose help the non-state actors may be dependent. The assumption here is that these supporting actors are often states, which should render attribution easier.³⁸ In cyberspace, solutions are sought in the combination of technical, cognitive and behavioural expertise to help lift the fog of anonymity and enable effective responses.³⁹ It is argued that cross triangulation of the digital footprint, geographical origin, modus operandi, as well as geopolitical intent, renders attribution in cyber space in fact possible in the fast majority of cases.⁴⁰ Adversarial interest is also singled out as being particularly relevant in the attribution process.⁴¹ Additionally, cyber-attacks intended to cause serious damage are more likely than not to be accompanied by non-cyber measures, which should also help identify the potential perpetrator.⁴² Lack of political will may be a bigger obstacle than technical limitations. It is pointed out, for instance, that Obama’s administration was well aware of the identity of the election meddling perpetrators in 2016 but nonetheless decided not to ascribe responsibility publicly so as to avoid further escalation.⁴³ It is also argued that the attribution problem can be bypassed by heavier reliance on deterrence by denial. Deterrence by denial in the cyber context can be further enhanced by military, political or economic measures to secure physical infrastructure and supply lines.⁴⁴ Attribution in the cyber domain is thus certainly more complex but authors argue that obstacles can be solved with the appropriate amount of expertise and will.

Myriad solutions to attribution problems in other domains are also proposed. In space, CDD authors focus not only on the hardening of satellite assets to bolster deterrence through denial; they also suggest the strengthening of situational awareness through monitoring capabilities that enable attribution; the assessment of geopolitical risk based on analysis of strategic intent and space capabilities; and the traditional exploitation of human intelligence sources.⁴⁵ In the terrestrial military domains, it is argued that attribution is progressively less of a problem. Attribution of actions executed by irregular forces can exploit data from social media, photos and position tracking applications.⁴⁶ Western countries were thus able to identify and attribute Russian troop movements near the Ukrainian border during the summer of 2014. Likewise, the US was able to quickly ascribe the 2019 hostile activities in the Persian Gulf to the Iranian Revolutionary Guards units.⁴⁷ North Korean missiles launches over the past decade were also time and again detected by US satellite systems.⁴⁸ Finally, attribution of actions outside these military domains can also be enhanced, it is suggested, by tracing overall patterns. Authors point out for instance that one diplomatic visit of a foreign official may not be significant, but when placed in a broader picture, and when combined with other actions, it may allow for the identification of an overall pattern of coercive activities.⁴⁹ On a more practical note, Linda Robinson et al. suggest that hybrid campaign analysis units that can expose systematic patterns and generate more holistic threat pictures, will contribute to cross domain attribution capabilities.⁵⁰ In sum, authors in the refinement strand suggest that attribution challenges can be addressed and overcome largely through the implementation of a series of practical recommendations.

3.3 Threat Credibility and Proportionality

The issue of how to render deterrent threats credible in CDD is made more complicated by the inherent disproportionality of responses across domains and instruments of power.⁵¹ In short, decision makers lack agreed-upon guidelines for proportional responses to the wide array of potential hostilities in CDD.⁵² This is different from within-domain deterrence as Thomas Schelling’s captured in his observation that “there is an idiom in this interaction, a tendency to keeps things in the same currency, to respond in the same language, to make the punishment fit the character of the crime.”⁵³

The conversion mechanism between violent and non-violent actions and their effects is seen as the biggest hurdle to threat proportionality.⁵⁴ Using violence against non-violent hostilities such as theft, espionage, infiltration or election meddling is likely to be seen as disproportionate by many. This is further exacerbated by the multitude of state and non-state actors, each of which may have different beliefs about the appropriate conversion ratio between violent and non-violent measures.⁵⁵ As one scholar puts it, “while the United States could threaten to retaliate against cyberattacks asymmetrically through economic sanctions or military threats, there is a significant chance that such actions would appear escalatory, disproportionate, or otherwise inappropriate to the American public or the international community.”⁵⁶ Furthermore, actors operating through cyberspace are likely to have different degrees of tolerance for escalation risks because of their “anonymity, invulnerability, and global flexibility”.⁵⁷ This exacerbates the proportionality asymmetry because it is not clear how individual actors and groups appraise the severity of cyberattacks. Moreover, retaliatory threats involving actions in cyberspace may have significant second and third order consequences. Their ultimate proportionality is thus hard to assess beforehand.⁵⁸ Additionally, propaganda, infiltration, espionage, economic sanctions and stock market manipulations tend to produce their effects slower than the implements of violence on land, on sea, in the air or in space.⁵⁹ Ultimately the conversion ratio between violent and non-violent measures is unclear because the former tend to have more direct and immediate effects while the latter tend to rely on more gradual and second order effects.

