Deterrence by delivery of arms: NATO and the war in Ukraine

ABSTRACT

During the course of the war in Ukraine, various actors have employed a unique type of deterrence by denial: namely, the threat to deliver arms. NATO leaders have committed—through rhetoric and deeds—to continue to deliver weapons to Ukraine if Russia escalated the war in order to deny Russian success. Not only is this type of strategy undertheorized, but it also challenges the distinction between direct and extended deterrence that has been central to deterrence scholarship. In deterrence by delivery of arms, the patron deters not by threatening to fight or deploy forces, but by committing to send weapons. However, the strategy also requires the protégé’s ability to fight. Studying deterrence by delivery of arms opens up understudied areas of deterrence (by denial), and provides a useful opportunity to consider how prominent concepts contribute to the research but also at times limit it.

KEYWORDS:

• Deterrence by denial
• extended deterrence
• the war in Ukraine
• NATO
• Russia

The war in Ukraine has called into question various prominent assumptions in international relations research—including in the research on deterrence. During the course of the war, one of the main strategies NATO members have executed is to deliver weapons to Ukraine. Among other things, this sends a deterrent message that Russia’s efforts to win are worthless given NATO’s commitment to assist Ukraine by providing it with the weapons it needs. As I show below, NATO leaders have framed this arms delivery to Ukraine as part of a deterrent strategy. This strategy, which I refer to as “deterrence by delivery of arms,” challenges two main, interrelated tendencies in the research of deterrence.

The first is the bias in traditional research toward the study and theorization of deterrence by punishment rather than deterrence by denial. Deterrence by punishment emphasizes the costs the putative challenger would face if it executed the undesired activity, whereas deterrence by denial aims to convince putative challengers that they will fail to achieve their goal (Snyder, 1961). I suggest that the more limited research on deterrence by denial (compared to deterrence by punishment) has created some blind spots with regard to identifying and theorizing about actors’ employment of this strategy.

In recent years, there has been a trend to push forward in developing, categorizing, and theorizing about the concept of deterrence by denial (e.g., Adamsky, 2021; Borghard & Lonergan, 2023; Gearson, 2012; Libicki, 2021; Lupovici, 2023; Smith & Talbot, 2008; Wilner, 2021; Zilincik & Sweijs, 2023). According to Wilner and Wenger (2021, pp. 4–5), this shift is due mainly to the transformation of the international security system and emerging regional and global security challenges. However, despite this growing interest in the study of deterrence by denial, this research still lags behind research on deterrence by punishment (Wilner & Wenger, 2021, pp. 2–7), and scholars have yet to fully acknowledge deterrence by denial through delivery of arms as a distinct category. While a number of scholars have started to consider these dynamics, their studies have been very brief (Lanoszka & Becker, 2023, p. 175) or done in the context of deterring terror attacks (Mezzell, 2019; Trager & Zagorcheva, 2005, p. 109) or as part of a policy analysis (Mitchell, 2015, pp. 124–125).¹ In other words, this sub-category is undertheorized and requires further elaboration.

The second deterrence tendency that “deterrence by delivery of arms” strategy challenges is the traditional distinction between direct and extended deterrence (Huth, 1988, p. 424). The former concerns the issuing of threats to defend the territories of the defending actor, while the latter concerns extending the defense to the territory of an ally. The war has shown that these two categories are not exclusive options but could be combined in different ways. The logic behind the strategy of deterrence by arms delivery undercuts the assumed dichotomy between direct and extended deterrence. On the one hand, the strategy relies on the capability of the protégé to fight; from this perspective, it overlaps with direct deterrence (by denial). On the other hand, because the capabilities are delivered by a patron (or a collective actor) to defend a territory that is not in its control, it overlaps with extended deterrence. Furthermore, these two elements are interrelated, as threatening delivery of arms is ineffective unless the protégé is able to use them.

The next sections answer three key questions. In the first, I address the question, What was the consensus in the academic literature on deterrence prior to the war in Ukraine? In this section, I elaborate on the traditional scholarship of deterrence as briefly noted above. The second section considers the question, What precisely did we get wrong about deterrence? In this section, I provide a brief empirical discussion of the threats issued, which I conceptualize as a new category of deterrence by denial, namely deterrence by delivery of arms. I explain why this strategy was adopted² and show how this type of deterrence challenges the traditional views of deterrence as presented in the first section. In the third section, I consider the question, Why were we wrong on deterrence? I assert that the prominence of traditional concepts and distinctions of deterrence scholarship limited researchers’ ability to conceptualize this type of deterrence. I conclude by considering the promising research agenda this new concept of deterrence offers.

