Also available as a PDF:

Introduction

Although deterrence theory has demonstrated utility in maintaining limited peace between states, its applicability for preventing attacks from non-state actors is less clear. It is important to recognise that it is ‘practically impossible to prevent all terrorist attacks’ even with effective intelligence and defences (Elbahy, 2019, pg.46). This essay will instead argue that deterrence has significant relevance for the realistic goal of reducing kinetic, conventional attacks by violent non-state actors (VNSA) as part of a broader strategy to target the causes of violence. This broader strategy should include aid, economic development, and political conflict resolution to remove the root causes of VNSA. Deterrence should be part of strategies that decouple VNSA’s perceived link between terrorist means and achieving political ends. However, this essay will focus specifically on what relevance deterrence has to this broader counter-VNSA strategy and not on the broader political and economic strategy itself and the causes of VNSAs. 

Common use of deterrence is rife with imprecise language, such as referring to Trident submarine-launched missiles as ‘The Deterrent’, which confuses military instruments with their intended effects (Gray, 2010, pg.278). While there is no commonly agreed-upon academic definition of deterrence; nevertheless, a prominent definition is ‘the persuasion of one’s opponent that the costs and/or risks of a given course of action … overweigh its benefits’ (George and Smoke, 1974, pg.11). There is also an increasing blurring of deterrence and coercion in policy, where coercion is inflicting ‘pain for the purpose of dissuasion by the credible threat to inflict more pain’; however, this essay will not be engaging in this debate (Gray, 2010, pg.281). 

This essay will analyse deterrence theories’ assumptions and conditions before summarising the factors considered to have made it effective in supporting European peace and preventing certain types of attacks during the Cold War. This essay will not specifically cover VNSA’s acquisition of Weapons of Mass Destruction (WMDs), which may be the step-change in relative power that takes VNSAs from enduring criminal risk to an existential communal threat. However, it will explain why deterrence contributed to limited European peace and acknowledge the essential contribution of the nuclear dimension with WMDs. I will subsequently analyse the emergence of non-state deterrence and its utility in deterring different parts of VNSA networks. Then key challenges raised against VNSA deterrence will be assessed 3 to evaluate why deterrence theory has relevance in wider counter-terrorism strategies to limit VNSA attacks.

Deterrence theories’ assumptions and conditions

Deterrence theory has long been part of international politics; however, its importance increased significantly after recognising nuclear weapons as a new class of weapon (Elbahy, 2019, pg.44). This is due to their unparalleled destructive ability and consequences of use, difficulties of defending against them, and their role in strategic coercion through threatening civilian targets (Paul, 2009, pg.2). This made it much harder for actors to persuade themselves that a nuclear attack could be survived with a net strategic benefit compared to conventional threats (Gray, 2010, pg.279). Furthermore, nuclear weapons’ symbolic recognition by most actors as a different class of weapon changed the substantive effects of their use as they fulfilled a signalling function to other actors, and this symbolic recognition added to their destructive potential in altering actors’ cost-benefit calculations. 

During the Cold War, deterrence was the leading theoretical and policy framework in the bipolar international system. This theory postulated that adversaries could be deterred if actors can shape their perceptions of the costs or benefits of a course of action (Kroenig and Pavel, 2012, pg.22). This results in a cost-benefit analysis where actors rationally interpret deterrent threats by calculating their resolve to bear these costs against perceived benefits from acting

Deterrence theory is based upon several assumptions. First, actors are rational, conducting costbenefit decision-making calculations (Paul, 2009, pg.5). Second, for deterrence to succeed, deterrer’s must be able to understand their adversary sufficiently to target their high-value items (Elbahy, 2019, pg.44). Third, it must be possible to make adversaries genuinely believe that deterrer’s have the resolve to actually use their capabilities (Gray, 2010, pg.279). Fourth, deterrence becomes more effective if adversaries perceive the ‘red lines’ that will cause the deterrent threat’s implementation (Elbahy, 2019, pg.44). This assumed perception requires good communication amongst adversaries because there is no utility in surprising an adversary with unforeseen retaliation when signalling alone may deter this action (Wirtz, 2012, pg.9). This final assumption is contested because it may lead to ‘salami tactics’ where limited probes ‘achieve objectives over time without triggering a deterrent response’ (Wirtz, 2012, pg.9). Consequently, some state’s doctrine, such as France, was used to signal adversaries and 4 embraced ambiguity over defining their vital interests or ‘red lines’ to make adversaries more cautious. While politicians, especially in Israel, with clearly communicated thresholds often sought to add to their ‘reputational bank account’ for resolve, in preparation for convincing adversaries (Gray, 2010, pg.279).

