Article body

In a context of increasing control in public expenditure, and facing new demands for services from citizens and new challenges, the government is looking for solutions to continue to deliver public services while reducing its costs. This evolution is based on the emergence of several problems. First, the government no longer has the financial and human means to deliver high quality public services. The gradual development of new public services, followed by the economic liberalization of the 1990s and the construction of Europe, have amplified the tendency to no longer automatically ensure public service activities through public entities but to entrust them to private companies. Recently, the government has also been seeking to transfer some of its ill-administered missions to the social economy (Harrisson & Boucher, 2011). Another trend comes from the third sector itself, which develops networks in order to modernize a deficient state and to change modes of governance.

Second, the public services organization is not (or not any more) adapted to the complex and multifaceted nature of certain social issues. These social problems are “wicked” social problems that cannot be satisfactorily solved by the activity of a single actor (Sørensen & Torfing, 2013). Historically, public administrations have been organized vertically around a function (e.g. labor, housing, health, etc.) and not around a problem (e.g. social exclusion). The public organization reduces the interactions between these functions; therefore, the ability to act on cross-functional issues is limited (Enjolras, 2010). This organization has led to fragmentations. To confront these problems, public actors are looking for solutions. Since the results of New Public Management policies are not satisfactory, new forms of public management and public governing are explored. The government first transferred some missions to the communities. Communities, whose resources are also shrinking, are seeking partnerships to support their actions, particularly with third sector partners and even with citizens. As a result, territorial innovation initiatives are increasingly being recognized. These initiatives no longer focus only on economic development and the attractiveness of the territories, but also in recent years on the social field (Klein et al., 2009; Hillier et al., 2004,). The recognition of social innovation is noticeable on several scales[1].

The aim of this research is to analyze social innovation dynamics of multi-actor innovation networks with a public service aim. These networks involve a variable number of public and private agents, in order to co-produce social innovation. This research is based on empirical case studies carried out within the European Co-VAL project. The analysis is illustrated by four French social innovation networks. This article is organized into four sections. In the following section, we address the issue of the emergence of wicked problems and how the public administration is shifting to manage these issues. We examine the literature on collaborative innovation, the issue of territorial proximity, and Actor-network theory (Akrich et al., 1988, 2006). The second section explains the methodology. The third section provides an analysis on the importance of collaborative innovation in the innovation process as well as a discussion on the nature of proximities and the territorial dynamics of the French case studies. This analysis is enriched by the study of controversies that occur in relation to the main functions of the social innovation network. The last section is devoted to the conclusion.

Conceptual framework

This paper is based on an original conceptual framework structured around three theoretical approaches, each providing different insights into the understanding of social innovation networks: the collaborative innovation approach, the literature on territorial proximities, and Actor-network theory. This conceptual combination is justified by the nature and scope of the subject: the analysis of social innovation networks involving multiple actors and organizations from different backgrounds (professions, culture, status, etc.), and mobilizing numerous ‘objects’ (legislation, method, Committee, infrastructure, etc.) (see Figure 1). Contributions from the literature on collaborative innovation characterize the type of collective governance at work in these networks and the way in which this type of governance affects the innovation process. The literature on territorial proximity provides a framework for analyzing the nature of proximities that are created, consolidated, or loosened through interactions between actors, resources, and assets on a territory. Actor-network theory (ANT) completes the understanding of decision-making relating to the innovation trajectory. It enables researchers to carry out inquiries and analysis based on relational ontology. By identifying the ‘interessement’ processes, changes and controversies associated with the construction of the social innovation network, this theory reveals the role of human and non-human “actants” in the development stages of the network and the innovation process, as well as their implications. The ANT provides both a theory and a methodological combination to trace the evolution of innovation processes.

