How can systemic perspectives on defence capability development be strengthened?

ABSTRACT Over the last thirty years, suggestions for how to develop defence capability have developed rapidly. However, supporting theory and structured concept development lag behind. Despite this imbalance, countries need to continuously spend resources on defence development. This study identifies central challenges in relation to the scientific perspectives and approaches needed to support the development of defence capability. The results show that the support for developing interactions between technology and social components is especially weak and that relevant supporting theories and methods from related fields are not considered. This study also shows that it is important to be able to address these questions from various perspectives and not to be limited by a specific scientific tradition. Finally, this study also identifies a possible emerging cluster of reports on capability-related research that provide a base for a much-needed cross-disciplinary approach to the development of defence capability.


Introduction
Today, our societies experience technology development at a pace never seen before. This is a challenge, as rapid technology development places a strain on operational development (Finlan 2021) and causes a need and drive towards more holistic and systemic perspectives on military capability, a so-called "whole of society" perspective (De Spiegeleire 2011). Although there are approaches such as capability-based planning (De Spiegeleire 2011;Walker 2005), the ability to make informed decisions in regard to the development of defence capability is questioned, and decision-makers "tend to focus their attention on extrapolations of known technical capabilities, with little regard to associated social facets" (Adlakha-Hutcheon et al. 2018). This increasingly complex situation affects not only military organizations but also defence innovation and industry (Cederberg, Wallén Axehill, and Herzog 2019;Molas-Gallart 2010). We therefore find the question posed by Molas-Gallart (2010) to be more pressing than ever before: Do we have the knowledge, approaches and tools needed to strengthen development for defence in an ever more dynamic, interconnected and, thus, complex world?
Our aim here is to study existing scientific perspectives and approaches and their relation to the development of defence capability. Support for the development of defence capability has only received limited academic attention (Hannay and Gjørven 2021;NATO STO 2021). In this study, defence capability should be understood as a system that involves several components, such as organization structure, competence, processes, and several different technical subsystems, with defence purposes. To limit our scope, our main interest is support for defence idea generation and concept development, i.e. the initial steps in the development of defence capability. Furthermore, in this study, we view defence as an activity including both military defence and comprehensive national defence aimed at preparing for war, avoiding war, and fighting wars.
Given our assumed need for holistic and systemic perspectives, the research approach and discussion are based on a lens of general system approaches developed from the system perspectives that are captured in the System perspective section. This study also highlights the effects of the "whole of society" approach with three examples representing different settings for the development of defence capability in an everwidening context. The Results section presents an overview of research that may support the development of defence capability. Finally, the findings are discussed, and conclusions are presented.

System perspective
Today, system perspectives provide theoretical tools for addressing complexity in various fields. Systems theory is an integrated element in social science and engineering and therefore provides valuable tools for addressing topics in both areas.
The organizations that perform military tasks and operations are systems that need to have a balance among organization structure, procedures, and technical systems. Consequently, the systems of interest in this study are complex systems with substantial organizational and social components. We find that many of the important aspects are captured in several different parallel generic system perspectives, including the following: • Sociotechnical systems are widely used to describe complex systems that have interdependent parts and comprise separate but interdependent technical and social subsystems (Baxter and Sommerville 2011;Mumford 2006). • A System of systems is defined as a collaborative system of systems (Maier 1998), i.e. systems forming an "assemblage of components which individually may be regarded as systems, and which possesses two additional properties: [i] Operational Independence of the Components . . . [and] [ii] Managerial Independence of the Components" (Maier 1998). • Heterogeneous hybrid systems are understood as systems consisting of elements of various kinds, such as natural objects, technical artefacts, human actors and social entities such as organizations and the rules and laws governing their behaviour (Geels 2004;Kroes 2009). • Engineering systems are characterized by a high degree of technical complexity, social intricacy, and elaborate processes and are aimed at fulfilling important functions in society (de Weck, Roos, and Magee 2011).
According to Mumford (2006), sociotechnical methods and approaches such as these can contribute to a better understanding of how human, social and organizational factors affect how work is done and used. This counteracts technocentric approaches that do not sufficiently address the complex relationships among organizations, people and technology (Baxter and Sommerville 2011). However, care must be taken to address how sociotechnical understanding can contribute to improving systems (Baxter and Sommerville 2011;de Weck, Roos, and Magee 2011;Maier 1998).
The system perspectives above are parallel and cannot together create a coherent perspective or together contribute to a detailed single model. This conclusion is also supported by the fact that there are competing definitions of a sociotechnical system that are based on different perspectives and aims (Kramer and Moorkamp 2016).