Deterrence theory is based upon several assumptions. First, actors are rational, conducting costbenefit decision-making calculations (Paul, 2009, pg.5). Second, for deterrence to succeed, deterrer’s must be able to understand their adversary sufficiently to target their high-value items (Elbahy, 2019, pg.44). Third, it must be possible to make adversaries genuinely believe that deterrer’s have the resolve to actually use their capabilities (Gray, 2010, pg.279). Fourth, deterrence becomes more effective if adversaries perceive the ‘red lines’ that will cause the deterrent threat’s implementation (Elbahy, 2019, pg.44). This assumed perception requires good communication amongst adversaries because there is no utility in surprising an adversary with unforeseen retaliation when signalling alone may deter this action (Wirtz, 2012, pg.9). This final assumption is contested because it may lead to ‘salami tactics’ where limited probes ‘achieve objectives over time without triggering a deterrent response’ (Wirtz, 2012, pg.9). Consequently, some state’s doctrine, such as France, was used to signal adversaries and 4 embraced ambiguity over defining their vital interests or ‘red lines’ to make adversaries more cautious. While politicians, especially in Israel, with clearly communicated thresholds often sought to add to their ‘reputational bank account’ for resolve, in preparation for convincing adversaries (Gray, 2010, pg.279).

Given these assumptions, classic deterrence theory further requires several conditions to be successful; a deterrer should have sufficient capability to act (Paul, 2009, pg.2); the threat should be credible, with deterrer's possessing sufficient ‘resolve’ to act, or at least the appearance of this; the deterrer should be able to communicate with specific or generic opponents who must understand the explicit or implicit threat (Bowen, 2006, pg.59); and the adversary’s decision-making calculations must be influenced sufficiently not to act (Trager and Zagorcheva, 2006, pg.91). 

To achieve deterrence, actors can impose costs or deny benefits from acting. Deterrence by denial involves ‘hardening targets’ to make attacks appear too costly to attempt, seeking to convince adversaries that their resolve to bear costs will not overcome the imposed costs with a net benefit (Trager and Zagorcheva, 2006, pg.91). Whereas deterrence by punishment threatens ‘unacceptable costs’ in retaliation if adversaries act in certain ways, requiring sufficient understanding of adversaries and what they value to deter successfully (Kroenig and Pavel, 2012, pg.22). Both strategies intend to shift adversaries’ perceptions of their cost-benefit calculus not to act, with an assumption being that threats are identifiable to be deterred (Iverson, 2011, pg.2). Often actors use both denial and punishment strategies simultaneously. For example, during the Cold War, NATO’s strategy for defending Europe involved threatening Soviet high-value targets with nuclear retaliation if they attacked conventionally, while denying them an easy military victory if they attempted to invade with NATO’s conventional forces (Wilner, 2013, pg.746). 

Additionally, the ability to target an adversary’s high-value items was important for upholding deterrence, with two doctrines emerging. Counter-force doctrine was developed to limit a nuclear war by targeting military capabilities rather than civilian targets, however, it was associated with denial and a first strike to destroy an adversary’s capability (Nakao, 2021, pg.1- 2). Whereas, counter-value doctrine maintains deterrence by punishment through targeting an adversary’s high-value items with a secure second-strike capability that could survive a counter-force first strike (Nitze, 1976, pg.197). These are concepts are generally associated with deterrence between state actors and have less utility when applied to VNSAs.

How does deterrence theory explain ‘peace’ and the prevention of certain attacks during the Cold War

The Cold War is the most commonly used example of deterrence theory in action (as much as a highly complex, 40-year competition by competing international alliances can represent a single theory being ‘in action’), and it contributes to explaining European peace between states and the prevention of certain kinds of attacks. There are several definitions of peace, with negative peace being ‘the absence of violence and war’, while positive peace is ‘the integration of human society’ – this analysis will focus on the former definition (Martín, 2005, pg.46).

Cold War deterrence is not universally applicable even to state actors, as these were two nuclear-armed superpowers engaging in deterrence for decades. Therefore, similar approaches may not be effective against future adversaries (Payne, 2008, pg.4-5). Deterrence was significantly easier with this dyadic relationship between two main adversaries. The objective was to prevent nuclear conflict between known actors whose objectives and decision-making was generally understood (Elbahy, 2019, pg.45). Morgan argues that nuclear war’s absence suggests that deterrence was successful, even if it failed to prevent lesser conflicts between powers (Morgan, 2012, pg.86). Zagare, meanwhile, argued that nuclear war had such high assumed costs for both actors that actions which could have caused mutually assured destruction (MAD) became ‘irrational’ (Zagare, 2004, pg.112). This suggests Cold War deterrence succeeded because nuclear weapons risked irrationally high costs for both actors if they crossed communicated thresholds. 

However, this negative peace was limited geographically to Europe and preventing nuclear attacks. Direct or proxy conventional conflicts persisted outside of Europe in Korea, Vietnam, Angola and elsewhere, while espionage and subversion continued on the continent, demonstrating deterrence’s limitations (Krepon, 2003, pg.15-18). This is described as the stability-instability paradox, where Snyder identified links between ‘strategic nuclear stability and more conflict at lower, levels of escalation’ (Rauchhaus, 2009, pg.260). This suggests that deterrence with only conventional weapons will be partial, rather than absolute deterrence provided by nuclear weapons to ensure states’ survival. Long-term deterrence stability resulted from actors having secure, well-targeted and inherently credible threats because adversaries could incrementally improve their understanding of each other over time and avoid crossing thresholds when competing (Kilgour and Zagare, 1991, pg.326). Relationships like these may be more difficult to develop against non-state actors which lack the nuclear dimension of 6 absolute destructive threat. In summary, deterrence did not prevent all types of attacks during the Cold War. Nevertheless, its unique context allowed for capability, credibility, understanding and resolve to be developed and communicated over time. This allowed for long-term, generally stable deterrence that contributes to explaining European peace and the prevention of nuclear attacks. 