Even when it comes to conversion within single instruments of violence or against similar targets, proportionality assessments are not necessarily straightforward. For example, in space, the problem of proportionality is exacerbated by the differences in value which the individual actors tend to place on the same assets. The US is much more dependent than China on its satellites, both for military and civilian purposes. Therefore, the simple cost-benefit equation of destroying one satellite for each one destroyed by the enemy is asymmetric and therefore disproportional.⁶⁰ In fact, the costs incurred by the US are disproportionately higher.⁶¹ It is argued that this undermines the credibility of US threats to harm space assets of states that do not rely on these systems in equal measure.⁶² Finally, attacks against targets in and through new domains may cause considerable collateral damage which again further complicates proportionality assessments. For example, retaliation against space objects may cause debris which can threaten both friendly and hostile activities in outer space.⁶³ Alternatively, threatening terrestrial attacks in response to hostilities against satellites may be deemed disproportionate because the former may result in human casualties while the latter is likely to produce only material damage.⁶⁴ In this regard, authors point at patterns of failed deterrence when it comes to deterring less destructive hostilities.⁶⁵

In tackling proportionality and credibility in CDD, scholars propose various solutions. In general, authors discuss strengthening cross domain deterrent postures by explicitly formulating cross domain threats in deterring domain specific actions, for instance by including conventional or even nuclear responses to enhance the credibility of threats seeking to deter attacks on critical assets in cyber space and space.⁶⁶ Some treatments suggest that a degree of proportionality can be established by focusing on the effects of specific actions rather than on the specific instruments used in this process.⁶⁷ Schneider, for example, speculates that cyber sabotage of a radar system can be countered proportionately by the electromagnetic jamming of a similar target. However, as she notes, this is likely to work better with direct, kinetic effects than with less direct, and less tangible effects. Smeets and Lin point out that states can build up credibility by regularly deploying a capability in practice. Actors with a clear track record of using particular capabilities, whether violent or not, may be able to develop sufficient reputation to offset the lack of credibility posed by the instruments themselves.⁶⁸

CDD is seen as particularly relevant in the context of cyber deterrence. It is argued that cyber deterrence also requires a broad mix of military, diplomatic, economic and legal measures,⁶⁹ synchronised within an overall deterrence posture. To bolster credibility, “cyber deterrence needs to be a well-integrated defence component that is in tune with non-cyber policy initiatives, and to accomplish this, policymakers need to juxtapose carefully cyber deterrence means and ends to those involved in broader defence policies”.⁷⁰ For the sake of credibility, cyber deterrence improvements need to be “mutually reinforcing”, to have the potential to surprise the adversaries as well as to flexibly manoeuvre between both denial and punishment options. Furthermore, some argue that states are likely to consider truly destructive cyberattacks as regular acts of war, which should make threats of conventional military retaliation credible, as international law already allows such responses when the principles of necessity and proportionality are adhered to.⁷¹

Some scholars are also optimistic about the credibility of other non-violent measures. It is argued that election meddling too can be deterred by the threats of economic sanctions targeted against energy, banking and defence sectors.⁷² Additionally, in response to serious threats posed by authoritarian governments, Western democracies can threaten to disrupt the former’s protected information sphere and to leak sensitive information about the regime’s misconduct to the foreign public.⁷³ Finally, as Jervis reminds us, it is necessary to realize that “threats need not be completely credible in order to be effective”: it may be enough for threats to be probable rather than certain, no matter whether one employs violent or non-violent measures; “credibility is not an objective, nor is it a property of the person or state making the threat. Rather it is ‘owned’ by the target.”⁷⁴ This underscores that conversion rates ultimately hinge on the perception of the beholder.