A systems perspective of defence capability development
Military organizations are equipped and trained for a specific set of tasks in the event of a national emergency or war. The challenges they face today, more than ever, are characterized by uncertainty and risk, and the interactions with the external society are many and varying. Thus, it is important to acknowledge that society plays an integral role in defence (Jore 2019), often illustrated with the use of the term "whole of society" in relation to military capability (De Spiegeleire 2011) and comprehensive national defence (Berzina 2020).
Several concepts are concerned with understanding and structuring the ability to solve military tasks and operations. Examples include combat power (U.S. DoD 2020), fighting power (DCDC 2014), joint functions (U.S. DoD 2020), warfighting capability (Fitzsimmons 2007;U.S. DoD 2008), elements of combat power, DOTMLPF and similar concepts (Australian MoD 2014; U.S. DoD 2008; Yue and Henshaw 2009), and military power (Biddle 2010). These concepts are typically aimed at highlighting the importance of understanding military capability as a function of both physical and abstract components and therefore have connections to the concept of military capability (Andersson et al. 2015;De Spiegeleire 2011;Grabis, Zdravkovic, and Stirna 2018;Kuikka, Nikkarila, and Suojanen 2015;NATO Standarization Office 2019;Smith and Oosthuizen 2012). Analogously, in the area of crisis management, the implications of conceptualizing crisis management capability are also discussed (Cedergren and Tehler 2014;Lindbom 2020).
Although there are links to capability definitions, as shown above, research on systems for defence with a clearly stated sociotechnical approach is not common (Bakx and Nyce 2015). However, examples include command and control: a sociotechnical perspective of human factors in defence (Walker et al. 2008); the links between intelligence innovation in practice and system of systems dynamics (Svendsen 2015b); the need for future assessment described by Adlakha-Hutcheon et al. (2018) andNATO STO (2021); and the accountability of military AI discussed by Verdiesen, Santoni de Sio, and Dignum (2021). Therefore, given the lack of a common or specific system perspective in defence research, we use a lens developed from general system approaches to ensure that we capture the existing links, even when the links are to different or competing system approaches.
Previous work has shown that the development of new systems often focuses on the technical system and on objective requirements (de Weck, Roos, and Magee 2011). However, "the failure of large complex systems to meet their deadlines, costs, and stakeholder expectations are not, by and large, failures of technology" (Baxter and Sommerville 2011), and there are challenges to meeting high demands on aspects that are emergent system properties (Andersson 2018;La Porte 1996;Liwång 2020).
Therefore, based on how capability is understood here and using the abovementioned lens of general systems approaches, we argue that there is a need to consider social, political, sociopolitical, economic, and technical aspects when developing systems for defence. Figure 1 presents a proposed generic way of illustrating capability system development and interactions among different development measures, the existing capability (the system), and the surrounding society.
Our outlook is based on the idea that we are in the beginning of the development of a capability system, which is labelled "Now" in Figure 1, and we look at the foundations and conditions for performing the related actions and processes.
The processes included in system development in Figure 1 represent a society's ability to develop defence capability. Such an ability for systems development is central and has many parallels to the sectoral sociotechnical system of innovation described by Geels (2004). Based on the systems perspectives described above and our view on capability system development illustrated in Figure 1, we argue that knowledge in the following areas is needed to develop defence capability: • Area 1: Understanding of the Need for new capability, including links to Expected tasks and contexts, and of effects from change during development due to interactions with the existing Capability system under development and its interactions with Societies. • Area 2: Support for the technology development process. • Area 3: Support for the social development process. • Area 4: Support for the interaction development process.