The emergence of non-state deterrence

While determining whether deterrence has failed is simple because the action the deterrer attempts to prevent occurs, it is harder to determine the extent of deterrence’s success. The USSR’s collapse allowed analysis of deterrence’s impacts on decision-making, but this empirical evidence is not available to actors currently engaged in deterrence (Bowen, 2006, pg.59). Although Elbahy argues that our understanding of deterrence theory from the Cold War needs updating, the fundamental concept of managing threats and controlling others’ behaviour without directly fighting them remains sound. The dyadic adversarial relationship has been replaced with multiple adversaries, different objectives, context, and capabilities, which deterrence theory must reflect (Elbahy, 2019, pg.45). This fourth wave of deterrence focuses on asymmetric threats and broadens the concept (Knopf, 2010, pg.1). Currently, it is hypothesised that VNSAs still require state or significant industrial support for creating potent, deliverable WMDs, but this will likely change in the future with technological proliferation (Kapur, 2009, pg.109-111). This will present new challenges for deterring WMDs; however, this essay will focus on deterring conventional, kinetic attacks by VNSAs.

Deterrence was a key pillar of Cold War national security strategies but is now a necessary but insufficient component of strategies to reduce terrorist attacks by disrupting and deterring all parts of VNSA networks (Kroenig and Pavel, 2012, pg.33). Deterring VNSAs will be partial rather than the Cold War’s absolute deterrence through MAD. States attempting to prevent all VNSA attacks regardless of scale set an unrealistic goal because, for operational and fiscal reasons, there is no perfect deterrent or defence against VNSA attacks (Iverson, 2011, pg.20). Scale is a challenge for denial strategies, as not all targets can be hardened sufficiently to deter attackers. Meanwhile, the intelligence resources needed to find and deter all VNSA networks through punishment make it unfeasible. Domestic populations need to accept the possibility of small-scale attacks while resources are expended to deter and prevent catastrophic attacks (Iverson, 2011, pg.20).

Deterrence works best when the defender is protecting vital interests because this increases coercive credibility. Thus, highly costly forms of interstate violence, like nuclear war, are easier to deter than lesser conventional attacks (Wilner, 2013, pg.772). Therefore, while states combat VNSAs to protect citizens, the balance of resolve is skewed as there is no existential threat to motivate states while VNSAs have relatively higher resolve. States also cannot threaten VNSAs with annihilation due to their dispersed characteristics, unlike the Cold War’s highlycredible retaliatory threats (Kroenig and Pavel, 2012, pg.24). This contextual change and asymmetry between actors has reduced deterrence’s credibility, with dimensions of asymmetry including; actor’s relative balance of resolve; divergent value-systems from political and cultural dissimilarities, causing divergent constraints on action; and the gradually-closing capability gap (Bowen, 2006, pg.55). Rather than communicating uncredible threats of annihilation, Israel demonstrates its resolve and capabilities to pursue long-run ‘cumulative deterrence’, convincing adversaries that destroying Israel is impossible or involves costs that exceed any expected benefits. This interpretation of deterrence changes notions of success from absolute deterrence of attacks to coercive success ‘on the fringes of adversary behaviour’ (Wilner, 2013, pg.747). Therefore, while deterrence is no longer as credible in isolation due to the lack of existential threat to states and VNSA’s characteristics, it is still important for reducing the number of attacks and contributing to a wider strategy. 

A significant difference in deterring VNSAs compared to the Cold War is the numerous, varied adversaries, each with independent decision-making. Furthermore, many VNSAs are composed of distinct factions with contrasting objectives, and in some cases, like Al-Shabaab, there is infighting over these objectives, creating greater uncertainty (Elbahy, 2019, pg.46). VNSAs can be broadly divided into two groups based on their goals. Groups with positive goals seek to ‘advance the welfare of an existing population or territory’, such as the IRA in Northern Ireland; whereas those with negative goals seek ‘maximal violence and destruction’, such as Al-Qaeda, and are perceived as harder to deter (Iverson, 2011, pg.11). While some groups possess objectives that fall into both categories, this distinction allows for different deterrent strategies to be applied. ‘Traditional’ terrorist groups like the IRA and ETA attack targets associated with their political objectives. These groups are often open to negotiation and inducements, and may self-deter from mass-casualty attacks. Therefore, deterrence is often a short-term delaying tactic to allow other approaches to work (Bowen, 2006, pg.63-64). Whereas global franchise VNSAs, like Al-Qaeda with its reach, fanaticism and ambitious 8 revisionist goals, are extremely uncommon, thus states should deter collaboration with these groups.

