Mapping the integrative field: taking stock of socio-technical collaborations

Abstract

Responsible innovation requires that scientific and other expert practices be responsive to society. We take stock of various collaborative approaches to socio-technical integration that seek to broaden the societal contexts technical experts take into account during their routine activities. Part of a larger family of engaged scholarship that includes inter- and transdisciplinarity as well as stakeholder and public engagement, we distinguish collaborative socio-technical integration in terms of its proximity to and transformation of expert practices. We survey a variety of approaches that differ widely in terms of their integrative methods, conceptions of societal context, roles, and aspirations for intervention. Taking a handful of “communities of integration” as exemplars, we then provide a framework for comparing the forms, means, and ends of collaborative integration. We conclude by reflecting on some of the main features of, and tensions within, this developing arena of practical inquiry and engagement and what this suggests for integrative efforts aimed at responsible innovation.

Keywords:

• socio-technical integration
• collaboration
• transdisciplinarity
• public engagement
• technology assessment

1. Introduction

Throughout Europe, North America, Australasia, and elsewhere, the integration of social, ethical, cultural, environmental, and other “societal” considerations into scientific and technical practices has been a noticeable theme within recent research and development policy mandates (Fisher and Rip 2013; Macnaghten, Kearnes, and Wynne 2005), programs (Owen 2014; Rodriguez, Fisher, and Schuurbiers 2013), and program planning (Ommer et al. 2011; Fisher 2014). Departing from past policies for societal research on science, these policies suggest more collaborative models of societal research with science. Meanwhile, practical scholarly engagement with expert and technical practices is itself framed in terms of collaborating with and directly influencing these same practices. We take stock of this emerging field of engaged scholarly collaborations aimed at socio-technical integration, develop a framework for comparing integrative methods and goals, and reflect on the results in light of responsible innovation.

Although socio-technical integration can take many forms, it is generally characterized by the reciprocal and productive combination of opposing, diverging, or previously segregated contextual dimensions of expert practices in the pursuit of consequential, often policy-relevant, ends. It is part of a larger family of participatory research approaches that include scholarly engagement (Felt 2014; Hackett and Rhoten 2011), ethical, legal, and social implications/aspects (ELSI/ELSA) research (Juengst 1991), laboratory studies (Hess 2001), team science (Stokols et al. 2008), applied ethics (van Gorp 2005), technology assessment (van Eijndhoven 1997; Guston and Sarewitz 2002), inter- and transdisciplinarity (Bergmann et al. 2012; Frodeman, Klein, and Mitcham 2010; Wickson, Carew, and Russell 2006), and public engagement (Macnaghten, Kearnes, and Wynne 2005; Wynne 2011), among others. Like many of these, engagements aimed at socio-technical integration tend to confront challenges arising from differences in language, practice, values, power, and other potential sources of “incommensurability”, “incongruence”, or “discordancy” (Bracken and Oughton 2006; Eigenbrode et al. 2007; McCormick et al. 2012; Rabinow, Bennett, and Stavrianakis 2009; Sievanen, Campbell, and Leslie 2011). As we define it, however, collaborative socio-technical integration is distinct from these related approaches primarily insofar as it involves close, transformational interaction with scientific and technical experts.

Such collaborative integration appears as a central component of visions for “responsible (research and) innovation” (Guston et al. 2014; Owen, Macnaghten, and Stilgoe 2012; von Schomberg 2013), “anticipatory governance” (Barben et al. 2008), “reflexive governance” (Voss, Bauknecht, and Kemp 2006), and the notion of “convergence work” (Stegmaier 2009). These visions concern the explicit integration of societal considerations into scientific and technical practices as knowledge systems and technological trajectories evolve, and they have precedents in earlier calls (Bronk 1975; van Eijndhoven 2000; Mitcham 1994), programs (Hollander and Steneck 1990; Juengst 1991), methods (Friedman 1996; Schot and Rip 1997), and frameworks (Williams and Edge 1996). The reinvigoration of established approaches and the proliferation of more recent ones, however, provide an opportunity to take stock of the collaborative pathways to integration. Accordingly, we outline the main features of collaborative socio-technical integration in terms of its multi-dimensional interactions with expertise. We then take as exemplars a number of “communities of integration” to illustrate a conceptual framework that facilitates comparison while capturing a diversity of relational (forms), methodological (means), and normative (ends) orientations found within this integrative field. The paper concludes by offering reflections on synergies, tensions, and future research directions in relation to responsible innovation.¹