Collaborative innovation and social innovation

The boundaries of the concept of social innovation are very different depending on the authors (Cloutier, 2003; Hillier et al., 2004; Dandurand, 2005). In this research we use the encompassing definition of “social innovation” from the TEPSIE EU project: Social innovations are “New approaches to addressing social needs. They are social in their means and in their ends. They engage and mobilize the beneficiaries and help to transform social relations by improving beneficiaries’ access to power and resources” (TEPSIE, 2015). Social innovation and public innovation are related terms. Some authors consider them to be synonymous (Sørensen & Torfing, 2013). Social innovation is linked to public innovation as “the value created accrues primarily to society as a whole rather than private individuals” (Phills et al., 2008, p.36). More and more “contemporary public policy problems carry characteristics of “wicked problems” (Termeer et al., 2015, p.680), such as ageing populations, immigrants, energy security or homelessness (Brown et al., 2012). Wicked problems are difficult to tackle because they have many multi-layer interdependencies, often multi-causal and with conflicting goals (Australian Public Service Commission, 2007). Thus, wicked problems cause serious challenges to public actors. The traditional bureaucratic structures are not appropriate for these complex problems, as these structures operate through single functional ‘silos’ (Brown et al., 2012). Therefore, tackling wicked problems requires stakeholders to coordinate their approaches (Termeer et al., 2015).

Figure 1

Conceptual framework and scope of the study

Conceptual framework and scope of the study

-> See the list of figures

Weak incentives to innovate and the complexity of contemporary problems contribute to the emergence of social innovation networks. According to the literature, innovation networks that mobilize a set of heterogeneous actors (third sector actors, public actors, private actors, citizens...) have more chance to innovate - as multi-actor meetings accelerate the processes of mutual learning (Sørensen & Torfing, 2011). These various actors find areas of common interest in order to co-create social innovation (Windrum et al., 2016). Multi-actor collaboration improves all phases of the innovation process (Sørensen & Torfing, 2013). Innovation networks may be based on collective governance, which corresponds to “the processes and structures of public policy decision making and management that engage people constructively across the boundaries of public agencies, levels of government, and/or the public, private and civic spheres in order to carry out a public purpose that could not otherwise be accomplished” (Emerson et al., 2011, p. 2). The complexity of wicked problems is a major driver for the development of collaborative governance in the context of social innovation with a public service objective. This need for cooperation is noticeable in the public management literature in which the understanding of service innovation has evolved from traditional public administration, New Public Management to New Public Governance (see Desmarchelier et al., 2020). The social innovation networks we are studying are trying to establish new modes of governance.

Territorial Proximity and innovation

The literature devoted to territorial development has focused on the dynamics of territorial innovation, first on the basis of business clusters (see the abundant literature on the theory of innovative environments, industrial districts, and specific resources). More recently, the promotion of the territory is based on various institutionalist theoretical studies on collective action (Decouzon et al., 2015), proximity, and territorial development (Torre & Gilly, 2000; Boschma, 2005).

The institutional approach of the territory integrates the historical dimension, non-market relations, and the interplay of institutions (formal and informal) into their analysis (Gilly & Leroux, 2005, p.258). Territorial dynamics connect constraining or facilitating factors at a macro level, such as the institutional framework, with forms of coordination at the micro level, such as embedded relations between agents (social proximity) or institutional arrangements (e.g. norms and values embodied in specific exchange relations). However, this approach considers that institutional mechanisms are never one-sided nor definitive (Gilly & Leroux, 2005, p.261). Local and extra-local institutional compromises are provisional, and local spaces can adapt themselves or act on macro-economic and macro-institutional structures.

Territorial dynamics versus actors’ relationships

These studies explore two types of causality: on the one hand, analysis of territorial dynamics from a “heritage” perspective. This fieldwork studies the impact of the territory on the actors; on the other hand, another piece of fieldwork analyzes the consequences for the territory of the relationships between actors that are established on the territory.

The first piece of fieldwork focuses on cumulative and collectively capitalizable assets (Boschet & Rambonilaza, 2010). For proximist authors, development is a matter of the construction and activation of resources (Colletis et al., 1999). These authors used Williamson’s distinction between generic and specific assets. The specificities of the territory often determine the configuration of actors wishing to be involved in local projects (Bonnin & Grémy, 2004). The territory is a vector for the construction of collective identities that facilitates the realization of social links between individuals where the market fails to coordinate them (Gingras, 2004). Thus, social innovations are strongly rooted in the territory from which they have emerged and require “partnership-based governance”, (Chanut-Guieu et al., 2014).

Studies that explain the influence of actors on the territory indicate that the local development process is reliant on the coordination of actors. The ability of actors located in a territory to control its evolution must be driven by a collective dynamic (Boschet & Rambonilaza 2010) which organizes the combination of resources and assets, and implements collective strategies likely to influence the economic future of the territory (Colletis & Pecqueur, 2005).