Most counterterrorism strategies focusing on addressing violence’s root causes are only effective in the long-term, while religious terrorism and suicide attacks have increased (Trager and Zagorcheva, 2006, pg.88). The Western response has been a rise in targeted killing of highvalue individuals to disrupt VNSAs at low risk to states (Crandall, 2013, pg.630-633). This potentially permanent policing operation lies below the threshold of conventional warfare to ‘mow the lawn’ by disrupting rising threats (Niva, 2013, pg.185). Targeted killing does not address the underlying social or political problems causing terrorism, and this offensive military force is expensive and should not be relied on as an effective long-term strategy. Therefore, without addressing terrorism’s root causes, deterring or eradicating individual VNSA groups will not eliminate terrorism due to the multiplicity of threats (Elbahy, 2019, pg.46-47). Instead, strategies to deter VNSAs should combine diplomatic, military, political and economic methods in a comprehensive strategy. When deterring multiple groups, states can use ‘broad deterrence’ against all groups with generic threats and low-cost actions such as blocking members’ assets and travel, while seeking to deter cooperation with groups viewed as the most dangerous by communicating they will focus resources on the most dangerous group. Regardless of how highly-motivated groups are, this disincentivises cooperation between groups and actors may self-deter to be perceived as less dangerous (Trager and Zagorcheva, 2006, pg.106-108). Strategies like this can decrease more dangerous VNSA’s capacity while preserving resources to use directly against non-deterred or more dangerous groups. Even if only temporary deterrence is achieved, this can allow states to use offensive strategies or long-term policies to degrade VNSAs (Trager and Zagorcheva, 2006, pg.122). 

How deterrence theory can be utilised against different parts of VNSA networks

Deconstructing VNSA networks enables tailored threats to be communicated to different actors through direct strategies against actors themselves or indirect strategies targeting what networks value (Kroenig and Pavel, 2012, pg.24-27). This requires a developed understanding of actors, which is resource-intensive. Therefore, generic deterrent threats should be issued, with resources reserved for attributing, targeting and punishing actors when deterrence fails (Trager and Zagorcheva, 2006, pg.108-109). Attacks should be viewed as a lengthy process 9 that can be deterred at every stage from target selection to media exploitation (Bowen, 2006, pg.54), with capabilities, motivation and opportunities to attack deterred at every stage (Iverson, 2011, pg.6). States must also demonstrate resolve to increase threats’ credibility and not capitulate to deter terrorist tactics. For example, the US gained a perceived reputation for capitulating by leaving Lebanon after Hezbollah’s 1983 Beirut barracks attack, and removing troops from Saudi Arabia in 2003 was perceived as a concession to Al-Qaeda (Trager and Zagorcheva, 2006, pg.103). The sections below will use examples to demonstrate how deterrence has relevance against all parts of VNSA networks.

Deterring States

Not all VNSAs require state support, but some act as proxies, require state-provided resources, or safe havens to operate. State sponsors are vulnerable to direct response strategies, such as President Bush’s threats after 9/11 and in 2005 that the US would not discriminate between state sponsors and VNSAs which led to actors reconsidering their relationships (Kroenig and Pavel, 2012, pg.27). Deterrence by denial assumes that the adversary’s capabilities are known, while technological proliferation makes the knowledge and capabilities needed to conduct attacks and develop WMDs increasingly anonymous and easier (Iverson, 2011, pg.4-5). Therefore, effective attribution methods would increase this threat’s credibility and deter states supporting VNSA and WMD development by holding any state sponsor accountable for VNSA conventional attacks or any VNSA WMD attack (Kroenig and Pavel, 2012, pg.27). 

Meanwhile, punishment-based coercion’s effectiveness in limiting existing state sponsorship depends on the proxy-patron relationship, including the affinity between their goals and level of dependency. If VNSA relationships are not essential to survival, states will be coerced, such as Pakistan’s President Musharraf ostensibly denouncing the Taliban post-9/11 because of US coercion (Wilner, 2013, pg.750-751). Furthermore, anonymity makes punishment harder; therefore, developing attribution capabilities for a wider range of conventional weapons will deter state supporters and deny VNSAs access (Iverson, 2011, pg.12). However, not all VNSAs can be deterred through direct response strategies targeting state support, such as those in lawless regions or controlling the host state’s decision-making apparatus to shape policy, such as Hezbollah’s power over Lebanese policy. Consequently, such exceptions require alternative deterrent methods (Wilner, 2013, pg.750-752).