Research approach
This study is performed in two steps. In the first step, scientific fields that are identified as possibly supporting capability development are qualitatively described. The fields described are as follows: • From the theoretical background: Generic system perspectives and Concepts for military capability. • From the description of related research: Traditional military system understanding, Traditional views on technology development, Understanding of technical knowledge, Design science, Safety as a holistic emergent system property, Constructive technology assessment, Human factors engineering, Organizational psychology, and Innovation systems.
The fields are selected to cover capability system development, as illustrated in Figure 1. Therefore, documents have been included if they relate to one of the described Areas 1-4 and provide a representative description of each field. The focus was to capture a sufficiently broad description of each field to provide an understanding of how each field could support capability development. The aim is not to represent a systematic review of the fields. The first step is finalized with a citation analysis investigating the connections between the described research approaches. In the citation analysis, the research is grouped based on the central theme of each source with the aim of identifying relationships within capability research and between capability research and other fields. The study can address capability development, as defined by Figure 1, and the relationship between capability development and the fields described. However, we cannot claim to provide insight into the different fields, as the study of each field is not sufficiently systematic. The strength of the results of the citation analysis is tested with a sensitivity analysis in which input data are systematically changed through addition and removal in a stochastic manner.
In the second step, each of the four areas that were previously identified as necessary for developing defence capability, Areas 1-4, serve as the outset and framework for a qualitative analysis. In this step, the identified and described scientific perspectives and approaches described in Step 1 are related to their respective possible role in supporting the development of defence capability.
In general, the understanding of what type of different considerations are needed in capability development is too narrow (De Spiegeleire 2011). Therefore, to avoid repeating this mistake, we have included three examples of defence capability development contexts that are presented in Examples of capability development contexts and used to exemplify the implications of the "whole of society" defence capability. The three examples represent military defence, civilian defence and societal expectations for defence. Military autonomous underwater vehicles (Example 1) are subsystems used today but are also an area characterized by fast technology development that promises to increase the effectiveness of military systems. Electric grids (Example 2) represent an area of societal need and a type of civilian defence system central to modern society but not developed for current or future defence needs. The relationship between defence development and societal values (Example 3) represents a societal perspective of the development of defence and capability and highlights internal as well as external, such as political and public, expectations for the development and configuration of future defence systems.
The three examples, Examples 1-3, are used in the second step to highlight activities and needs within Areas 1-4.

Examples of capability development contexts
Example 1: Military maritime expectations of a shift towards system-of-platforms solutions Today, military unmanned systems at sea perform various tasks at various levels of conflict in applications such as mine countermeasures, intelligence, surveillance and reconnaissance, and anti-submarine warfare, and the level of autonomy is increasing (Scharre 2018;Sparrow and Lucas 2016). Therefore, Maritime Autonomous Systems (MAS) can appear in a broad spectrum of scenarios, and military MAS are affected not only by conventional laws and regulations but also by the Law of Armed Conflict, so a broad palette of social and philosophical issues needs to be considered (Johansson 2018). However, knowledge regarding the actual operational circumstances and design of unmanned ships and vehicles is lacking (Wróbel, Montewka, and Kujala 2017).
Technology development in relation to autonomy, sensors, sensor fusion, batteries, etc., has created new possibilities for vehicles and vessels (Maguer et al. 2018). These technical possibilities coupled with effects such as fewer or no crew on board change aspects such as safety, risk, endurance, range, and size. Therefore, at sea, there are potential emergent system properties, such as operation in a remote, underwater, or hostile location. However, it is challenging to identify how the system of systems should be organized to reach its best potential and deliver a suitable capability, especially when the organization is accustomed to only extrapolating from known technology (Adlakha-Hutcheon et al. 2018).

Example 2: Emerging defence challenges for the electric grid
The systems that distribute electric power throughout our countries and continents are important for the economy, stability, and security. The infrastructure of a nation's electric grid is determined by decisions made long before today's societies and threats emerged. The vulnerability of United States electric grids to physical disruptions of the system providing electricity has long been recognized (U.S. Congress Office of Technology Assessment 1990). This vulnerability has increased in recent years because the infrastructure has not expanded in relation to a growing demand, which means that the system's redundancy has decreased (National Energy Policy Development Group 2001). Additionally, assessments indicate that the threat of human attacks has increased, and the US Committee on Science and Technology for Countering Terrorism states that "electric power systems must clearly be made more resilient to terrorist attack" (National Research Council 2002); examples of technical advances to this end include papers such as Salmeron et al (Salmeron, Wood, and Baldick 2004). In addition, research indicates that the development of a dependable grid today requires understanding the grid's dependency on its cyberinfrastructure and its ability to tolerate potential failures (Sun, Hahn, and Liu 2018). Scholars also note that the cyber-physical relationships within the smart grid need to be examined, and possible attacks need to be specifically reviewed to determine how the grids should be protected (Sridhar, Hahn, and Govindarasu 2012).