To deal with the proportionality issue as it relates to violent instruments, a generic solution that is proposed is to rely on a set of strategies to resolve the proportionality issue in different contexts. Anthony Juarez for instance lists counter-force, counter value, tit for tat, denial and ambiguity as potential options.⁷⁵ It is also argued that the supposed asymmetries in interests and values as related to space should not be overrated. For example, while some nations may not be as dependent on the satellites for their military utility, they may still value them highly for economic, cultural or prestige reasons and will therefore consider them vital assets.⁷⁶ With respect to the US it is said that it can credibly threaten retaliation against attacks aimed at its space assets everywhere precisely because its space assets are so important.⁷⁷ To deal with the disproportionality issue, it is recommended to focus on the overall effects rather than on specific instruments. In the context of space, this should involve a broad menu of “kinetic or non-kinetic attacks on adversary command, control, communication, intelligence, surveillance, and reconnaissance (C3ISR) and reconnaissance, surveillance, targeting, and attack (RSTA) assets in the land, air, and sea domains”.⁷⁸ Overall, the recommendations from authors who are concerned with the effectuation of CDD in this refinement strand focus on establishing proportionality and increasing credibility through the adoption of a combination of these practical measures.

3.4 Signalling

Signalling in the cross-domain context is more complex for two reasons which are closely related to establishing proportionality. First, it is harder to relate signals about particular actions in one domain to anticipated reactions in another in line with Schelling’s previously cited observation. Moreover, while signals relying on military instruments may resonate more than those relying on non-military instruments, they also come with higher risks of misunderstandings. For example, a signal of resolve to respond to cyberattacks by moving platforms for the launch of conventional or nuclear weapons may be easily interpreted as a preparation for hostilities rather than as an adjustment of the deterrence posture.⁷⁹ Conversely, signalling purely in cyber space may be difficult, because unlike in other domains, the relevant infrastructure of that domain is not under the exclusive control of the government.⁸⁰ Consequently, signals may get lost or be ignored by the adversaries.⁸¹ Striking the right balance between over signalling on the one hand and under signalling on the other thus constitutes a paramount challenge to communication in the cross domain context. Second, a number of modern instruments and tactics are effective precisely because they are secret. This implies a serious trade-off for the signaller who runs the risk of losing the advantage yielded by the capability the moment it signals its possession. After all, it allows adversaries to devise effective countermeasures, which is especially pertinent in cyberspace, but also applies to hybrid operations.⁸² This then raises the question of how to signal true capabilities while maintaining their utility for prospective hostilities.

CDD authors discuss an assortment of, once again, predominantly practical measures to meet these signalling challenges. For example, it is argued that the issue of tying threats across domains can be tackled by synchronized signalling at different levels of conflict. At the political level, signalling takes the form of public and private communication as well as norm development, at the strategic level it conveys the developments of doctrines about the actions and reactions, and at the tactical level it contains the actual application of particular forms of power to demonstrate resolve and capabilities.⁸³ The US successful orchestration of this kind of effort across different levels to signal its discontent with Chinese espionage activities during the Obama administration is a case in point.⁸⁴ To signal the relationship between different domains is thus possible but it requires the synergistic employment of communication across more levels than previously.

To tackle the issue of secrecy versus effectiveness, several suggestions are offered. One approach, which builds upon the recognition of the temporary nature of cyber capabilities noted by several scholars, may rely on building up a redundant portfolio of those capabilities, some of which can then be regularly used to demonstrate cyber capabilities.⁸⁵ The logic behind this option is that cyber weapons by their very nature are transitory—they lose their effectiveness over time because cyber vulnerabilities are exposed and patched.⁸⁶ Therefore, they can be disposed of for signalling both capability and resolve without losing their effectiveness. Other scholars suggest alternative ways that bypass the issue altogether by public advertisement of attribution technologies.⁸⁷ This way actors signal both their will and capabilities to allocate blame if necessary alongside announcement of the type of weapons and/or attacks they consider to be the most threatening. It is also argued that cyber weapons in fact possess signalling advantages compared to traditional instruments on the grounds that they can be used in a demonstration of force without starting the conflict they seek to prevent because they do not necessarily involve violent, kinetic effects.⁸⁸ That quality renders them sufficient to signal intent while avoiding escalation.⁸⁹ Additionally, states can rely on a combination of public speeches and real action to signal their cyber-capabilities. More states have been openly talking about the possession of sufficient cyber capabilities in recent years, some of whom have followed up with actions, such as Russia in Ukraine and the US in the context of its strategy of persistent engagement.⁹⁰ Some actors may be willing to signal more than others because of their strategic culture.⁹¹ Signalling one’s capabilities may even not inevitably lead to the loss of effectiveness, because not all adversaries are able or willing to patch the revealed vulnerabilities,⁹² Moreover, cross domain signalling by military, political or economic measures may alleviate the problem with clandestine capabilities because conventional forces, diplomatic pressure and economic sanctions do not lose their effectiveness once exercised in an adversarial relationship.⁹³ CDD refinement authors thus conceive a combination of these practical measures to facilitate signalling across domains and solve the trade-off between secrecy and effectiveness.