Deterring Foot Soldiers

Foot soldiers within VNSA networks can be deterred by tactical denial strategies, like hardening targets, which are useful for actors who will not be deterred by retaliatory threats. For example, Foot soldiers are willing to bear significant costs because they highly value operational success. This makes denial strategies effective in deterring attacks and disincentivising recruitment if attacks are seen to cause limited damage or inglorious life imprisonment. With evidence from ideologically-motivated negative VNSA groups that suicide bombers were concerned their martyrdom would be jeopardised if they did not kill enough of their enemies (Kroenig and Pavel, 2012, pg.28). This could be enhanced by advertising the number of failed attacks and emergency response capabilities or communicating direct denial threats to shape adversaries’ perceptions of the probabilities of success (Kroenig and Pavel, 2012, pg.28-30). While protecting every conceivable target is impossible, it can shift attacker’s methods or targets. This is more effective with VNSAs who already demarcate between ‘hard and soft targets’; for example, some Kurdish VNSAs condemn attacks on civilians and focus on security forces (Wilner, 2013, pg.769). Security measures should be intended for deterrence rather than defence against ‘endless numbers’ of attacks by credibly raising the probability of failure, so attacks are not attempted (Kroenig and Pavel, 2012, pg.28- 30). Thus, deterrence has utility in limiting the number and types of attacks without the prohibitive costs of attempting a perfect defence.

Deterring Enabling Actors

Strategic denial can deter VNSA’s leaders by putting their political goals at risk. Groups with positive goals can have these known high-value items credibly threatened, such as Hamas’ locations in Gaza. Meanwhile, negative groups can have direct and indirect costs imposed on their members and goals, such as through Israel’s public-private partnership to limit terrorist attacks’ media coverage to deny VNSA’s benefits. Therefore, even if attacks are successful states will not comply, and this is only more effective against more committed VNSAs. This method seeks to break the link between terrorist means and sought political ends (Wilner, 2013, pg.753-758). For example, if the US can credibly convince Al-Qaeda that its military presence in the Middle East will continue regardless of continued attacks, it may deter Al-Qaeda’s leaders by disincentivising attacking Americans in favour or targeting ‘Apostate Regimes’ in the region (Kroenig and Pavel, 2012, pg.31-32). While implementing specific denial tactics, 11 like radiation detectors at borders, will deter certain types of attacks by risking strategically valuable nuclear material being confiscated to disincentivise leaders from certain courses of action (Kroenig and Pavel, 2012, pg.28). 

Meanwhile, if sufficient resources are committed, leaders and financiers can be found and deterred with direct retaliation as they are generally less willing to bear personal costs. Even political challenges to punishing financiers located in other states can be overcome if sufficient diplomatic resources are committed (Trager and Zagorcheva, 2006, pg.96). An example of direct retaliation is the UK’s 2006 ‘anti-glorification’ law threatening radical clerics with imprisonment, causing them to cease inciting or leave the UK. This demonstrates how appropriate punishment will deter different actors, although this example raises civil liberty concerns for liberal states (Kroenig and Pavel, 2012, pg.24). Finding enablers is also easier after an attack, so developing credibility for punishing these actors deters an essential part of VNSA networks. 

Deterring VNSA’s Supportive Communities

Punishing terrorism’s supportive communities can indirectly change VNSA behaviour from the bottom up. Israel widened the associated costs of VNSA involvement to include suicide bombers’ families through punitively demolishing their homes. This reduced suicide attacks by affecting attacker’s cost-benefit calculations but raises legal and moral concerns, limiting Western actors’ credibility for these threats (Wilner, 2013, pg.756-757). VNSAs driven more by ideology than political goals can still have these nonmaterial objectives denied by supporting opposing scholars arguing suicide contradicts religious teachings or publishing shameful VNSA actions. This deters actors by alienating them from supporting communities (Kroenig and Pavel, 2012, pg.32). This strategy of deterrence by delegitimization aims to decouple VNSA’s perceived link between using terrorist means and achieving their political ends (Levi and Stein, 2015, pg.411). As such, it is crucial to improve understanding of actors and their goals to allow for deterrent threats to be communicated effectively (Bowen, 2006, pg.59).

Challenges against applying deterrence theory to VNSAs

A key challenge is that some VNSAs are considered ‘irrational’ and therefore unresponsive to the cost-benefit calculations deterrence requires (Trager and Zagorcheva, 2006, pg.87). This strategic culture, where individuals are willing to sacrifice their lives for ideological reasons, often contrasts with Western individualised cultures, contributing to criticisms of irrationality. However, the applicability of irrationality to VNSAs is contested, with some literature concluding that groups such as Al-Qaeda conduct cost-benefit analysis and have valuemaximising behaviour. For example, using suicide attacks to generate mass casualties are signs of calculated decision-making to maximise an attack’s effects (Telleen, 2008, pg.69). This criticism is often linked with the idea that many terrorists are so highly-motivated and willing to bear suicidal costs that they will not be deterred by threats of punishment (Pape, 2005, pg.5). Thus, the exposure of VNSA groups to losses will be unlikely to prevent individuals from joining, provided that these losses are perceived to have furthered the group’s cause (Elbahy, 2019, pg.45). However, suicide terrorism is a ‘minor sliver of terrorist activity’, with many terrorist actors susceptible to lethal threats (Wilner, 2013, pg.755). Furthermore, foot soldiers are concerned not only with whether attacks succeed, but how successful they will be, with concerns that martyrdom could be jeopardised if they do not kill sufficient enemies. Therefore, foot soldiers’ high value of operational success means denial strategies can convince them operations will likely fail (Kroenig and Pavel, 2012, pg.28). Furthermore, highly-motivated actors will still be sensitive to their political goals being held at risk rather than their ‘life or liberty’ (Trager and Zagorcheva, 2006, pg.88). Therefore, through a combination of denying political goals, imposing costs by hardening targets, and targeted punishment threats against enabling actors, states can still deter attacks by VNSAs despite their high resolve.

