Introduction

Technical superiority is essential for successful military operations: "a small edge in performance can mean survival" (Alic et al. 1992). This is why the defense industry continues to propose increasingly high performance systems, and from the Manhattan Project to combat aircraft, passing through communication systems, it has significantly contributed to technical progress, especially after World War II.

Beyond the security aspect, contribution to technical progress is one of the arguments advanced by the industry to highlight the positive effect of arms expenditure. Indeed, due to tight budget constraints in developed countries and increasing costs of defense materials, the impact of defense on the overall economic performance of a country has come under scrutiny; the driving role played by defense technological innovation within national innovation systems seems to be an argument for maintaining this expenditure.

On the other hand, since the late 1980s, the technologically pioneering role attributed to the defense industry has been challenged; this marked the end of the spin-off paradigm (Alic et al. 1992). In pure economic terms, it was more difficult to justify military expenditure, and the relation between military and civilian domains appeared under a new light. Consequently, a long-term view was proposed of how military technological spin-offs to the civilian domain alternate with civilian technological absorptions in the military field (Dombrowski et al. 2002).

At this point, a duality emerged and captured the interest of the scientific community.The simplest definition of this concept is undoubtedly the one proposed by the French Ministry of Armed Forces, according to which it "must make possible military and civilian applications" (Ministre de la défense 2006). Nevertheless, this definition does not cover the full complexity of the concept of duality, which today retains several senses, none of which gathers consensus, both from academic and operational perspectives.

Upon its emergence in the 1980s, duality was presented (notably in the United States) as a means enabling civilian sectors to benefit from military Research and Development (R&D) expenditure (Quenzer 2001; Uzunidis and Bailly 2005). Duality is then to a certain extent an argument that goes against the existence of a crowding-out effect associated with defense expenditures compared to civilian expenditure in R&D.From then on, the relations between defense production and civilian production became a major field of analysis for defense economists, and duality a widely employed concept. It is the focus of many works (Gummett and Reppy 1988; Alic et al. 1992; Cowan and Foray 1995; Molas-Gallart 1997; Kulve and Smit 2003; Mérindol and Versailles 2010) and facilitates the understanding of connections between the Defense Industrial and Technological Base (DITB) and the rest of the economic sectors. The development of underlying principles of duality would be an opportunity to improve the economic and technological performance of military expenditure and justify its economic legitimacy. Indeed, by supporting the synergies between civilian and military innovation, duality is a means to reduce the cost of defense policy and improve the innovation capacity of a country.

Nevertheless, an opposing view on duality has progressively emerged and has taken a parallel development path. Its supporters perceive the rapprochement between defense innovation and civilian innovation as a risk of disseminating military technologies in general, and weaponry systems in particular (Alic 1994; Tucker 1994; Bonomo et al. 1998; Meier and Hunger 2014). According to this paradigm, on the one hand, duality weakens the capacities of States to control defense technology dissemination, making it easier for enemy or unallied powers to acquire it. On the other hand, military technologies are this way made available to non-State groups, which would then pose a new threat for the States. From this perspective, duality would lower the performance of military expenditure as a guarantee for peace and would pose a risk for global security and economic stability.

Besides these two macroeconomic approaches, there is a later microeconomic perspective on duality, which is seen as an opportunity for defense companies to diversify their activity. Although the aeronautics sector is a pioneer in this field, today almost no industrial sector involved in the military field is free from a dualization of the market, and duality is now key to the strategy of defense companies (Depeyre 2013; Mérindol and Versailles 2015a).

System integrators in particular are leading this rapprochement between civilian and defense fields (Prencipe 1997, 2000; Gholz 2002; Sapolsky 2003; Hobday et al. 2005; Lazaric et al. 2011). Given their specificity, they have to aggregate an increasing number of technologies that are not always exclusively owned by defense manufacturers (for example, semiconductors or telecommunications) and must be able to appropriate or "absorb" technologies that are nowadays not necessarily intended for military application. Conversely, while system integrator skills were originally developed within the defense industry, they are now widespread in many large civilian companies. Due to this competence, such manufacturers, particularly those with access to high technologies, can integrate in their production a broad technological spectrum, which partly originates in the military field. Therefore, due to technology transfers, companies in both defense and civilian sectors benefit from technical advances in various sectors.

From a broader perspective, this dualization can be interpreted as a rapprochement of civilian and military production systems (Guichard 2004a, 2004b; Guichard and Heisbourg 2004; Serfati 2005, 2008; Bellais and Guichard 2006). In 1995, the U.S. Congressional Office for Technological Assessment defined duality as a process through which the Defense Technology and Industrial Base (DTIB) and the broader Commercial Technology and Industrial Base (CTIB) merged into a single National Technology and Industrial Base (NTIB) (US Congress 1990). In its most integrated sense, duality is then defined as an organization aimed at joint defense-civilian technological and industrial production. In the absence of a border between defense technology and civilian technology (if it never existed), the two sectors have an opportunity to cooperate in the research and development of technologies in order to take maximum advantage of overall competences and knowledge previously divided between two environments.

According to this approach, situations such as civilian material being used in a military context, off-the-shelf purchases by the Defense Ministry or, conversely, a technology initially intended for defense being appropriated by an industry, no longer fall under the umbrella of duality. The latter is only defined in terms of commonality, synergies and technological coherence between technological systems and "meso-sectors", according to the approach proposed by Guichard (2004a, 2004b). The challenge is then to classify technologies in order to evaluate duality. If uses are no longer considered key factors for duality, then it is possible to reduce the bias of the analysis linked to fluctuations in the acquisition policies of Defense Ministries. Moreover, while uses are essential in assessing the criticality of a technology for defense operations, they provide no explanation for a potential technological transversality. How a technology is used gives no indication on its technological characteristics. In this case, an essential distinction lies at the basis of this analysis. The dual use of a technology (market-related duality) should be distinguished from dual innovation (production-related duality).

A second theme approached in addition to duality, and deriving from it, is that of technological innovation as such. When studying innovation, the definition proposed by the second edition of the Oslo Manual can be used, namely: "Technological product and process innovations (TPP) comprise implemented technologically new products and processes and significant technological improvements in products and processes. A TPP innovation has been implemented if it has been introduced on the market (product innovation) or used within a production process (process innovation)" (OECD 2005). By this definition, it is the very essence of innovation to provide companies with a competitive edge. This definition resumes the position supported by Porter (1985), who presents it as key to company competitiveness. Companies willing to maintain sustainable competiveness on a constantly evolving market must have innovation at the core of their strategies.

Moreover, companies are at the center of the innovation process: seizing technological opportunities is a first step that must be followed by protecting the advantage thus obtained, which is key to capitalizing on it (Teece 1986). A company can implement several protection regimes, with various performance levels in terms of degrees of appropriability (Dosi 1988). Six appropriation instruments are commonly identified (Levin et al. 1985): patents, secrecy, lead time, effects of the learning curve, duplication cost and time and the efforts involved in sales and high-quality services. While patents are acknowledged as an efficient product innovation appropriation mechanism, secrecy, lead time and the effects of the learning curve are considered as efficient for process innovation protection. The latter are...