Example 3: Defence development and societal values
In technology development, it is suggested that there is a cocreation, or "intraaction," i.e. a process in which subject and object, matter and meaning are formed and transformed such that one cannot look at one component without addressing the other (Barad 2003). There is no one way of conceptualizing social values since different conceptualizations provide different perspectives. According to Kenter et al. (Kenter et al. 2019), this creates a need to address "complex, wicked problems, where facts are uncertain, stakes are high, and decisions are urgent." Furthermore, decisions about the development of new systems are inherently normative, and using different lenses results in competing ontological views (Kenter et al. 2019;Savaget et al. 2019).
When a norm critical lens is assumed to consider these aspects, the task becomes to study how values that are intrinsically embedded in society (Geels 2004) are applied to material artefacts and to identify the manner in which technical artefacts pose special risks in terms of reinforcing undemocratic social norms (Michelfelder, Wellner, and Wiltse 2017). These considerations are especially important in the area of defence, where these influences are often ignored. When technologies cocreate human actions, they give material answers to the ethical question of how to act (Verbeek 2006). Therefore, one could certainly make the argument that technology development is not only influenced by societal values but can also influence and change societal values in return, which stresses the importance of understanding the interactions between society and capability illustrated in Figure 1.

Description of related research
The development of today's defence systems understanding While operations research, which was developed during WWII to provide a "quantitative basis for decisions regarding operations under their control" (Morse and Kimball 1998), was able to support the use of technical systems, it became clear during the 1950s that scientific support was also needed to develop these systems. The RAND Corporation proposed systems analysis (Hoag 1956), which was then developed in parallel with major US development programs such as Polaris, Mercury and Gemini. Similar to systems analysis, systems engineering has its roots in systems theory and is "a way of looking at the 'big picture' (or entire life cycle) when making technical decisions" (Yang, Cormican, and Yu 2019). Systems engineering processes are commonly used today in the defence industry and in defence material acquisition (Andersson 2018). Systems engineering focuses on the life cycle of the technical system and qualitative requirements (Andersson 2018;Yang, Cormican, and Yu 2019). Today, there are also developed connections among systems engineering, system-of-systems and military intelligence innovation (Svendsen 2015a(Svendsen , 2015b. However, when analysts attempt to apply systems engineering and related concepts to an increasing number of areas in society, many problems involving human or social aspects are often laden with uncertainty, therefore increasing complexity, and could instead be best identified as ill formulated, "wicked," or "tangled" (Jackson 1995;Wognum et al. 2019;Yang, Cormican, and Yu 2019). Subsequently, soft systems methodology (SSM) was introduced and can be described as an approach that makes it possible for different stakeholders to reach an agreement on a commonly acceptable system description (Baxter and Sommerville 2011;Jackson 1995).
Concepts that are related to military capability include the concept of military utility, offering a capability-centric way to value a specific technology or technical artefact (Andersson et al. 2015) and approaches to defining a capability life cycle (Grabis, Zdravkovic, and Stirna 2018). The focus of these concepts is shifted from the system engineering perspective towards the capability system (Smith and Oosthuizen 2012). Related work is also performed under the term Concept Development and Experimentation (CD&E) (NATO 2021;Pikner 2015), which specifically addresses approaches for early phase system development.
In support of defence planning, there is a development related to scenario development and foresight aimed at reducing the uncertainty related to the future of warfare and adversaries (De Spiegeleire 2011; Garvey 2022). However, scenario development and foresight also need to be created with care to not introduce assumptions about the future that are too limited (De Spiegeleire 2011).