Established understandings of deterrence

Two main distinctions in traditional deterrence scholarship provide the context of this article: the distinction between direct and extended deterrence, and the distinction between deterrence by punishment and deterrence by denial.

Direct versus extended deterrence

Scholars make a clear distinction between direct deterrence and extended deterrence. According to the common view, direct deterrence is a strategy that aims to dissuade threats to the defender’s (i.e., the deterrer’s) territory, while extended deterrence aims to extend the scope of defense to the territories of an ally. As Huth (1988), in a well-cited definition, suggests, “The objective of extended deterrence is to protect other countries and territories from attack, as opposed to preventing a direct attack on one’s own national territory” (p. 424). Extended deterrence can be exercised by individual states or by a collective of states, such as NATO (Morgan, 2003, pp. 172–202).

This distinction has been used in considering the difficulties of successfully practicing extended deterrence. Scholars widely accept that deterrence success is based on three main conditions: capabilities, credibility, and communication. Defending actors need to hold capabilities that allow them to employ a deterrent threat, they need credibility regarding their willingness to use these capabilities when needed, and they have to communicate their capabilities and credibility to the putative challenger (Morgan, 2003, pp. 15–20). By juxtaposing direct deterrence with extended deterrence, scholars can emphasize the imminent challenges of practicing extended deterrence. In the latter, the deterrent threat is not inherently credible as it is in situations of direct deterrence, where actors defend their own territory and citizens (Huth, 1988, p. 423). In situations of extended deterrence, it is more challenging for defenders (e.g., the patrons) to communicate and signal their resolve as part of establishing their credibility (e.g., Wilner, 2018, pp. 413–414). Furthermore, the defender actors may need to put their citizens and territory at risk if they defend a third party, especially when facing a powerful putative challenger (O'Neil, 2013, pp. 15–16). Contrasting direct and extended deterrence also shows that special capabilities are needed to execute extended deterrence strategy. Since conflicts may take place outside of the defender’s territory, actors need delivery means and systems to successfully practice this strategy (e.g., Mehta, 2021, p. 967). Conversely, if a patron poses a credible threat backed by appropriate capabilities, this may give the protégé incentive to exploit the situation and adopt an adventurous policy. In such situations, the patron needs to choose whether to be dragged into the conflict (Snyder, 1997, pp. 181–183) or to fail to maintain its commitment to the protégé and jeopardize its credibility (e.g., Cooley & Nexon, 2016, p. 102).

Punishment versus denial

The second distinction in this context is between deterrence by punishment and deterrence by denial. Deterrence by the threat of punishment is based on a threat of retaliation, while deterrence by denial is based on threatening that the putative challengers will not succeed in achieving their goals (Snyder, 1961). As Wilner and Wenger (2021) argue, “[W]hereas punishment deters through the fear of pain, denial deters through the fear of failure” (p. 7). A notable example of deterrence by punishment is nuclear deterrence, the study of which contributed to sharpening the notions of this strategy (Carnesale et al., 1983, p. 44) and became prominent in the research and its development (Morgan, 2021, p. 18; Wilner & Wenger, 2021, p. 3). In a similar way, most scholars studying extended deterrence also focus on extended deterrence by punishment (Thränert & Kartchner, 2015, p. 157; see for example, O'Neil, 2013).

However, scholars also study the multiple ways in which actors utilize the strategy of deterrence by denial (Lupovici, 2023, pp. 5, 9–14), including explorations of strategies of extended deterrence by denial (Inwook & Park, 2019, p. 177; von Hlatky, 2015, p. 4). Some of these strategies include deploying military forces that can be used as a tripwire (Blankenship & Lin-Greenberg, 2022); using military forces to tactically and doctrinally integrate with local forces (Mitchell, 2015, p. 125); threatening to send forces in the case of an escalation (e.g., Vanaga, 2019, p. 171); and deploying missile defense systems (Haffa, 2018, p. 105). Nonetheless, despite this growing research, the study of the strategy of deterrence by denial through delivery of arms remains undertheorized.