Dimensions of proximity

In France, the school of proximity was built on the idea that the location in a geographical and social space influences the interactions of the actors (Torre & Gilly, 2000). The coordination of actors during collective actions, especially off-market, is based on several types of proximity between local actors. The number of dimensions varies according to the authors. Organized proximity is the capacity that an organization offers to induce interaction between its members (Torre & Rallet, 2005). In the framework of the Evolutionary Economic Geography, Boschma (2005) distinguishes cognitive proximity (knowledge sharing) from social proximity (belonging to a network), organizational proximity (belonging to an organization), institutional proximity (formal and informal shared institutions), and geographical proximity. The types of proximity may have slightly different contents depending on the author. For the purpose of this article, spatial proximity will refer to the “separation in space and the links in terms of distance. It integrates the social dimension of economic mechanisms, or what is sometimes called functional distance” (Torre & Gilly, 2000, p.12-13). Physical proximity is not enough to set up collective dynamics but makes social contact possible (Kirat, 1993). These actors develop non-geographical proximities such as institutional (or cultural) proximity, which “is based on the adherence of actors to a common space of representations and rules of action guiding collective behaviour” (Colletis et al., 1999, p.27-28). Organizational proximity refers to the interactions between actors within (or between) organizations. Groups of agents, which are a priori independent, set up coordination procedures in order to achieve the objectives on which they have agreed (Colletis et al., 1999). Organizational proximity can be independent of any spatial dimension (Tremblay, et al., 2002).

Insofar as the combination of these proximities may have positive or negative effects on the objectives of a network (Boschma, 2005), we will analyze the way in which these types of proximities facilitate or hinder territorial social innovation. The measurement of these proximities has been studied by different authors. Geographical proximity is identified by the spatial distance between two actors (Torre & Gilly, 2000; Pecqueur & Zimmermann, 2004). Cognitive proximity can be estimated by observing the patents filed by the actors (Breschi et al., 2003) as well as knowledge and methods (Hansen, 2014). Network analysis characterizes social proximity (Lazega, 1994), organizational proximity is identified by belonging to the same organization (Torre & Rallet 2005). Institutional proximity is more difficult to detect (Kirat & Lung, 1999; Hamouda & Talbot, 2018). It operates both at a macro level through the political and institutional framework that frames the interactions and the resulting common representations (Hamouda & Talbot, 2018), and at a meso level through belonging to the same institutional form (Balland, 2012). The informal dimension of institutional proximity can be transcribed by the notion of inter-organizational culture (Lu et al., 2016). Therefore, the tools for measuring institutional proximity are both formal (e.g. the creation of shared rules) and informal (e.g., common representations (Gilly & Lung, 2005), common cognitive maps (Torre, 2014)).

The Actor-network Theory

Actor-network theory (ANT) (Akrich et al., 1988, 2006) complements previous studies by taking into account, in the analysis of the social innovation network, not only social interactions between humans but also the active role of non-human objects and their interactions with humans. These elements are referred to as “actants” and each actant - whether human or non-human (technology, knowledge, report, places, devices, etc.) - is equally important to the network. This theory is based on a relational ontology framework in which it is no longer the specific entities that are important, but the relationships between them. This theory examines both humans and non-humans at the same level of analysis, according to a principle of human/non-human symmetry (Latour, 1991). ANT explains how actors form alliances and use artifacts to enlist other actors (Callon & Latour, 1981; Law, 1991). Based on analysis of the translation processes, the observation of divergent interests, controversies, and situations of trials of strength (Callon, 1986), this theory enables us to explain the innovation trajectory of the network and the situations in which projects fail or succeed. ANT assumes that conflicts and controversies are solved through the creation of networks of actors around a common solution that emerges from negotiations. These conflicts lead to the adaptation of rules. J-D Reynaud’s research (2003) examines collective territorial initiatives, and collective arrangements as a dynamic of agreements and compromises. The process of translation presents the idea that innovation does not emerge as the result of a linear development path but is transformed and modified along its unpredictable trajectory (Modell et al. 2017).

ANT is a theoretical-methodological approach. Without narrowing ANT to its analytical instrumental dimension, we will summarize the innovation process in a table using the different stages of the innovation process. Their meanings are as follows. The first stage seeks to identify the need for change, the actors involved in the innovation process, their interests, and the non-human actants. The problematization formulates the complex issue, which gathered all the actors with a focus on solving a problem for the benefit of all. The third stage (involvement) aims to convince the various actors concerned by the problem and to select the actants for their contributions, input, and expertise. The fourth stage (enrolment) takes the form of a generalized negotiation where actors can act and give their opinion in order to improve the overall process. The fifth stage (mobilization) amplifies local actions to disseminate them on a larger scale, thanks to the involvement of spokespersons who “translate” the different interests involved from one register to another.