General research fields related to capability development
Parallel to the development of ever more complicated technology and technical systems, there is also an academic and philosophical debate on the true nature of technological knowledge. The debate focuses on whether to embrace the assumption that technologies are value neutral and directed towards utility (Kroes 2009;MacKenzie and Wajcman 1999;Spicer 2003;Verbeek 2006;Wajcman 2009). An alternative approach is that knowledge and knowledge claims are socially constructed and therefore intertwined with a variety of perspectives and motives (Åsberg and Lykke 2010;MacKenzie and Wajcman 1999;Pinch and Bijker 1984). This approach has strong links to the challenges presented in Example 3 but could also be useful for explaining the failure of today's systems to meet stakeholder expectations.
Work related to the nature of technological knowledge, or technoscience, investigates the cocreation of technology and social norms while also identifying the ways in which values are transmitted in the design processes of technological systems and solidified in artefacts, therefore enabling value judgements, possible biases and power relations that often remain undetectable (Haraway 2013;Leese 2017;Michelfelder, Wellner, and Wiltse 2017;Verbeek 2006;Wajcman 2009). Therefore, it is argued that science, as well as technology, is tainted with the concept of objectivity as an ideal that conceals its biased nature (Faulkner 2001;Haraway 1988;Wajcman 2000).
Constructive technology assessment provides a applied approach for addressing social expectations when users, or citizens, are invited to discuss plans for future systems. This creates a fuller picture of the system, brings underlying values to the surface and allows all parties to learn from each other (Lennerfors 2019;Rip 2018). Sociocracy (and the related term holacracy) is another approach that has been suggested for personnel involvement in innovation and strategic decisions; the purpose is to integrate multiple viewpoints representing different disciplines and various layers of an organization (Lekkerkerk 2016).
One possibly relevant, but generic, perspective of the development, or design, of defence capabilities is that of design science. Design science considers the design of artefacts, including both physical and nonphysical man-made systems (Cedergren and Tehler 2014;Hevner et al. 2004). The aim is to be able to design although there are "vague requirements, changing environments, shifting stakeholders interests, unclear problem situations etc" (Johannesson and Perjons 2014), which are created with an initial focus on functional purpose and abstract functions before physical form is determined (Winter 2008), i.e. an abstraction hierarchy (Rasmussen 1985). According to Hevner (Hevner 2007), design science is a pragmatic science in the sense that it emphasizes relevance. Therefore, design science offers an explicit focus on the problem, or function and purpose, of the capability system. However, design science research has shown that it is particularly challenging to perform multidisciplinary design (Winter 2008).
Potential support can also be found in the area of safety as an emergent sociotechnical system property (Rasmussen 1985) and in descriptions of the interconnectedness between safety and security (Bakx et al. 2016;Jore 2019), in which it is proposed that it could be fruitful to view security as an emergent system property. However, Kramer and Moorkamp (2016) stress that in areas such as safety and human factors, organizations are viewed as sociotechnical systems, but they are only loosely related to the sociotechnical research tradition and differ substantially in theoretical background. The safety field also includes principles for engineering safety that cover principles such as inherently safe design, safety reserves, safe fail systems, and procedural safeguards (Möller and Hansson 2008). Accordingly, there is a strong interconnectedness between safety and security (Bakx et al. 2016;Jore 2019). However, it is also stressed that approaches developed in the area of safety cannot be applied to security issues without proper testing (Jore 2019;Liwång, Ringsberg, and Norsell 2013).
Human factors engineering addresses the design of devices and systems for human use to create, for example, controllability (Lützhöft and Vu 2018). Often, human factors engineering is discussed in relation to the sharp end, i.e. the relation between technology and a specific human operator, and not in relation to the development of capability concepts. However, examples of the latter exist and include the controllability of military artificial intelligence (Verdiesen, Santoni de Sio, and Dignum 2021). The wider field of organizational psychology addresses interrelated organizational problems such as decision processes, personal and interpersonal relations, organizational structure, and interactions with technology (Chernyshenko and Stark 2004) with the goal to "look at all parts of organizations and to develop a systemic view" (Schein 2015). Phenomenon such as cyberspace and cybersecurity are examples of areas of inquiry and practice in which opportunities for organizational psychology are identified (Dreibelbis et al. 2018).
In systems of innovations, the systems that create artefacts constitute the scope of analysis (Geels 2004). Systems of innovation can be defined on several levels, such as the national, regional, or sectoral levels. Geels (2004) makes four additions to sectoral systems of innovation. The first addition explicitly incorporates the user perspective. The second addition is a more distinct definition of systems, the actors involved in them, and the institutions that guide actors' perceptions and activities. The third addition is to make institutions, such as professional societies, governmental agencies, independent research and coordination organizations, and public service organizations, an integral part of the analysis. The fourth addition addresses issues of change from one system to another to "analyse long-term dynamics, shifts from one sociotechnical system to another and the coevolution of technology and society" (Geels 2004). These additions are crucial for the possibility of applying the sectoral system of innovation in defence because of the integral and unique role of user institutions, such as an armed force, in defence development and the innovation system as a whole. Research also shows that in the area of defence, the sectoral system of innovation has seen major changes, as responsibilities have moved between different types of actors in different ways in different countries (Hartley and Belin 2019).