Another challenge is that even if VNSAs were afraid of punishment, they lack a ‘return address’ for retaliation which constrains this, unlike states whose territory or population can be held ‘atrisk’ (Trager and Zagorcheva, 2006, pg.87). However, when states devote sufficient resources, they can find VNSAs’ members, objectives, and safe havens. Even if all group members cannot be found, states can almost always increase the group’s costs of achieving their goals. This may involve aiding their adversaries, imposing travel restrictions on members, or granting concessions to use as leverage to coerce VNSA’s behaviour. Despite concerns that granting concessions to VNSAs will encourage further attacks, deterrence relies on coercion rather than force to destroy a threat, and concessions with coercion allow for a deterrent equilibrium (Kroenig and Pavel, 2012, pg.31). Furthermore, if terrorists cannot be found, then the 13 indiscriminate use of force will often be counterproductive by alienating non-combatants and generating public sympathy; therefore, non-kinetic means such as deterrence should be used when possible (Rid, 2012, pg.140). Ultimately, only highly-motivated VNSAs must be found and threatened directly because generic threats will deter less motivated actors. This substantial motivation often makes these actors more susceptible to strategies targeting their political goals (Trager and Zagorcheva, 2006, pg.110).

The critique that deterrence by punishment and counterterrorism goals are incompatible has been described as the ‘defeat-deter paradox’ (Wilner, 2013, pg.742). This critique argues deterrence involves bargaining over adversaries’ behaviour, so those who are deterred need reassurances that they will not be punished if they comply (Jervis, 2009, pg.136). Firstly, deterrence can be a double-edged sword as it assumes a hostile relationship with an actor and its pre-emptive use could encourage an adversary to use force if the relationship was not previously adversarial (Telleen, 2008, pg.160-162). While secondly, an adversary that expects retaliation regardless of its actions will not be deterred because there is no incentive to comply. This critique treats VNSAs as homogenous units; however, punishment threats can be tailored and applied against different actors enabling terrorism but are not directly involved in fighting (Wilner, 2013, pg.743). 

Additionally, this critique does not account for ‘intra-war deterrence’, where tacit or explicit bargaining is used in an ongoing war when key thresholds have not already been crossed (Terrill, 2009, pg.4). This bargaining can occur over the war’s outcome or operational methods, with narrow deterrence of ‘particular types’ of operations within conflicts (Wilner, 2013, pg.762) allowing for limited conflicts to occur by deterring escalation, expansion or types of attack (Schelling, 1966, pg.135). For intra-war deterrence to succeed, deterrent threats must be conditional on behaviour, therefore actors must communicate credible guarantees that adversaries refraining from the specified activity will not be punished through ‘reassurance policies to complement deterrence’ (Kroenig and Pavel, 2012, pg.27). Furthermore, immediate deterrence can be used during an active conflict, and occurs when general deterrence fails, and additional denial threats are needed to deter a planned attack by increasing the risks of failure to change attackers’ cost-benefit calculations (Wilner, 2013, pg.762). Therefore, the defeatdeter paradox requires a better understanding of VNSA networks, target selection, and decision-making. This allows focused denial strategies and coercion through communicating threats to enabling actors concurrently with military operations targeting specific actors.

Conclusion

Deterrence was a vital component of understanding European peace during the Cold War and the limitations in geographic scope and types of attacks between actors. However, deterring VNSAs requires different methods and approaches and currently lacks substantial empirical evidence. Despite large numbers of diverse VNSAs with differing objectives, the underlying conditions of deterrence remain applicable. This essay has demonstrated that deterring all VNSA attacks is an unrealistic goal, but deterrent strategies have relevance in limiting the number and types of attacks.

The importance of intelligence to improve our understanding of adversaries cannot be overstated to get inside the ‘black box’ of adversary decision-making (Bowen, 2006, pg.59). This enables states to deconstruct VNSA networks, allowing tailored deterrence against different actors regardless of their resolve or anonymity. Generic denial tactics, such as domestic security measures, should be coupled with strategic denial methods against VNSA’s political goals and improved attribution against state and individual supporters. Meanwhile, communicated deterrent messages should include the intent to focus resources on finding and punishing actors who are not deterred or are perceived to be the most dangerous.

rrence against different actors regardless of their resolve or anonymity. Generic denial tactics, such as domestic security measures, should be coupled with strategic denial methods against VNSA’s political goals and improved attribution against state and individual supporters. Meanwhile, communicated deterrent messages should include the intent to focus resources on finding and punishing actors who are not deterred or are perceived to be the most dangerous. When deterring specific VNSAs, low-cost measures such as travel and financial restrictions can be applied to all their members, while specific threats can be targeted against strategic goals, operational success, or enabling actors to influence cost-benefit calculations. Deterrence can limit state support and prevent VNSA collaboration, reducing the spectre of WMD acquisition. Furthermore, by limiting certain types of attacks, deterrence contributes to national security goals while gaining time for long-term methods to be applied against different VNSAs. This essay has also demonstrated that common criticisms against applying deterrence against VNSAs are either unfounded or can be overcome with sufficient resources.