A number of suggestions have also been presented with respect to signalling outside of the cyber domain. King Mallory observes that signalling can rely on explicit moral Manicheism through clear verbal statement that there is no middle ground or grey zone in order to persuade adversaries that any kind of hostilities, direct or indirect, will lead to retaliation.⁹⁴ Signalling of both will and capability is also possible against hybrid intrusions, especially with rapidly deployable response teams of police and Special Forces which convey to the adversary that it is not likely to achieve its interests. Other means of signalling include implicit warnings reflected in changes in postures in combination with public statements.⁹⁵ Others suggest that “acts of retorsion” including economic sanctions and diplomatic coercion/isolation are perfect signalling instruments.⁹⁶ The authors in this literature have thus come up with a broad portfolio of signalling measures across all domains.

3.5 Escalation Management

The combination of issues discussed in relationship to attribution, threat proportionality and signalling makes escalation management much more difficult in CDD.⁹⁷ The attribution problem injects uncertainty into the deterrence relationship because it renders unclear under which conditions the deterring actor will deem it appropriate to escalate. Challenges associated with credibility and proportionality undermine basic tenets of successful escalation management simply because of the unpredictable dynamics across domains. Complexity of signalling further befuddles escalation management in practice because it is unclear whether signals are both sent and received. Accordingly, the diversity of escalation dynamics of cross domain deterrence is singled out as “a core analytic issue”.⁹⁸

An assortment of sources of instability for escalation management in CDD are discussed many of which are directly or indirectly related to issues addressed previously. First and foremost, there is no shared framework to describe and therefore manage escalation across domains.⁹⁹ Without such a framework “decision makers will have difficulty distinguishing between proportional and escalatory attacks and reprisals that cross from traditional strategic domains into these newer ones and vice versa”.¹⁰⁰ Second, there are many sources of instability when it comes to particular measures and weapons across domains. Western superiority in conventional weapons motivates adversaries to actively seek and exploit asymmetric and diverse measures with varying kinetic and non-kinetic effects and with differing degrees of proportionality.¹⁰¹ Some of the instruments and tactics operate across domains that cross potential thresholds faster than in the past.¹⁰² In this context, the use of unmanned and semi-autonomous systems and, in the future, other AI enhanced weapon systems may be particularly destabilizing.¹⁰³ Furthermore, the nature of the cyber and space domains and the character of technologies used in these domains may generate escalation risks through first-strike instabilities.¹⁰⁴ This renders these domains not only inherently unstable but also implies spill over effects to other domains in CDD.¹⁰⁵ Consequently, the anticipated effects are sometimes difficult to gauge before their actual employment.¹⁰⁶ Third, proportionality perceptions of actions in particular domains vary considerably from one actor to the next.¹⁰⁷ For example, Russia and China tend to see the integration of military, political and economic tools in a much more holistic fashion and for this reason they are likely to appraise the conversion rate between individual domains differently. As Adamsky explains in this volume and elsewhere the Russians combine nuclear, conventional and information measures to deter continuously and across domains.¹⁰⁸ Dean Cheng in this volume and elsewhere, describes the Chinese understanding of deterrence to involve “political activity and psychological warfare”.¹⁰⁹ Any combination of these three challenges may hinder attempts at successful escalation management in any particular conflict.

CDD authors once again have come up with a range of proposals how to address these issues. First and foremost, they agree that it is necessary to develop a shared framework which would encompass the expectations for escalation dynamics.¹¹⁰ There are several distinct approaches to the development of a shared framework. Some scholars point to the salient function of international law. Game theorist James Morrow, for example, argues that the developments in international law can constitute a first step towards the development of such a framework. Law alleviates the uncertainties about proportionality by explicitly stating what is acceptable, what is the appropriate response as well as how these actions relate across specific domains. As a coordination mechanism, law contributes to a common understanding of proportionality. Though it is unlikely to eliminate competition, it may channel hostilities into more manageable forms.¹¹¹ In this vein, others argue that an international cyber warfare convention would improve the prospects for both deterrence by punishment and by denial “by clarifying what counts [as] an act of cyber-aggression and what level of retaliation is deemed acceptable by the international community, an ICWC would thereby enhance states’ capacity to adopt and communicate an effective deterrent posture”.¹¹² Another perspective on framework development builds upon the notion of different kinds of escalation ladders, including a “provocation framework” to elucidate thresholds in “grey-zone” competition and improve escalation management by helping “policymakers understand the value of their actions and how reciprocal and proportional responses achieve strategic effect… ”.¹¹³ Such a framework is supposed to work as an explicit “declaratory policy” to signal both commitment and expectations of proportionality.