Some scholars have acknowledged that military support could serve to deter by denial: for example, in deterring terror attacks. Trager and Zagorcheva (2005), for instance, mention the ability of “a deterring state [to provide] economic and military aid to governments targeted by insurgents” (p. 109). Likewise, Mezzell (2019) critically analyzes the United States employment of security force assistance (SFA), which aims to support states that struggle with terrorism through “indirect deterrence.” Furthermore, Mitchell (2015) proposes not only to deploy military forces to help defending countries (thus enhancing their extended deterrence by denial), but also to supply them with defensive weapons. These deliveries would make “it harder for a stronger opponent to take territory by imposing costs on its forces during their initial attempts at incursion” (Mitchell, 2015, pp. 124–125). While I rely on these studies, they nonetheless reveal the need to further develop the concept of deterrence through delivery of arms and to acknowledge it as a sub-category of deterrence by denial.

Revelations from the war in Ukraine

The war in Ukraine and the strategy of deterrence by delivery of arms challenge two interrelated tendencies in the study of deterrence. The first involves extending the category of deterrence by denial to include “deterrence by delivery of arms.” The second involves challenging the sharp dichotomy between direct and extended deterrence.

Deterrence by delivery of arms as an extension of the concept of deterrence by denial

As the war in Ukraine has continued, NATO members’ practice of deterrence by delivery of arms has become more prominent. This strategy fulfills a number of goals simultaneously: It affirms a commitment to support Ukraine, avoids explicit direct involvement in the war,³ and issues deterrence threats. NATO leaders faced increasing pressure to assist Ukraine, but they also fear that becoming directly involved in the war would lead to a harsh Russian response.

Delivering weapons to Ukraine and promising to continue delivering them aimed to enhance Ukraine’s security and its ability to fight and defend itself. But it also had a clear deterrent rationale. The West’s commitment to assist Ukraine served as a message of deterrence by denial: That is, it signaled to Russia that, because of the West’s steady supply of military support to Ukraine, Russia would not be able to achieve its goals and therefore Russia’s escalation was useless. This logic has been implemented through an extensive and gradual delivery of arms, including tanks, combat vehicles, drones,⁴ artillery, launch rocket systems (MLRS), and air-defense systems, mainly by the United States, but also by Germany, the United Kingdom, and other European Union (EU) members. In addition, Poland and Slovakia have also delivered MiG-29 fighter jets to Ukraine (Axe, 2023). Experts estimated that between 24 January 2022 and 31 May 2023, military assistance to Ukraine was worth around 74.8 billion euros (Brown et al., 2023; Trebesch et al., 2023, pp. 35–41).⁵ Another crucial aspect of the military assistance has been the military training of Ukrainian forces personnel, including training in the United Kingdom with the assistance of a number of countries (Mills, 2023, pp. 22–23) and training in Germany by U.S. forces. In addition, Ukrainian forces received training for specific systems, such as defense systems (Lopez, 2023), and F-16 training for pilots, although such fighter jets have not been transferred to Ukraine (Pollard & Pawlak, 2023).⁶

Leaders framed these deliveries of arms in the context of deterrence by denial. For example, President Volodymyr Zelensky of Ukraine asserted in October 2022, “[T]he more weapons and ammunition we receive, in particular, artillery, drones, anti-aircraft defense, anti-tank and anti-ship weapons, the more tangible the responsibility for violation of international law will be for Russia and the less the aggressor will feel that he supposedly has room for escalation” (Zelensky, 2022a; see also Zelensky, 2022b). NATO leaders issued similar threats. For example, Federal Chancellor of Germany Olaf Scholz (2022) stated, “Putin will only seriously negotiate a peace when he realises that he cannot break Ukraine’s defences. That is why we are supporting Ukraine … The aim is rather to make it clear to Putin that there will be no victor’s peace.” Similarly, in August 2022 Estonian Foreign Minister Urmas Reinsalu explained the need to assist Ukraine, claiming that Putin would end the war only “because the price tag is going to be threatening his power mechanisms” (as cited in Miller & Ward, 2022).