Methodology

Our objective is to analyze, through the dimensions of proximity, the inquiries on innovation processes and collaborative governance, the process by which networks of actors develop social innovations with a public service aim (the innovation process) and create territorial dynamics (Figure 1). For this purpose, we have adopted a qualitative method based on case studies carried out within the European Co-VAL project. Case studies have been undertaken in different European countries and have been the subject of national reports (Merlin-Brogniart, 2019). Insofar as national contexts influence the administrative territorial division, as well as the territorial dynamics, we will focus on a single country, the French case.

Data collection

The analysis provides insights into the interrelationship between different proximity dimensions and the influence of these dimensions on territorial dynamics. The purpose is also to specify the controversies within the key functions fulfilled by the coordination of actants in the social innovation network. The criteria for selecting case studies were the following: each case study has a concrete social innovation, and an innovation network is required to realize and develop this social innovation. The case studies come from different societal fields (elderly care issues, long-term unemployment, migration, education, and environmental protection) and involve various types of innovations, innovation processes, and innovation networks. The collection of information was based on a comprehensive survey composed of semi-directive interviews, each lasting approximately one hour - recorded and transcribed - conducted from July 2018 to March 2019 with the key actors of each network, with a total of nineteen interviews. As Huberman and Miles (1991) suggest, these data were triangulated with existing secondary data on these innovation networks, observations, and internal documents (websites, reports, etc.). The interviews cover the history of the project and its development over time. The methodology is in line with the translation chain approach recommended by Callon (1986). The guideline of interviews was structured with five key dimensions in order to identify the type of innovation process of the case study, the type of innovation network, drivers and barriers of the innovation, institutional factors, and impacts on the project and type of performance. Given the focus of this paper (the innovation process) the part of the interviews regarding the impacts and types of performance has not been exploited. The first two dimensions and to some extent the two remaining dimensions have provided the empirical basis for this research. The reconstitution of the history of the social innovation network is based on the interviews and was cross-checked with secondary data. The different types of proximity were identified on the basis of the nature of measurement set out in the literature review. Some authors propose the analysis of discourse to identify the type of proximity (Hamouda & Talbot, 2018; Hansen, 2014). In this article, we use this method to identify the types of proximities mobilized by the various actors.

Case-studies presentation

We have chosen to focus on single-country case studies because the institutional and historical context of the country influences the development of social innovation networks (Merlin-Brogniart et al., 2022). France is classified as a corporatist regime in Esping-Andersen’s typology (1990). This country has a long democratic history with a strong tradition of state governance. Therefore, even if the country has decentralized corporatist structures and the government encourages multi-stakeholder networks, some rules maintain uniformity and control of territories, such as the principle of territorial equality, which required that each local experiment could be extended over the whole nation. This principle has led de facto to a limitation of experimentation. Local authorities are increasingly being encouraged by governments to mobilize territorial actors in order to bring about “social innovations” (Petrella & Richez-Battesti, 2010, Gallois et al., 2016). This change of perspective was noticeable in France in the 1990s in the context of New Public Management. Territorial development policies have then moved from a logic of positive territorial discrimination to an objective of competitiveness and employment growth (project logic). This new perspective mobilizes the involvement of local elected representatives who are gaining more weight as policy makers (Taiclet, 2011). In this context, co-production with the inhabitants and users of the territory is very far upstream, as part of institutional innovation (public-private user partnership, project mode management) (Lipovac & Boutonné, 2014). Another consequence of the country’s historical background is the perception of multi-actor social innovation networks – “primarily described as a solution for modernising the state and reducing costs by sharing financial and human resources and the expertise of actors” (Merlin-Brogniart et al., 2022). The institutionalist approach contributes to considering social innovation as a territorialized system (Richez-Battesti et al., 2012), which is not necessarily as prevalent in other European countries.