Attempts have also been made to further develop escalation ladders to establish the logic of escalation in the context of single domains,¹¹⁴ as well as in the interaction between different domains.¹¹⁵

Here, an interesting schism about whether to focus on instruments or on effects emerges. On one hand, it has been argued for cross domain frameworks to be based on the “severity of various military and non-military actions based on the full range of their anticipated effects, rather than assuming that military actions represent an escalation from non-military actions”.¹¹⁶ On the other hand, “cyber operations might not have the same saliency or emotional effect as conventional operations—even when they create the same physically destructive effects”.¹¹⁷ This second line of research, therefore, indicates that psychological effects vary across different instruments regardless of the physical damage these instruments cause.¹¹⁸ Relatedly, it is also possible that cyber instruments are “poor tools for escalatory purposes” because of the limited cost-generation potential of offensive cyber operations”.¹¹⁹ This echoes the observations that actors tend to deescalate rather than escalate in the cyber domain because cyber tools enable actors to release tensions by “sub crisis management manoeuvring”.¹²⁰ These practical ideas concerning the development of shared frameworks, whether alone or in some combination, thus seek to address problems associated with escalation management in CDD.

Authors working on CDD have also proposed several solutions to tackle the problems of destabilizing measures and of varying perceptions of proportionality. To deal with the former, it may be wise to avoid offensive activity with specific weapons (nuclear) and against specific targets (command and control).¹²¹ Additionally, the vulnerable assets should be better protected. Satellites should be dispersed across broad space and have their passive and active defences improved.¹²² Economic interdependence too, may have a stabilizing effect by motivating restraint in interactions. Finally, new domains tend to create mutual vulnerabilities which can incentivize prudence and caution out of fear for retaliation. Declarations of restraint as well as the developments of some basic thresholds for response are seen as time tested mechanisms.¹²³ This may prove particularly useful in the cyber and space domains. Other recommendations lean towards the opposite direction, with experts suggesting not to show restraint but rather to show resolve and the will to retaliate in order to establish escalation dominance up front.¹²⁴ How to combine the two contradictory approaches continues to be a pernicious problem, and requires future research and also practice to solve.¹²⁵ The ideal situation is the one in which each actor can exercise restraint but still radiate resolve.¹²⁶ These diverging recommendations imply that successful escalation management depends on the practical application in particular contexts rather than on general truths. Overall, the solutions for escalation management proposed in the refinement camp are of a predominantly practical nature. They build on, but do not extend, the logic of classic deterrence theory while offering a range of valuable practical insights how to effectuate CDD in today’s world.

4 Refinement and Reinterpretation—Expansion and Reconceptualisation

The second strand in the CDD literature argues that the character of contemporary challenges requires the broadening and deepening of our understanding of deterrence. Instead of offering practical recommendations on how to effectuate CDD in light of changing strategic conditions, authors propose theoretical and conceptual additions and innovations to existing concepts of deterrence rooted in deterrence by punishment and deterrence by denial. Some authors offer additional theoretical concepts to update deterrence; other authors in effect seek to reconceptualise deterrence in light of the nature of cross-domain challenges. This stems from the recognition and conviction that new domains require new approaches. Finding incremental practical fixes simply does not suffice, so they argue. It is thought that the traditional parameters that may have allowed deterrence to work in previous times, simply no longer hold in the context of today’s multipolar, connected and complex strategic environment. The greater diversity of actors that dispose of an even greater diversity of means that can successfully threaten each other in this environment either undermines deterrence or may even render it impossible.