Citation analysis
To provide for a snapshot of the scientific fields discussed, this study also analyses the relations between the identified and described scientific fields. The citation analysis is performed on documents referenced in this study and presented graphically in Figure 2. A group of connected documents is here defined as more than two documents, each of which has at least two links with other members of the group.
In the citation graph, Biddle (2010), Fitzsimmons (2007), De Spiegeleire (2011), and Walker (2005) relate to capability and are loosely connected but not enough to be classified as a group. From the results of the citation analysis, two results stand out: • No strong connections between different studies within capability development were identified. • No strong connections between studies in capability development and the other studied research fields were identified. One exception is the study by Lindbom (2020) on crisis management capability embedded in Group B.
This lack of strong connections for studies related to defence capability exists in this material although the other areas, which have fewer studies, show connections both within fields and between fields. The identified lack of connections between theories and between theories and methodology for the development of defence capability highlights a need for more research, especially multidisciplinary research. A sensitivity analysis was performed to assess the strength of the results of the citation analysis in relation to the connection between studies in capability development. In the sensitivity analysis, the effect of removing any five documents from the original set and the effect of adding five not studied, but influential, documents 1 on defence capability to the original set are considered. There is no effect from the tested changes on the original set on the results identified above, and the analysis is therefore deemed robust enough to be used in this study.

Support for understanding the capability need (Area 1)
To a large extent, to understand this need is to understand the future and deal with substantial uncertainty. In describing the future through scenarios, there is a risk of creating too limited descriptions or point estimates (De Spiegeleire 2011) or building too much on known aspects of history (Garvey 2022). Creating foresight is therefore important and has been given some academic attention (Garvey 2022). Approaches such as the fighting power concept (DCDC 2014), the DOTMLPF construct (Australian MoD 2014; U.S. DoD 2008; Yue and Henshaw 2009), and the modern system view on military power (Biddle 2010) are typically applied with tight system boundaries that only consider military components and do not consider societal challenges, concerns or limitations. Additionally, since there is only a scattered understanding of what capability is, the possibility of describing the capability need in a coherent solution-independent manner is limited. Therefore, more development is needed to ensure that the approaches can be used constructively in the capability concept development phase to define and articulate the capabilities needed.
By avoiding an overly narrow system understanding and providing specific actions for creating knowledge about societal expectations, a fuller picture could be created, thus bringing underlying values to the surface. This could then assist in anticipating potential problems and expectations in relation to social sustainability and norm critical research (Example 3).
When major changes to the whole capability system cannot be excluded, such as in the concept development phase, continuous organizational changes, such as the sociotechnical systems engineering approach proposed by Baxter and Sommerville (2011), do not provide the support needed to identify changes building on a completely different system design. Especially when knowledge of the final properties of the system is limited, emergent properties must be treated dynamically throughout the development process affected by both organizational change and technical development. The possible increased effectiveness at sea when a system of different autonomous vehicles replaces systems centred around a crewed ship (Example 1), the vulnerability of electric grids (Example 2) when combining old and new components, or the effects on the interaction between civilian society and military organizations as a result of building on outdated norms (Example 3) are examples of such emergent system properties. It is possible that design science provides generic support for capability design that facilitates a system perspective (Cedergren and Tehler 2014).
Additionally, the need to apply several different disciplines in the development of defence capability is currently not effectively supported by academia. This conclusion is supported by a lack of multidisciplinary research and due to ineffective methods of communicating between different fields (Baxter and Sommerville 2011;Rasmussen 1997;Winter 2008). Many of the desired emergent system properties, such as the robustness of an electric grid (Example 2) and of a multiplatform concept (Example 1), reside between different disciplines.
In sum, the weakness of today's approaches primarily lies in the fact that they are more directed towards the extrapolation of today's systems or rest on a weak scientific background and lack support for multidisciplinary considerations.

Support for the technology development process (Area 2)
Understanding how the capability need interacts with technical components has strong ties to systems engineering and sociotechnical systems engineering. Suitable requirements for technical components cannot be defined as constants (Baxter and Sommerville 2011). If it is assumed that requirements can be defined independently, there is a risk of locking the design to unproductive aspects and therefore of affecting the capability of the system negatively. For autonomous maritime systems (Example 1), the requirements for the technical system components will be heavily dependent on the organizational design. Substantial complexity is also added when there are numerous legacy system components to consider, such as within electric grids (Example 2) or large military organizations.
Flexibility, modularity and versatility are system qualities that are needed to meet future challenges (de Weck, Roos, and Magee 2011). However, it remains unclear how much flexibility is needed, and additionally, how do we validate that the proposed solution meets this need? There is a balance to be met since flexibility always comes with a cost. If we fail to meet such challenges, there is a risk that complex products will become useless in a fast-changing world.
In sum, systems engineering, sociotechnical systems engineering, and organizational change activities may provide suitable activities and processes for understanding the role technical components play in creating the capability needed. However, it is important that the approach chosen for the development of defence capability allows for development under a high degree of complexity, uncertainty, and change.