In this research, four social innovation networks have been investigated, encompassing different societal fields. The first case study examines the MAIA method (Method of action for the integration of healthcare and support services in the field of Autonomy). The aim of this method is to overcome the barriers of service system fragmentation regarding populations in vulnerable situations. The second case study examines the project “zero long-term unemployed territory” (TZCLD). The idea of the project is to be able to offer to any unemployed person in a given territory a job adapted to his/her know-how, paid the guaranteed minimum wage, based on an indeterminate duration contract (CDIs). Jobs are suggested by the unemployed persons on the basis their know-how, and the needs of the territory. To finance these jobs within a company, the project is based on an innovative mechanism: transfer to the new employers the social expenses regarding unemployed persons. The third case study examines the case of “The Fabrique saillysienne” (FS) whose objective is to implement, with the village’s residents, citizen projects for the village of Sailly-lez-Lannoy. These cooperative projects have a general interest objective. The inhabitants’ projects mainly have an environmental purpose. The last case study “the Melting Potes Program” (MP) relies on a civic service commitment offered by the Unis-Cités association to refugees. The aim of the program is to facilitate the inclusion of minorities in the French community by using civil service as a springboard and to eliminate prejudices against them.

Social innovation networks, innovation processes, and proximities

The analysis first focuses on the need for collaborative innovation in the development of social innovation networks with a public service aim. It examines the role of public, non-profit and private actors in the network, and sets out the types of innovation strategies. The following sub-sections examine the different types of proximities that these innovation networks attempt to build and discuss the territorial dynamics generated by proximities. The last section uses ANT to examine the changes and the actants involved when incidents occur in relation to the main functions of the social innovation network.

Collaborative innovation, innovation processes and Social innovation network

The first finding of our analysis is that social innovation with a public service aim could not have been achieved without the creation of an innovation network. The public actor has been essential in each case study, either as legislator and funder (TZCLD, MAIA, MP), or as initiator of the project (FS). The involvement of elected representatives as policy makers is different according to the case studies. Their role is essential when the territory is identical to the town. For example, within the Fabrique Saillysienne, the town hall plays the role of initiator and coordinator for logistical and sometimes financial support. The involvement of the elected representative is also tangible in the TZCLD program because the application to the program is carried out and justified by the city. For the MP program, without being the main “actant”, the city’s elected representatives have indirectly facilitated the emergence of the project through their policy of inclusion of minorities for several years. Finally, when the territory associated with the social innovation network is larger than the city, the role of the elected official is less important (e.g. MAIA). The public actor is also crucial in shaping public policies at a national level: the recognition by the public authorities of a social need and the acceptance or legitimization of the innovative solution are drivers for social innovation (Dandurand, 2005). In the cases studied, modifications of the legal framework were necessary to implement the project.[2]The third sector also has a prominent place in these networks either as initiator of the project (TZCLD, MP)[3], as a partner sharing the daily life of the users or as an actor enabling the project to be legitimized (TZCLD, MP, MAIA, FS). The participation of private actors can also be crucial to ensure the proper functioning of innovation networks, especially when they represent an essential link in the chain of actors in the network (Zafiropoulou, 2013). In the framework of New Public Management, citizens and users’ participation is seen as an important component of public management as co-producers (Osborne, 2006, 2010). In our case studies, citizens can be at the heart of innovation such as in the Fabrique saillysienne, which relies on citizens’ competences[4] - or to some extent - the users in the case of civic service insofar as they transfer knowledge to other users (MP). In the other case studies, users are not co-producers of the innovation.

Two types of governance and implementation strategies with different interaction modalities and implementation processes have been identified. The first implementation strategy for network emergence is the strategy “help it happen”; when cooperation does not occur spontaneously (De Stampa & Somme, 2012) an ascending approach is favored (i.e. top-down and bottom-up); the method is developed at national level but the deployment is left to the initiative of the local manager. The other case studies are part of a “bottom-up” strategy with the creation of tools ex nihilo, by experimentation at local level, followed by a process of dissemination based on knowledge sharing. The choice of implementation strategy depends on the type of social problem to be addressed. The bottom-up strategy, which corresponds to the unpredictable and self-organizing innovation process (Greenhalgh et al., 2004) is often used by private actors, or actors from the third sector to develop their own solutions, without inducing a model (Merlin-Brogniart et al., 2022). The “help it happen” implementation strategy usually corresponds to one of the fields historically assigned to the State such as health, education, utilities, social services, etc. Therefore, the public authority wishes to control the way in which the project is carried out. The initiator, whether it is a public or private actor, will be forced to collaborate with the government because the project requires public intervention (due to the scope of the project, the method of financing, a legislative change, staffing, etc.), as well as for its dissemination (Merlin-Brogniart et al., 2022).