4.1 Refinement of Traditional Concepts of Deterrence

Some of the additions are theoretical refinements. For instance, in order to deter across both old and new domains, concepts such as cumulative, punctuated and layered deterrence have been introduced. The concept of cumulative deterrence is based on Israel’s strategic experience. Israel has defended itself against a diverse spectrum of attacks conducted by state and non-state actors over a long period of time, partly by “attacking the rival repeatedly in response to specific behaviours, over a long period of time, sometimes even disproportionally to its aggressive actions”.¹²⁷ Or, as we put it in a previous publication on cross domain deterrence in the context of hybrid conflict, “cumulative deterrence conceptualises deterrence as a continuous, longer term process in which a one-off transgression does not spell failure but adversarial behaviour is shaped by the deterrer in a concerted effort.”¹²⁸ Within the framework of cumulative deterrence, deterrers understand the necessity of absorbing some attacks in order to prevent others. This marks a clear departure from a more absolutist notion encapsulated in traditional deterrence approaches aimed at deterring all attacks. The concept of cumulative deterrence may indeed be better suited to the less impactful but more frequent and ambiguous amalgamation of contemporary security threats and actors rather than to deterring the threat of a nuclear attack.¹²⁹ Another alternative is punctuated deterrence, which conveys punishment to address a series of actions and cumulative effects. The difference between cumulative and punctuated deterrence is that within the framework of cumulative deterrence, deterrers respond continuously over long time periods to single attacks, while in the case of punctuated deterrence they respond gradually over time and in a punctuated manner.¹³⁰ In the context of space deterrence, some authors have come up with the notion of layered deterrence, which includes a simultaneous combination of international norms, entanglement, retaliation, and denial of benefit which can be conducted across domains.¹³¹

4.2 Reinterpretation of Deterrence by Denial

In trying to come to terms with the nature of today’s strategic challenges, authors have also sought to expand on traditional concepts of deterrence by denial and punishment, even trying to merge the two into one mechanism. Recent years have seen the introduction of notions such as offensive denial and resilient denial, punishment through stigmatization, and entanglement, as well as the substitution of dissuasion for deterrence. With respect to deterrence by denial authors have introduced the distinction between tactical and strategic denial.¹³² Tactical denial refers to denying the adversary the prospect of attaining the direct impact of a particular hostile action, while strategic denial refers to denying it the political benefits that it expects to derive. While the former still aligns with traditional conceptions of deterrence by denial, the latter constitutes a significant broadening of the concept. Yet also tactical denial has been significantly expanded, most importantly by including offensive pre-emptive action. Traditional deterrence theoreticians assumed that denial is inherently tied to defensive measures, whether active or passive ones. The complexity and the opportunities presented by today’s strategic landscape domains have led them to theorise ways in which offensive action can be used to deny the adversaries the means to conduct offensive action. This, again, is discussed most often in relation to the cyber domain.¹³³ It is also observable in strategic practice, as the US has started to pursue its persistent engagement, which is it seeks to “defend forward” by preventively denying the adversaries their means for the conduct of hostile operations.¹³⁴ The underlying logic, however, as such holds for every other domain in which offensive means can degrade the adversaries’ capabilities to fight before the actual hostile interaction takes place. It is possible to conceive of denial in more traditional domains as encapsulating pre-emptive or preventive strikes against adversarial military capabilities.¹³⁵ Similarly, Israel has relied on a strategy of cumulative attrition in order to deter its enemies from carrying out immediate attacks by denying them the capability to do so.¹³⁶ Besides, capabilities in other domains tend to rely on cyber measures to varying degrees hence the use of offensive denial may impact land, naval, air and space domains as well. Overall, this approach recognizes the fact that the adversaries’ capabilities and their will to fight may be dependent upon each other and thus by denying the opportunity to use those capabilities is also likely to degrade their will to fight.

Where it comes to strategic denial, resilience is singled out as a key component.¹³⁷ Resilience is conceived as the ability to absorb the direct impact of the hostile activity in question without suffering any long-lasting impact. While originally proposed in the context of deterring terrorist attacks, recent scholarship proposes that resilience can be a strategic asset across multiple domains of competition and may be effective against both state and non-state actors.¹³⁸ Ultimately, strengthening resilience is envisaged as a cross domain effort because its objective is to prepare whole societies in a cross sectoral approach to withstand adversarial activities. Once attained, resilience then signals to the adversaries the futility of carrying out potential attacks by nullifying the potential benefits to be derived from a broad spectrum of hostile measures. To deal with the ever-increasing complexity of contemporary actors and domains, deterrence by denial has thus been conceptually stretched by including new approaches that include other types of effects.