Support for the social development process (Area 3)
System characteristics are heavily dependent on the system's social component characteristics, such as flexibility, power, deception, deterrence, risk, proportional measures, and ethics. Doctrine concepts with potentially non-straightforward (complex) interactions with organizational development include areas such as new competencies and the anticipated, but unrealized, properties of new technology in terms of unarticulated possibilities of autonomous systems (Example 1) and how doctrine expectations for the flexibility of the electric grid (Example 2) should manifest in organizational design.
Additionally, the very different scientific discourses between some "paradigms" of social science in relation to the role of technology in our societies (Aradau 2010) highlight the importance of understanding aspects that cannot be measured in relation to capability but that will affect the decisions made about capability and how it is perceived (Example 3). Organizational psychology and human factors can be used for organizational design. However, the focus is primarily on design in relation to the organization or the immediate technology, not on organizational development from the perspective of a capability need.
In sum, many perspectives describe social challenges to defence development. However, very few of these perspectives provide input on how to perform development.

Support for the interaction development process (Area 4)
Generally, many important system properties related to electric grids, MASs or meeting social expectations relate to emergent system properties originating from an increase in the number of interconnections between system components. Innovation is an important driver of change, yet exactly how innovation power should be harnessed is not given.
There is also a need for knowledge exchange between stakeholders, including potential future stakeholders and society at large, in accordance with Geels (2004) discussion of sociotechnical innovation systems. Innovation may need to align with changes in organizational structure, tactics, doctrine, and behaviour. This cannot be achieved without a suitable shared system understanding.
Other challenges unique to the field include the fact that defence systems cannot (or can only with extreme difficulty) be tested in the intended operational conditions and contexts and especially not in the degraded system states that often are a reality when solving defence.
The contribution of a specific technology to a system for defence may be the same for all nations on some measurable performance scale. However, social structure and context differ substantially between nations. Therefore, the actual value of the contribution of a specific technology to a defence capability is not the same over different contexts.
To understand foundations for the development of defence capability, it is not feasible or suitable to focus on physical systems (De Spiegeleire 2011). Other approaches need to be in focus. See, for example, the suggestions by Rasmussen (1985), the work in risk governance (Cedergren and Tehler 2014), and the developments achieved in relation to CD&E. However, research, especially in relation to capability concepts for defence, is limited.
In sum, the interactions between technology and social system components are complex and ever changing. It is unlikely that we will ever be able to meet this challenge by identifying a fixed definition of interfaces between system components. Rather, this is an area of important trade-offs that must be based on a foundation of knowledge shared among stakeholders on the challenges and weaknesses of the system. Such shared knowledge provides the only way for design decisions to be made continuously and creates a robust way of meeting the ever-changing and context-specific interactions between technical and social components. However, our study shows that the scientific foundation for this is weak and that the practice in systems engineering advocates for fixed and measurable requirements for technology development.