Social innovation networks and the measurement of proximity dimensions

Our proposal is based on the idea that by producing social innovation, social innovation networks with a public service aim change the territorial dynamics. Indeed, responding to complex needs leads to sharing knowledge, information, and expertise on the social issue and to developing new common structures to overcome silo problems. The resulting new structures and relationships (proximities created, destroyed and reconstructed) are established at local level. Due to their nature, spatial proximity and organizational proximity are easily identified through indicators). Other types of proximity are only partially visible through indicators. In order to have a better understanding of the different proximities, we carried out an analysis of the discourse of each interview. An illustration of each dimension of proximity is proposed below with case study illustrations (Appendix 1). Each type of proximity is identified by an acronym: Spatial proximity (SpP), Social proximity (SP), Institutional proximity (IP), Cognitive proximity (CP), and Organizational proximity (OP), and a number for each verbatim.

Creation of Institutional proximity and social proximity

The aim of the social innovation networks studied lies in restructuring or creating an interaction between professions or between citizens. Social needs require a local - integrated - and coordinated response. Various kinds of problematic and innovative solutions have been proposed in each network. For example, each local MAIA network develops its own culture and therefore its own institutional proximity. The networks generate social links at two levels. Some innovation networks (re)-create social links among the target groups (MAIA, MP, FS, TZCLD) (SP1, SP2, SP3, SP4, see Appendix 1). Social proximity is also generated by the social links established between partners (SP5, SP6) who are either citizens (FS) or professionals (MAIA, TZCLD, MP). The creation of such proximity tries to overcome the problems of silos by creating direct relationships between partners of different professions who do not usually work together. This objective is expressed in the development of an informal institutional proximity between the actors (cultural, habits). Sometimes, the partners share a rejection of the common formal institutional environment (IP1, IP2). Institutional proximity is also reflected in the establishment of a new culture embedded in the conceptual innovation, such as the notion of “additional” or “complementary” employment in the TZCLD project (IP3), or refers to the co-responsibility between professionals toward users (IP4).

The setting up of social innovation networks - Organizational proximity and Cognitive proximity

Organizational proximity is achieved by the way the social network operates. The structures created facilitate collaborative governance (also visible in Table 1’s enrolment stage -ANT). In the MAIA project, committees composed of various actors (a tactical table and a strategic table) allow collaboration spaces. In the TZCLD project, a local “Job-Oriented Company” is created to hire the unemployed, and in each of the authorized territories, a local steering committee in which the local employment partners and the long term unemployed are comprehensively mobilized by the committee, is also implemented.

Organizational (but also cognitive and institutional proximity) is also reflected by the creation of new professions that ensure the coherence of the system through the commitment of each actor in this territorial project (OP1, OP2). For example, the MAIA method creates value via the professional dynamics generated through the harmonization of professional practices (CP2) and supervised by a dedicated professional - the MAIA pilot.

Cognitive proximity is achieved through the creation of new methods (CP1, CP2, CP3). The question of cognitive proximity is also important in the context of social innovation networks, not to develop an innovation but to understand it and to reduce informational and expertise asymmetries. Social innovations require a certain level of understanding of the system in which the innovation network is embedded (CP4, CP5). This global understanding is also necessary in the case of TZCLD where assessing the indirect costs is crucial for the sustainability of this network. On a more general level social innovation networks also rely on the capacity of each partner to understand and learn about the job done by the other partners on the targeted population, as well as the constraints of their work. These efforts may reduce potential conflicts and should lead to strengthening the cognitive proximity of the partners (see last section - ANT).

Territorial dynamics and spatial issues: the choice of the territory

Spatial proximity generally corresponds to the boundaries of a city or districts of a city. However, some territorial divisions are more complicated to manage. Therefore, understanding the notion of territory is essential for the success of the social innovation network (SpP1, SpP2). For example, professionals often assimilate the territory to the area in which they are working. These differences in territorial understanding may be problematic for ensuring the coherence of the service provision system as it becomes difficult to work on a service continuum (e.g. MAIA). Moreover, territories often overlap. The challenge relies on acceptance by the multi-actor network of territorial redefinition., For the MAIA project, each MAIA pilot must reflect on the population and not on the territory of the local holder of the project in order to propose a complete and harmonized service adapted to the local territory.