U.S. officials were also prominent in connecting military assistance to deterrence by denial. For example, in July 2022, Jake Sullivan, U.S. National Security Advisor to President Biden, asserted that making Ukraine strong was a way to ensure that Russia’s invasion of Ukraine was a strategic failure for Putin, meaning that Putin should “be denied his objectives in Ukraine” (as cited in Goldberg, 2022, emphasis added). Furthermore, U.S. Secretary of Defense Lloyd J. Austin explicitly referred to the aim of military assistance to enhance Ukraine deterrence, stating, “We’re also looking ahead to provide Ukraine with the capabilities that it will need for deterrence and self-defense over the longer term” (Department of Defense, 2022). Similarly, U.S. Secretary of State Antony Blinken (2023) emphasized “building up [Ukraine’s] defense and deterrence.” In an interview with David Ignatius, Blinken made the connection between delivering military assistance and deterrence clear, stating, “The importance of maneuver weapons isn’t just to give Ukraine strength now to regain territory but as a deterrent against future Russian attacks.” According to Ignatius, this indicated a move toward a clear deterrence strategy based on giving “Ukraine the tools it needs to defend itself” (Ignatius, 2023). In this respect, delivery of arms enhances the capabilities of the “protégé” and thus contributes to its deterrence (see also Lanoszka & Becker, 2023, p. 177).

These deliveries of arms (and assistance in military training) clearly manifest the idea of intra-war deterrence. As Wilner (2013) explains, “intra-war deterrence suggests that even in the context of an ongoing conflict, adversaries can use threats to influence their opponent’s behavior and delimit the scope, nature, and ferocity of their military engagement” (p. 763). In this case, these practices of intra-war deterrence can be understood as intra-war deterrence by denial, aiming to convince Russia to avoid escalation and establishing future deterrence.⁷

These dynamics also demonstrate that the deliveries of arms and military training aim not just to compel Russia. While a strategy of deterrence intends to prevent an opponent from doing something, a strategy of compellence intends to coerce an opponent to do something (i.e., threatening to defeat Russia in order to coerce its withdrawal from Ukraine). While the deliveries of arms admittedly aimed to achieve compellence, they also aimed to establish deterrence (by denial). A number of scholars have suggested that an actor could apply a strategy of deterrence (i.e., intra-war deterrence) and a strategy of compellence simultaneously (Wilner, 2013, p. 745 ftn. 15). In addition, successful compellence could serve the strategic goal of future deterrence (Kirchofer, 2017, p. 25; Sperandei, 2006, p. 261; Tor, 2017, p. 95, ftn. 8). In this respect, potentially defeating Russia and forcing it to withdraw would have important implications for deterring it from attempting to attack Ukraine in the future as part of a logic of cumulative deterrence. Cumulative deterrence is based on repeated retaliations and military victories that aim to establish the defender’s deterrent posture and to communicate its resolve (Bar-Joseph, 1998, pp. 156–157; Stein, 1996). Scholars have even started to acknowledge the strategy of cumulative deterrence by denial (Kirchofer, 2017), achieved through defensive successes causing the challenger repeated failures (Sawyer, 2021, p. 111; Wilner, 2021, pp. 50–51).

Furthermore, the officials’ statements presented above also clearly demonstrate the connections between the war in Ukraine and deterrence (rather than just compellence). The statements show that the delivery of arms is itself the deterrent message to be communicated. Comparing this case to other cases of delivery of arms emphasizes its uniqueness. For example, in both the Iranian supply of arms to Hizbullah (Wilner, 2018, p. 417) and the delivery of arms through the US SFA (Mezzell, 2019, p. 137) intending to enhance deterrence, the delivery of arms was conducted covertly and therefore could not clearly communicate the deterrent threat.

As discussed above, scholars typically refer to a number of ways through which actors employ extended deterrence by denial. The most prominent is through the patron’s active engagement in sending, or threatening to send, military forces. However, for reasons discussed below, in the war in Ukraine the threat was the delivery of the weapons. This is a strategy of deterrence by denial, since, as a number of officials have stated, the idea was to make it clear to Russia that it would not succeed: Fighting would be worthless because NATO states would continue to arm (and to provide military training to) Ukraine. As a result, Russia could not win the war and, therefore, should stop fighting.