4.3 Expansion of Deterrence by Punishment: Norms, Delegitimisation and Entanglement

Scholars have also proposed a broader gamut of measures encompassed under deterrence by punishment. The traditional concept is expanded to encompass deterrence through norms, delegitimisation and entanglement. Punishment through norms seeks to convince potential transgressors not to engage in certain behaviour by presenting them with the prospect of social costs.¹³⁹ It seeks to alter the cost calculus of those who do not abide by the positive standards of behaviour, while deterrence by taboos seeks to do the same to those who engage in hostilities that are generally seen as off-limits. Breaking any of these two standards risks incurring not only a domestic backlash but also the loss of international prestige and ostracisation which is detrimental to vital interests of both state and non-state actors. Deterrence by association expands on that logic. It constitutes “a political mechanism in order to ‘call-out’ poor behaviour and strongly condemn such actions publicly, by those with the right authority, because it acts as a clear signal to others in the community of actors what is right and wrong behaviour”.¹⁴⁰ This extended version of deterrence by punishment is increasingly being discussed in the context of deterrence in new domains and in relation to both state and non-state actors but is equally applicable to any other domain.¹⁴¹

A second alternative strategy, delegitimisation, is loosely based on the logic of punishment as it aims “to raise the costs of participating in terrorism by challenging the normative, religious, and socio-political rationales individuals rely upon when participating in violence”.¹⁴² Authors in this strand also argue that this approach allows for the classification of both particular instruments and particular targets as unacceptable. The traditional deterrence literature also addressed this, but it may be even more relevant in the cross-domain context, because the new context makes it possible to channel the conflict into more manageable domains. In some cases, such as with nuclear threats, the focus on the stigmatisation of particular weapons may be more effective. For instance, the stigmatisation of biological, chemical and nuclear weapons which has developed gradually during the last century, was closely connected to the destructive nature of these weapons.¹⁴³ This logic may be applicable to space deterrence too if it is accepted “that encouraging behavioural norms regarding the peaceful use of space—and thereby increasing the political stigma of using weapons in space—is desirable…” because “even relatively weak political stigmas can deter attacks in space for players with something to lose.”¹⁴⁴ It is plausible, for example, that attacks against satellites should be discouraged by the development of an appropriate normative framework.¹⁴⁵ In other cases, such as the cyber domain, targets rather than instruments may warrant more attention.¹⁴⁶ Deterrence through norms may thus adhere to the original logic of deterrence by violent punishment but certainly stretches its scope. It relies on a broader concept of punishment by including the social and psychological costs in order to deter actions from engaging in certain behaviour.

A third way in which traditional concepts of punishment are stretched revolves around entanglement.¹⁴⁷ Entanglement relies on fostering interdependence between actors and contributes to deterrence success by shaping the cost calculus of potential adversaries, as suggested by Joseph Nye. The assumption is that actors entangled in mutually dependency relationships will refrain from launching attacks because they themselves will incur costs too. It works by persuading potential adversaries that the continuation of the status quo is in their own interest, hence they should be reluctant to launch an attack in the first place.¹⁴⁸ The logic of entanglement, in the cyber domain and beyond, works by “adding more factors into the deterrence cost calculus— economic, political and diplomatic, for instance—then an adversary can be entangled…since they would have to suffer the consequences in other areas of their relations.”¹⁴⁹ Essentially, entanglement operates by “mutual establishment and recognition as well as perception management of benefits both in the present and over time”.¹⁵⁰ Or, to put it another way, entanglement works by persuading the relevant actors that they are “stakeholders in cyberspace” which should motivate them to exercise restraint in offensive behaviour.¹⁵¹ Due to their mutual interdependence, this kind of deterrence is most often discussed in the context of the overall Sino-American relationship.¹⁵² But that logic may also apply to the space domain because attacks against commercial satellites can impede international trade and finance.¹⁵³ Deterrence through norms and deterrence through entanglement are thus seen as necessary expansions in today’s globally connected world. The theoretical innovations offered expand the scope of deterrence by taking a more holistic view of the overall incentive structure of potential targets of deterrence and including less tangible factors such as identity and social belief systems into consideration as well as non-military dimensions to affect the cost-benefit calculus of potential adversaries.