Comprehensive counterterrorism strategies should aim to disrupt attacks, degrade VNSA networks and counter ideological support, and deterrence has an important role in each of these areas (Kroenig and Pavel, 2012, pg.24). While new technologies, globalisation, growing anonymity and increasing multipolarity in the international system will present new challenges, at present, deterrence can be applied to VNSAs to influence their decision-making. Therefore, deterrence has significant relevance for limiting the number and types of conventional, kinetic attacks by VNSAs as part of a broader strategy.

Bibliography

Bowen, W., 2006. Deterrence and asymmetry: non-state actors and mass casualty terrorism. Contemporary Security Policy, [online] 25(1), pp.54-70. Available at: https://doi.org/10.1080/1352326042000290506 [Accessed 5 December 2021].

Crandall, C., 2013. If You Can't Beat Them, Kill Them: Complex Adaptive Systems Theory and the Rise in Targeted Killing. Law Journal Online, Seton Hall Literature Review, 43 (2). pp.595-642. Available at: https://heinonline.org/HOL/Page?collection=journals&handle=hein.journals/shlr43&id=608 &men_tab=srchresults [Accessed 15 December 2021]

Elbahy, R., 2019. Deterring violent non-state actors: dilemmas and implications. Journal of Humanities and Applied Social Sciences, [online] 1(1), pp.1-12. Available at: https://www.emerald.com/insight/content/doi/10.1108/JHASS-05-2019-003/full/pdf [Accessed 26 November 2021].

George, A. and Smoke, R., 1974. Deterrence In American Foreign Policy: Theory and Practice.. [ebook] New York: Columbia University Press, pp.8-12. Available at: https://hdl.handle.net/2027/heb.00720 [Accessed 10 December 2021]. 

Gray, C., 2010. Gaining Compliance: The Theory of Deterrence and its Modern Application. Comparative Strategy, [online] 29(3), pp.278-283. Available at: https://doi.org/10.1080/01495933.2010.492198 [Accessed 27 December 2021]. 

Iverson, D., 2011. The Ring of Gyges: Anonymity and Technological Advance's Effect on the Deterrence of Non-State Actors in 2035. [online] Maxwell: Defence Technical Information Centre, pp.1-21. Available at: https://apps.dtic.mil/sti/citations/ADA565069 [Accessed 5 December 2021]

Jervis, R., 2009. Deterrence, Rogue States and US Policy. In: T. Paul, P. Morgan and J. Wirtz, ed., Complex Deterrence: Strategy in a Global Age. [online] Chicago: University of Chicago Press, pp.132-148. Available at: https://www.google.co.uk/books/edition/Complex_Deterrence/oZOsiAlOEiQC?hl=en&gbpv =0 [Accessed 30 December 2021]. 

Kapur, S., 2009. Deterring Nuclear Terrorists. In: P. Morgen, T. Paul and J. Wirtz, ed., Complex Deterrence: Strategy in the Global Age, 1st ed. [online] Chicago: University of 16 Chicago, pp.109-125. Available at: https://books.google.co.uk/books?id=oZOsiAlOEiQC&printsec=frontcover#v=onepage&q&f =false [Accessed 29 December 2021].

Kilgour, M. and Zagare, F., 1991. Credibility, Uncertainty and Deterrence. American Journal of Political Science, [online] 35(2), pp.305-334. Available at: http://www.acsu.buffalo.edu/~fczagare/Articles/Credibility,%20Uncertainty,%20and%20Det errence.pdf [Accessed 24 December 2021].

Knopf, J., 2010. The Fourth Wave in Deterrence Research. Contemporary Security Policy, [online] 31(1), pp.1-33. Available at: https://doi.org/10.1080/13523261003640819 [Accessed 11 January 2022]. 

Krepon, M., 2003. The Stability-Instability Paradox: Misperception, and Escalation Control in South Asia. Asia Policy Papers. [online] The Henry Stimson Centre, pp.15-18. Available at: https://www.stimson.org/2005/stability-instability-paradox-south-asia/ [Accessed 24 January 2022].

Kroenig, M. and Pavel, B., 2012. How to Deter Terrorism. The Washington Quarterly, [online] 35(2), pp.21-36. Available at: https://www.tandfonline.com/doi/full/10.1080/0163660X.2012.665339 [Accessed 20 December 2021].

Levi, R. and Stein, J., 2015. The Social Psychology of Denial: Deterring Terrorism. New York University Journal of International Law and Politics, [online] 47(2), pp.409-438. Available at: https://heinonline.org/HOL/P?h=hein.journals/nyuilp47&i=417 [Accessed 11 January 2022].