Acknowledging this type of deterrence and attempting to explain its adoption reveals interesting dynamics. On the one hand, a number of elements contributed to the need for NATO members to issue deterrent threats to defend Ukraine. Zelensky put pressure on NATO members to deliver weapons to deter Russia and deny its achievements (Zelensky, 2022a, 2022b). In addition, NATO is an organization constituted through deterrence identity and rituals, and it relies heavily on practices of deterrence to validate and justify the collective self (Mälksoo, 2021, pp. 70–71) and its member’s ontological security (see Lupovici, 2016). On the other hand, NATO members refrained from issuing direct deterrent threats to actively intervene and to send forces to defend Ukraine territory because of the fear of Russian retaliation (e.g., Putin, 2022). Western officials, such as German Chancellor Scholz, stated that NATO should avoid direct military confrontation with Russia, making it clear that such deterrent threats would not be issued (Arndt & Horovitz, 2022, p. 48). NATO members even refrained from making clear attempts to deter Russia from using nuclear weapons against Ukraine⁸ and instead issued a deterrent threat to prevent a situation where such an attack would affect NATO states (Arndt & Horovitz, 2022, pp. 30–31) or, more generally, where Russia would invade NATO territory (e.g., Biden, 2022). The deterrent threat to continue delivering weapons to Ukraine to deny Russia an achievement was a practical compromise. It allowed NATO to assist Ukraine and to narrate the discourse of deterrence, which is important to NATO members’ ontological security needs. But it also allowed NATO to avoid a direct confrontation with Russia following Russia’s deterrent threats.

Deterrence by delivery of arms and the interrelations between direct and extended deterrence

Along with reflecting a previously overlooked type of deterrence by denial, the case of the war in Ukraine also, and more importantly, challenges the distinction the scholarship prominently assumes between direct and extended deterrence. From this perspective, the deterrence being practiced in the war lies between direct deterrence and extended deterrence and is based on characteristics of both.

Foremost, the protégé’s ability to fight by itself enhances the effectiveness of the deterrent threats and its credibility. From this perspective, this type of deterrence by denial shares characteristics with direct deterrence. In fact, the massive military assistance to Ukraine ramped up only after the Ukrainians demonstrated their fighting capabilities (Dickinson, 2023; NATO Parliamentary Assembly, 2022, pp. 1, 19). One of the concerns had been that advanced weapons delivered to Ukraine would be captured by the Russians and thus end up in the hands of the putative challenger (e.g., Deen, 2022). But the strategy goes beyond direct deterrence since the capabilities also include the protégé’s ability to attain support from external actors (a patron or collective actors) that aim to extend the defense to the protégé’s territory. In this way, this strategy also overlaps with key aspects of extended deterrence. However, while traditionally these practices of extended deterrence do not exclude the option of the protégé actively taking part in defending its territory and deterring an attack, the patron bears most of the burden of using force if needed. For example, European defense during the Cold War was based on extended deterrence provided by “the most powerful NATO countries, but mainly via the United States,” although European partners were pushed to contribute to the defensive efforts (Zapfe & Vanaga, 2019, pp. 36–38). In contrast, deterrence by delivery of arms, as evident in the war in Ukraine, diverts the burden of using force to the protégé. The military training of Ukrainian forces also fits well with this rationale.

Limitations of deterrence scholarship

I suggest that a number of factors have limited scholars’ ability to identify this type of deterrence by denial or to challenge the dichotomy between direct and extended deterrence.

First, deterrence scholars’ have shown a strong bias toward studying deterrence by punishment—either direct or extended—and especially the conditions needed for this strategy to succeed (Stein & Levi, 2015, p. 411; Wilner & Wenger, 2021, pp. 2–7; Zilincik & Sweijs, 2023, p. 249). The limited research on deterrence by denial provided fewer incentives or tools for scholars to pursue research on this type of deterrence or identify how it could be exercised. Much effort has been made over the years to study deterrence by punishment. While the ways actors implement deterrence by punishment differ based on the means available to the defenders (e.g., conventional, cyber, or nuclear means), in all cases the strategy shares the same logic of threatening to do something that will inflict costs on the putative challenger. In other words, deterrence by punishment has clear conditions for deterrence success, even though it is often difficult to achieve it—as is evident in the debate over cyber deterrence (by punishment) success (e.g., Nye, 2017). In contrast, defenders can practice the strategy of deterrence by denial in various ways and based on different logics and means. Therefore, unlike with deterrence by punishment, there is no accepted theory on the conditions needed for deterrence by denial to succeed (Lupovici, 2023, p. 3; Wilner & Wenger, 2021, p. 7). Scholars, as noted above, have only recently started to acknowledge, theorize about, and categorize this strategy. From this point of view, identifying a strategy of deterrence by denial would have much more difficult before scholars started to theorize and categorize its various types. Furthermore, deterrence scholarship tended to focus on questions of the success of deterrence rather than on the adoption of the strategy (Lupovici, 2021). Since few theoretical and methodological tools were available to explore deterrence by denial, identifying new types of deterrence by denial did not serve the research goal of studying deterrence success.