The two causalities presented in the literature (the heritage perspective and the influence of actors on the territory) have been identified in the case studies. The interactions created between the actors lead to the determination of the territory and the service offer but, at the same time, the networks rely on the existing networks and resources in order to take advantage of existing social links, as in MP and TZCLD (SpP3, SpP4). It should be noted that territorial resources differ from specific assets (Colletis & Pecqueur, 2005) by the process of valorization, which allows the resource to be definitively identified. Therefore, the role of the actors is central because they must organize and coordinate themselves to integrate the resource into the local development dynamic (Grison et al., 2018).

The nature and size of the chosen territory is also an essential element for the success of the project. The analysis of territorial needs can only be carried out on a local scale. In addition, confidence and social bonds are more easily achieved on a reduced scale. Networks of associative partners at the local level may be different from the national network of partners. Therefore, it is necessary to consider territorial specificities. Significant differences were found between rural and urban areas. Rural territories seem to be more appropriate to some measure as they are better defined and social cohesion often already exist (social proximity), whereas urban areas have blurred boundaries and may pose problems of equality between citizens of the city. Thus, the responsibility of public action should be entrusted to the closest competent entity of the people directly affected by the action (ETZCLD, 2018).

Social innovation processes and Controversies

ANT complements the previous approach to proximities by specifying the nature of actants and identifying controversies related to the main functions of social innovation networks. The innovation process is summarized in a table describing the different stages of the innovation process for each case study (Table 1).

The previous section outlined that social innovation networks produce a function of participation in a community (creation of institutional and social proximities). This function is based on a process of “social” involvement of the users. The FS involves citizens in the life of the city, TZCLD reintegrates the long-term unemployed into working life and society, MP integrates migrants into society, MAIA reintegrates the elderly into social life and re-establishes their rights when necessary. In the case studied, the ‘interessement’ of actants is founded on the benefit for the users and for mutualization of skills between non-profit, public, and private organizations. These networks reduce asymmetries of information (between elected representatives and citizens (FS), toward minorities (MP), or between different actors working in silos (MAIA, TZCLD, MP). However, there are still information asymmetries concerning the users as they are generally not involved in the innovation process (MAIA, TZCLD, MP). Nevertheless, the resulting service makes the users more active compared to traditional service (TZCLD, MP). Only the FS places the citizen at the heart of the innovation process. Beyond this problem of information asymmetries, controversies concerned the legitimacy of the actants (e.g. not all the elected representatives were convinced by the ability of citizens to take charge of collective projects in the FS -which leads to discussions). The legitimacy of the MAIA pilot and the holder have sometimes been challenged in local MAIA projects.

Social innovation networks have a role of organizing the heterogeneous actants around a main function (health, unemployment…) and bringing coherence to the offer (organizational and cognitive proximities). This role is visible within the problematization stage (Table 1): Social innovation networks seek to provide a collaborative offer when the public or the private offer is deficient. Three of the four case studies are related to solving wicked problems (silo problems) that can only be tackled through cooperation between stakeholders (see Context, Table 1). The last one (FS) seeks to develop a new method of citizen participation. The negotiation meetings are characterized by the presence of facilitators - also “translators”- who seek to ensure that the needs and constraints of each actor are taken into consideration. However, the risk of relying on volunteering is the self-exclusion of some actors. The controversies identified are the following: In the MAIA case, controversies cover a series of misunderstandings (e.g. as the method was funding via the Alzheimer Plan, some people thought that other diseases were not concerned; the use of the term “house” (“maison”) in the first acronym results in an understanding of the method as a structure instead of the development of coordination between actors). Somme et al. (2014) reported some barriers linked to excessive power given to doctors and to acute care, professional cultural resistance, as well as imbalance of power. Controversies also concerned the legitimization of the holder in the MAIA project (e.g. at the experimental stage a private holder led to conflict of interest). The MAIA pilot found it difficult to mobilize actors due to their lack of time and the fact that some actors are already involved in other networks. The greatest controversy of the MAIA project is the existence of similar programs, which raises doubts about the need to build this network, which resulted in 2022 in reunification with other networks to form a coordination support facility (named DAC). In the TZCLD project, legislation has first been an obstacle to the development of projects (e.g. the principle of territorial equality), and controversies occasionally lie in the choice of eligible territories. In the FS, a Charter was created from the outset to avoid inappropriate behaviors. As the project relies on the volunteering of inhabitants, their lack of time and motivation sometimes endangered the sustainability of the project. In the MP project, differences in equal treatment have been noted depending on the territory (the Civic Service Agency was not uniform in the processing of files). Moreover, some national private funders have been reluctant to put money into helping migrant communities.