Discussion
The social construction of technology is a process that is subject to human decisions and influenced by the specific context in which they are situated. Not only acknowledging that these influences exist but also actively taking them into account positively affects technological innovation (Leese 2017). Especially with this in mind, this analysis shows that there is a shortage of scientific perspectives and approaches for understanding the development of defence capability and how different components of the sociotechnical capability system interact to create defence effects.
Numerous tools for addressing specific aspects of capability development exist. Many of these relate to the development of technical components; very few tools or approaches offer a link between different system aspects, and even fewer, if any, bridge different system aspects between different disciplines.
By examining the four areas of the analysis together, it can be identified that design science offers a relevant generic perspective of design based on a focus on functional purpose and abstract functions before physical form, and the capability concepts describe the structure of a capability. However, the link between design science and any of the capability perspectives is not articulated, and there is not a common systemic understanding among the capability perspectives.
There is a need for a theoretical as well as methodical foundation for the processes and interactions related to capability system development. To date, relevant supporting theories and methods from related fields have not been considered in full, and each discipline and approach only address a limited part of capability system development. Identified examples include the following: • Systems engineering provides an approach to the development of an identified measurable requirement in relation to a technical component and addresses the life cycle of that component. • Organizational psychology and human factors provide an understanding of and approaches for the development of organizations and human interfaces. • Critical technology studies describe societies' views on technology. However, the research does not specifically address development or how theories and knowledge can support design. • Capability models describe the links between capability and the technical and organizational components of a specific system (but not development). However, the models lack basic scientific support. • CD&E more specifically addresses the development of a capability system. However, CD&E also lacks basic scientific support.
The lack of connections between theories and between capability concepts, as shown in Figure 2, makes even the most basic application attempts risky and highlights a need for both more research and especially multidisciplinary and systematic applied work.
The three examples, the emerging capability promises of maritime unmanned systems (Example 1), the protection of grids for the distribution of electric power (Example 2), and the expectations from society regarding the ability for defence organizations to solve future tasks (Example 3), albeit representing different challenges, all would benefit from focusing on the development of emergent system properties such as flexibility and versatility and the interactions between technology and social aspects. In particular, there is no support for understanding and addressing the indirect effects of social change, as exemplified in Example 3.
Many of the identified challenges have already been highlighted in different studies, but what hinders the utilization of these lessons for effective development? While conducting this study, we noticed at least three main reasons: the fragmented landscape of capability research hinders coordinated efforts and even efforts that support each other, different studies approach these challenges from different disciplines and do not see a shared way ahead, and the lack of a common system understanding hinders the development of a common starting point. These reasons could explain the relative success seen in the area related to defence technology development, systems engineering, and system-of-systems, where the disciplinary starting points and system perspectives are well defined enough to foster a common development.
However, this study also identifies a possible emerging network of studies that collectively address military capability, systemic perspectives and other themes identified here. This network includes Biddle (2010), Fitzsimmons (2007), De Spiegeleire (2011), and Walker (2005. These studies may be a possible starting point for further in-depth development.
Engineers need to look beyond the idea that the importance of technology and technical solutions is defined by its measurable performance. However, social scientists need to meet them halfway and address the development process specifically. To not follow a specific scientific tradition and the ability to tackle these questions with several perspectives is not a flaw; it is a feature. Research that supports such a feature is too seldom performed and much needed.

Conclusions
For the development of defence capability, there is no one uniform suitable approach. There is a need to further develop an understanding of how to develop technical components under considerable uncertainty. This must be done while also addressing the development of social components, such as organization structure, doctrine, and tactics, that create defence development. However, support for the development of interactions between technology and social aspects specifically for the development of defence capability is lacking. The focus must be on emergent system properties such as flexibility and versatility of the future capability system.
In research on defence capability, relevant supporting theories and methods from related fields are not considered. Although the development of technical components is the area with the most developed approaches, engineers need to look beyond the idea that the importance of technology and technical solutions is defined by its measurable performance. However, social scientists need to meet them halfway by more specifically addressing how to create suitable development. The present study shows that it is important to be able to tackle these questions from several perspectives and not be limited by a specific scientific tradition. Therefore, not having a ready-made single disciplinary perspective of the development of defence capability is not a flaw; it is a feature. Efforts in which different scientific disciplines jointly approach complex questions are needed in the development of military capability.  (2020). Adding these five documents to the original set add a total of 251 documents to the citation analysis.

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

Notes on contributors
Hans Liwång is an Associate Professor at the Department of Systems Science for Defence and Security of the Swedish Defence University in Stockholm. Dr Liwång also is a Researcher at the Centre for Naval Architecture at KTH -The Royal Institute of Technology. Dr Liwång holds a PhD in Shipping and Marine Technology from Chalmers University of Technology and has been working with Defence and Security for more than twenty years.
LTC Kent E Andersson is a Senior Lecturer at the Department of Systems Science for Defence and Security of the Swedish Defence University in Stockholm. LTC/Dr Andersson also is an active officer in the Swedish Air Force. LTC/Dr Andersson holds a PhD in Military Sciences from the National Defence University of Finland and has more than ten years of experience as a lecturer in military-technology and more than twenty years of experience as an officer in the Swedish Armed Forces, mainly from working with development of command and control systems supporting tactical and joint command levels. CDR Therese Tärnholm is a PhD candidate at the Centre for Naval Architecture at KTH -The Royal Institute of Technology. CDR Tärnholm also is an active officer in the Swedish Navy and has more than twenty years of experience as an officer in the Swedish Armed Forces, mainly from working with development of naval and underwater systems.