Second, and following the previous point, the tendency to focus on deterrence by punishment also emphasized the importance of the distinction between direct and extended deterrence. This is especially prominent in the context of nuclear deterrence, as the ultimate deterrent threat could not be exercised by a protégé but rather only through the “umbrella” provided by a willing patron that could issue an extended deterrence threat (O'Neil, 2013). Once again, this distinction served scholarship’s interest in studying deterrence success but also helped to reproduce the distinction between direct and extended deterrence.

Third, while concepts and dichotomies are useful analytical tools, they nonetheless simplify reality and limit our ability to encompass complicated empirical situations and nuances (Gustavsson & Hallin, 2014, p. 570). This is a common challenge in research in general. However, the development of deterrence theory emphasized and reproduced these dichotomies, such as that between direct and extended deterrence. For example, many reviews of the scholarship have used this distinction as a key characteristic of deterrence scholarship (Danilovic, 2001, p. 99; Lupovici, 2010, p. 721; Mazarr, 2018, pp. 16–17). Nonetheless, this dichotomy also came to constitute how scholars think about the field. Not surprisingly, guidance on how to analyze empirical research asks scholars to specifically distinguish between these elements, suggesting that “any effort to generate a valid data set on deterrence must specify whether a case is an example of … extended or direct deterrence (or both)” (Harvey & James, 1992, p. 33). The point, however, is not only that deterrence can be both extended and direct (as during the Cold War), but also that these are interrelated in various ways, as the war in Ukraine demonstrates.

Conclusion

This article showed how the war in Ukraine challenges conventional understandings of deterrence theories. First, it diverts our attention to a new type of deterrence by denial: deterrence by delivery of arms. While this strategy overlaps with other types of deterrence, it is unique in that the threat is the ability of external actors to (continue) to support a protégé that is struggling with a putative challenger. More importantly, acknowledging this type of strategy challenges a key distinction in deterrence scholarship between direct and extended deterrence.

I suggest that acknowledging this type of deterrence and how it is used in the war in Ukraine offers a promising research agenda. First, it furthers recent attempts to theorize and map deterrence by denial, thus demonstrating the feasibility of and need to identify how this strategy can be adopted and the conditions under which it works. Some scholars have started to elaborate on conditions for success of deterrence by denial, albeit in different contexts (e.g., Gearson, 2012, p. 188; Libicki, 2021, p. 196, 198; Wilner, 2021, p. 54). But the research still lags behind the study of success of deterrence by punishment. One direction to advance the research of deterrence by denial is to acknowledge the different ways it is used. For example, as the war in Ukraine shows, the capabilities of both the patron and the collective actor delivering the weapons are important: that is, having sufficient and appropriate weapons to deliver and also a means of delivery. For this reason, in the war in Ukraine, access to the Polish border was fundamental (Hinshaw et al., 2022; Kranz, 2022). Likewise, the skills of the protégé are also crucial. If the protégé cannot use the supplied weapons, deterrence could not be effective or credible.