Martín, F., 2005. Critical Analysis of the Concept of Peace in International Relations. Peace Research, [online] 37(2), pp.45-59. Available at: https://www.jstor.org/stable/23607773 [Accessed 24 January 2022].

Morgan, P., 2012. The State of Deterrence in International Politics Today. Contemporary Security Policy, [online] 33(1), pp.85-107. Available at: https://doi.org/10.1080/13523260.2012.659589 [Accessed 28 December 2021].

Nakao, K., 2021. Denial and punishment in war. Journal of Peace Research, [online] pp.1-14. Available at: https://doi.org/10.1177%2F00223433211009765 [Accessed 28 December 2021]. 

Nitze, P., 1976. Deterring Our Deterrent. Foreign Policy, [online] (25), pp.195-210. Available at: https://doi.org/10.2307/1148029 [Accessed 28 December 2021].

Niva, S., 2013. Disappearing violence: JSOC and the Pentagon’s new cartography of networked warfare. Security Dialogue, [online] 44(3), pp.185-202. Available at: https://www.jstor.org/stable/26302245 [Accessed 28 December 2021]. 

Pape, R., 2005. Dying to Win: The Strategic Logic of Suicide Terrorism. New York: Random House, pp.4-9.

Paul, T., 2009. Complex Deterrence: An Introduction. In: P. Morgen, T. Paul and J. Wirtz, ed., Complex Deterrence: Strategy in the Global Age, 1st ed. [online] Chicago: University of Chicago, pp.1-16. Available at: https://books.google.co.uk/books?id=oZOsiAlOEiQC&printsec=frontcover#v=onepage&q&f =false [Accessed 10 December 2021].

Payne, K., 2008. Deterrence and Coercion of Non-State Actors: Analysis of Case Studies. [online] Fairfax: National Institute for Public Policy, pp.1-43. Available at: https://nipp.org/wp-content/uploads/2021/05/Summary-Report-of-Deterrence-and-Non-StateActors.pdf [Accessed 5 December 2021].

Rauchhaus, R., 2009. Evaluating the Nuclear Peace Hypothesis. Journal of Conflict Resolution, [online] 53(2), pp.258-277. Available at: https://journals.sagepub.com/doi/10.1177/0022002708330387 [Accessed 24 January 2022].

Rid, T., 2012. Deterrence beyond the State: The Israeli Experience. Contemporary Security Policy, [online] 33(1), pp.124-147. Available at: https://doi.org/10.1080/13523260.2012.659593 [Accessed 10 January 2022].

Schelling, T., 1966. Arms and Influence. [ebook] London: Yale University Press, pp.126- 139. Available at: https://www.jstor.org/stable/j.ctt5vm52s.7 [Accessed 30 December 2021].

Telleen, P., 2008. Deterrence and Nuclear Terrorism. American University Research Papers. [online] Washington D.C.: American University, pp.54-76. Available at: https://www.proquest.com/docview/304684671/FED2FE0ACCEB4776PQ/1?accountid=107 92 [Accessed 2 January 2022].

Terrill, W., 2009. Escalation and Intrawar Deterrence During Limited Wars in the Middle East. [online] US Army War College Press, pp.1-22. Available at: https://press.armywarcollege.edu/monographs/622/ [Accessed 30 December 2021]. 

Trager, R. and Zagorcheva, D., 2006. Deterring Terrorism: It Can Be Done. International Security, [online] 30(3), pp.87-123. Available at: https://www.jstor.org/stable/4137488 [Accessed 20 December 2021].

Wilner, A., 2013. Fencing in Warfare: Threats, Punishment, and Intra-war Deterrence in Counterterrorism. Security Studies, [online] 22(4), pp.740-772. Available at: https://www.tandfonline.com/doi/abs/10.1080/09636412.2013.844524 [Accessed 27 December 2021].

Wilner, A., 2015. Contemporary Deterrence Theory and Counterterrorism: A Bridge Too Far. New York University Journal of International Law and Politics, [online] 47(2), pp.439-462. Available at: https://heinonline.org/HOL/P?h=hein.journals/nyuilp47&i=450 [Accessed 28 December 2021]. 

Wirtz, J., 2012. Deterring the Weak: Problems and Prospects. Proliferation Papers. [online] Paris: IFRI Security Studies Centre, pp.7-28. Available at: https://www.ifri.org/sites/default/files/atoms/files/pp43wirtz.pdf [Accessed 27 December 2021].

Zagare, F., 2004. Reconciling Rationality with Deterrence. Journal of Theoretical Politics, [online] 16(2), pp.107-141. Available at: https://doi.org/10.1177/0951629804041117 [Accessed 28 December 2021].

About the Author

Charlie Bradbury is a public sector strategy consultant at MH&A and an Army Reserve officer. He is a governing member of the Fightclub UK wargaming experimentation group and teaches at Lancaster University's Security Institute. He gained a First in BA (Hons) Politics and International Relations at Lancaster University and a Distinction on the challenging MA Applied Security Strategy at Exeter University.