Another function of the social innovation network is emphasized to a greater extent by the use of ANT rather than proximity: Social innovation networks connect the various actors with a view to lobbying public authorities: Some of them are set up partly to obtain public (or private) funding, or to change the law. In the TZCLD project, one of the controversies was linked to competition: to convince the government, the TZCLD project created a committee to ensure non-competition of the job created by the JOC with existing jobs. The lobbying also concerned private or non-profit partners. For example, in the MP project, managers had to convince reluctant partners to trust migrants for civic service tasks.

Thus, the development of social innovation networks involves a dynamic of agreements and compromises that shapes the trajectory of innovation and requires the creation of dedicated methods and tools.

Table 1

The five stages of the innovation process and controversies according to ANT

The five stages of the innovation process and controversies according to ANT

-> See the list of tables

Conclusion

In order to analyze the way in which social innovation networks with a public service aim create new processes of collaborative innovation, we proposed a framework combining collaborative innovation literature, the ANT, and the insights of proximity theory. Given the wicked problem, and the dissatisfaction with existing solutions, some actors are seeking to create collaborative solutions based on networks including public, private, and associative actors working on the same social problem. Our case studies have revealed two types of implementation strategies for network emergence and innovation diffusion: the “help it happen” and “bottom-up” strategies, the choice of which depends on the type of social problem to be tackled, and its scope. In both implementation strategies, territorial regulation gives rise to different cooperative solutions depending on the specificity of the territory, the local networks already operating at a local level, and the diversity of actors who belong to the local network. The case studies highlight the fact that broad latitude is left to local managers to take into account the specificities of the territory, its resources, its institutions, and the existing networks of actors. For social innovation that does not fall within the historical public service, almost complete freedom of action is left to the actors, the network being framed just by principles of action. Among public actors, the role of local elected officials in innovation policies appears to be significant when territorial networks are built on the basis of the city or city districts. The public actor acts as legislator, funder, or initiator of the project. The third sector is essential - at an early stage - to ensure that the project will be taken into consideration by the public actor and to legitimize the project, being, in the eyes of the actors “guilty of nothing”. The users are mobilized in different ways in the innovation process depending on the nature of the project. They can be very poorly mobilized, supervised in their participation, or autonomous.

The objective of these networks is to create social and institutional proximity through interactions between professionals or citizens. The proximity of actors improves the provision of services to the targeted groups at a local level. These networks also seek to change certain formal rules, which hinder their project, or certain parts of their project, from being carried out. By creating spaces for discussion, by improving the awareness of the constraints of each actor and reducing asymmetries of resources, power, information, or knowledge between actors, this collaborative governance allows actors to imagine new ways of organizing networks in order to innovate. The organizational proximity created by the innovation network reflects the adjustments between professionals or between citizens that are necessary to ensure collaborative governance. Cognitive proximity is constructed around methods or know-how associated with network innovation.

Social innovation networks with a public service aim develop specific territorial innovation dynamics by relying on existing resources and networks that ensure social links. However, in a reverse movement, the creation of new interactions between the “actants” of the network shapes the boundaries of the territory and the complex service offering. Therefore, territorial regulation through social innovation networks leads to a community of actors (and actants) and a range of solutions that may differ according to the territory.

Territorial dynamics do not occur without controversies. By identifying ‘interessement’ processes and controversies within the essential functions generated by social innovation networks, the ANT reveals the constantly evolving nature of the relationships between actors within the translation process. This analysis reveals various types of controversies (misunderstandings, problems of stakeholder legitimacy, inadequate regulations, cultural or professional resistance) that have contributed to shaping the trajectory of social innovation’.

The context of social crisis and increasing control in public expenditure has surely accelerated the experimentation of new forms of collaboration and innovation processes and has facilitated the prompt recognition of the capabilities of territorial innovation networks. The case studies indicated that the choice of territory and the associated population is essential for the success of social innovation. The existing dynamics are linked to the territories of professionals, who do not necessarily know the territory of the other institutions with which they would be expected to collaborate. Thus, each territorial delimitation will have an impact on the dynamics of the actors. These innovation networks have many challenges to face. In particular, local project leaders must be credible in order to ensure the commitment of the local actors and the sustainability of the project.