Second, challenging the clear distinction between direct and extended deterrence provides a useful point of departure for a broader research agenda. It is possible to examine and theorize the various ways actors are involved in practices of deterrence where the behavior of the protégé and the patron goes beyond what is expected in traditional models. In a similar way, scholars can contemplate additional cases of actors that are relying on strategies of deterrence by delivery of arms, which will help to justify the importance of this category and the need to reconsider the dichotomy between direct and extended deterrence. A well-known example is the military assistance of the United States to Israel during the 1973 War: the airlift “to Israel was a clear demonstration of US ability and determination to assist a threatened client … and for reasons of national prestige and reputation resolved to outdo the rival superpower’s performance” (Levey, 2008, p. 496). In other words, it aimed mainly to enhance U.S. direct and extended deterrence. This example differs from the case of the war in Ukraine, in which the delivery of arms aimed to dissuade challenges to the protégé rather than to the patron. Similarly, the relations between Iran and Hizbullah also demonstrate deterrence through delivery of arms that aim to enhance the deterrent posture of the patron rather than of the protégé. The cases also differ in that the case of Iran and Hizbullah concerns dynamics of deterrence by punishment and not of deterrence by denial. According to Wilner (2018, p. 421) Iran “designed a subtle deterrent threat against Israel that serves the interests of its militant proxies” by rearming them so they would be able to retaliate if Iran’s nuclear facilities were attacked (Wilner, 2018, p. 423; see also Akbarzadeh, 2016, p. 134). Nonetheless, these dynamics reflect the feasibility and need to further explore, trace, and theorize the practices of deterrence by delivery of arms and the possible interrelations between direct and extended deterrence.

Third, this strategy has some advantages from a policy-oriented perspective. For example, as discussed above, it allows a defending actor to be involved but to avoid a direct confrontation with the putative challenger. It also allows actors to tune the deterrent message by controlling the types, amount, and pace of arms deliveries or the arms threatened to be delivered. Thus, there are more ways to control the escalation and to issue measured responses. In this respect, gradual delivery of arms could demonstrate resolve and capabilities as part of cumulative deterrence. While scholars debate over the effectiveness of cumulative deterrence, challenging the ability of putative challengers to learn from previous rounds about the resolve of a defending actor (e.g., Lupovici, 2016, pp. 20–21; Tang, 2005), others suggest, nonetheless, that capabilities could be learned through their employment in previous rounds (e.g., Tang, 2005, pp. 39, 49). Deterrence through the delivery of arms could thus clearly signal the capabilities: the protégé’s ability to both acquire arms and effectively use them. What still remains as a challenge is the patrons’ resolve—whether they would continue delivering the arms in the future. Furthermore, deterrence by delivery of arms also helps to address the challenge of situations of a too-resolute extended deterrent threat. The fact that deterrence is achieved through a combination of the external actors and the protégé’s capabilities limits the protégé’s willingness to independently adopt an adventurous strategy.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Amir Lupovici

Amir Lupovici is a senior lecturer in the School of Political Science, Government and International Affairs at Tel Aviv University. His research interests include deterrence, cyber (security), securitization and ontological security. He is the author of The Power of Deterrence: Emotions, Identity and American and Israeli Wars of Resolve (Cambridge University Press, 2016). Lupovici's research has also been published in journals such as the European Journal of International Relations, International Studies Quarterly, Review of International Studies, Contemporary Security Policy, International Studies Review, International Studies Perspectives, and Foreign Policy Analysis.

Notes

1 It should also be noted that Wilner (2018) refers to a strategy of deterrence through delivery of arms, but as further discussed below, he focuses on mechanisms of deterrence by punishment.

2 Focusing at this stage on explaining the adoption of the strategy of deterrence in the war, rather than on its success, is also a methodological solution (Lupovici, 2021, p. 1683), given that its effectiveness is to be revealed in the future.

3 See also Lanoszka and Becker (2023, p. 175) for a similar argument.

4 On the importance of drones to Ukrainian military efforts, see Chávez and Swed (2023) and Kunertova (2023).

5 For updated information collected by Kiel Institute for World democracy, see https://www.ifw-kiel.de/publications/data-sets/ukraine-support-tracker-data-17410/

6 Scholars and officials have already acknowledged the contribution of military training and exercises for deterrence by denial, where “the military exercise is the message” (Clem, 2018, p. 132; see also Crawford, 2009, pp. 283–284; Mälksoo, 2021, p. 62, 69). For example, in NATO Summit of July 2016, “NATO agreed on an adaptive approach on deterrence, based on more military training and exercises” (Rîjnoveanu, 2019, p. 264). As such military training and exercises enhance the protégé’s capabilities, and demonstrate the commitment of the defending actors who supply the assistance.

7 I would like to thank two anonymous reviewers for pushing me to further elaborate on the concept of intra-war deterrence, and specifically how it is part of “intra-war deterrence by denial” and affects cumulative deterrence, as further discussed below.

8 For further elaboration on Russia’s threats to use nuclear weapons, see Arceneaux (2023).