4.4 From Deterrence to Dissuasion

Finally, authors argue that deterrence of contemporary threats requires expanding classical concepts of deterrence not just in terms of the ways and means but also in its very nature. Taking stock of the theoretical additions and innovations to address today’s challenges, they argue that our common understanding of deterrence needs to be reconceptualised or rather that fundamental features that were mentioned in the classic deterrence literature require much greater emphasis. They argue that deterrence will have to focus both on persuasion and dissuasion and include both positive and negative incentives in order to prevent adversaries from engaging in undesired behaviour. Dissuasion, for example, can be seen within a broader approach to deterrence as a form that includes both threats and inducements but also “reassurances and benefits that make a world without aggression more attractive”.¹⁵⁴ The advantage of dissuasion is that it can be pursued “through international institutions, treaties, economic sanctions, raising reputation costs, soft balancing, and diplomatic engagement”.¹⁵⁵ Dissuasion, a subset of what can be termed “compliance seeking efforts’’, is supposed to include not only negative but also positive measures and it can work both by increasing the attractiveness of particular options and by decreasing the desirability of others.¹⁵⁶ While these ruminations may seem to be a classic case of concept creep, it is worth noting that they can also be considered a rediscovery of insights already coined by classical deterrence theorists. After all, in the early 1960s Glenn Snyder defined deterrence as “the power to dissuade” which is done by “the implicit or explicit threat of applying some sanction if the forbidden act is performed, or by the promise of a reward if the act is not performed” so that it constitutes “a process of influencing the enemy’s intentions, whatever the circumstances”.¹⁵⁷

5 Conclusion

Strategic concepts emerge in particular strategic contexts to deal with specific challenges in a given period. Some strategic concepts wither away once the strategic environment evolves, others persist but are adapted. Our review of the CDD literature finds a thriving scholarly and professional debate about the use and utility of deterrence in the context of today’s cross domain challenges. Our review reveals significant continuities but also significant changes in the insights offered by the CDD literature compared to the preceding waves in the deterrence literature. Deterrence has been cross domain in character since its early beginnings, prompting some to pose the question whether CDD is nothing more than old wine being served in new bottles. Accordingly, Gartzke and Lindsay start their 2019 edited volume by asking “whether CDD provides any additional analytical traction beyond classical notions of deterrence…” because “deterrence in practice has always dealt with …different military services with different nuclear, conventional, and unconventional weapons, together with various diplomatic, economic, and cultural instruments of national power.”¹⁵⁸ The continuities with traditional deterrence literature are indeed considerable: traditional concepts of deterrence by punishment and denial are still part and parcel of the strategic lexicon; the literature keeps returning to favouring conditions of successful deterrence including the communication of credible threats of cost imposition which is rooted in robust capabilities and will. At the same time, there is certainly no stasis in the CDD literature. As demonstrated in the review offered in this chapter, significant developments can be found both in terms of practical application and theoretical innovation. This speaks to the idea that CDD is more than just old wine being served in new bottles. Overall, our review warrants three main conclusions.

First, authors have spent considerable effort on the practical application of key tenets of traditional deterrence theory in the context of contemporary strategic challenges. This has resulted in an assortment of innovative ideas predominantly focused on practical measures and opportunities to deal with challenges related to attribution, threat credibility and proportionality, signalling and escalation management.

Second, authors have also come up with a number of theoretical advancements. In addressing today’s strategic challenges, they have refined and expanded on traditional concepts of deterrence by stressing that successful deterrence should be envisaged as a continuous process, by usefully differentiating in deterrence by denial between tactical and strategic impacts, by adding resilience to the other side of the denial coin; by incorporating social costs in the deterrence by punishment equation; and by complementing the traditional dominant focus on negative payoff structures with attention to the role played by positive incentive structures.

Third, in light of these refinements and expansions of the concept deterrence, the question is warranted whether this enlightened notion of deterrence is still in fact about the act of deterring an opponent or whether it in effect constitutes a reconceptualisation of the essence of deterrence by making it about dissuading but also persuading instead of deterring. After all, this expanded concept of dissuasion implies a more diverse range of instruments, both military and non-military, which can be used both as a stick and a carrot, both to compel and to deter, both to persuade and to dissuade, which brings it back to the broader coercive diplomacy literature from which it originally emerged.

Our own assessment finally is that dissuasion, rather than being akin to deterrence, is more fitting as an overarching concept which encompasses the various means and ways by which one can dissuade the adversary to abstain from the action.¹⁵⁹ As such it includes both positive inducements and negative threats. Dissuasion can thus work as an umbrella term for deterrence by denial and punishment, norms, entanglement, resilience and assurance. Given the salience of the hostilities conducted below the legal thresholds of international law as well as the inability or reluctance of states to respond to varied intrusion across all domains. This broader concept of dissuasion may be more appropriate in the context of the strategic challenges in today’s world.