2. Overview of the field

Collaborative socio-technical integration is the subject of explicit or implicit treatment in a wide range of literature, including case studies (Ribes and Baker 2007), projects (Nydal, Efstathiou, and Laegreid 2012), reports (Paletz, Smith-Doerr, and Vardi 2010), handbooks (Bijker and d'Andrea 2009), and manifestos (Balmer et al. 2012). Also discernable is a handful of sustained or coordinated programs that have given rise to several communities of practice for socio-technical integration (Cho et al. 2008; Fisher and Schuurbiers 2013; Friedman et al. 2013; Goorden et al. 2008; Rip and van Lente 2013). Although few comparisons of integrative roles and programs have been attempted (Boenink 2013; Calvert and Martin 2009; Rabinow and Stavrianakis 2013), a number of recent edited collections provide contours of the field. Some of these collections are organized around particular theoretical frameworks (Gorman 2010; Plaisance and Fehr 2010); others assemble various methodological approaches (O'Rourke et al. 2013), sometimes focusing on the laboratory as the site of integration (van der Burg et al. 2013; Doorn et al. 2013); while others emphasize normative aspects of integration (van de Poel and Verbeek 2006; Zuiderent-Jerak 2007) or its relation to policy and governance (Fisher 2011; Levidow and Neubauer 2014; Stegmaier 2009).

2.1. Delineating socio-technical integration

Rodriguez, Fisher, and Schuurbiers (2013) define socio-technical integration as

the explicit incorporation of activities devoted to broadening the social and ethical aspects that are taken into account during core scientific and engineering research and development (R&D) activities in such a way as to shape R&D pathways in socially desirable ways. (1126)

Fisher and Maricle (2014) define it as “any process by which technical experts account for the societal dimensions of their work as an integral part of this work” (Fisher and Maricle 2014). Notably, these definitions leave open the question of whether integration is accomplished by technical experts on their own or whether it takes the form of transdisciplinary and/or cross-sector engagements involving others. On the other hand, Guston's (2014) definition includes the humanities and social sciences: “Integration is the creation of opportunities, in both research and training, for substantive interchange across the ‘two cultures’ divide that is aimed at long-term reflective capacity building”. We preserve this meaningful distinction between integration as an activity proper to technical expertise and integration as an essentially cross-cultural endeavor by focusing on the idea of collaborative socio-technical integration, which we conceptualize in terms of three key characteristics: As we explain next, collaborative socio-technical integration focuses on relations of expert practices to their (often segregated) societal context, operates in close proximity to the expert practices in question, and functions to catalyze or support transformation of those practices in their societal context.

Accounting for context. As noted, the literature emphasizes bridging segregated or diverging categories, with “technical” denoting some form of specialized expertise and “societal” denoting important contextual dimensions of expertise that are either overlooked or neglected. Depending on the particular approach and the participating experts, the societal context can encompass cultural, ethical, political, environmental, linguistic, epistemological, axiological, and numerous other value dimensions. Collaborative integration seeks to work deliberatively, explicitly, and reflectively across variously conceived and constructed “socio-technical divides”. These include expert and lay ways of knowing (Pohl and Hirsch Hadorn 2007; Stokols et al. 2003), disciplinary distinctions between social and technical actors (Gregory et al. 2012), divergences between human value dimensions and technical rationalities (Lempert et al. 2006; Schon 1983), and intersections between science and policy/politics (Lempert, Popper, and Bankes 2003; Pielke 2007). Explanations for why these divides and limitations exist in the first place vary, but they tend to make reference to the nature and acquisition of expertise itself.

Close proximity. This emphasis on the contextual limitations of expertise implies a role – or rather roles – for others. Just as integration approaches tend to share an analytical distinction between societal and technical categories, so also do they tend to relate these categories through the work of intervention-oriented integration agents. While integration approaches differ in how they represent the roles of these would-be collaborators, they all tend to calibrate their interventions to the everyday practices, judgments, and experiences of the experts in question. Some see expertise as lacking a particular embodiment of ethics or values and emphasize the introduction of and sometimes advocacy for new content or capacities (Boenink 2013; Rabinow 2009; Shilton 2013). These approaches often portray the integration agent as an expert or representative who speaks with authority. Others see socio-technical divides as co-produced (Jasanoff 2004) and emphasize social, moral, or epistemological reflexive awareness of the focal experts (Doubleday 2007; Eigenbrode et al. 2007; Fisher and Miller 2009; O'Rourke and Crowley 2013; Tuma 2013). These approaches tend to emphasize the mutual learning that occurs on both sides of the collaboration and that enables the integration agent to articulate contextually specific questions, reflections, and insights. In all cases, integration approaches seek to situate themselves closely alongside and within the practice of technical expertise, moving in the same cultural, linguistic, and conceptual spaces and operating with “adjacency” (Rabinow and Bennett 2012) and “proximity” (Guston 2014) and doing so “midstream” (Fisher, Mahajan, and Mitcham 2006).

Practical transformation. Finally, collaborative integration emphasizes influencing and affecting change in how expert practitioners identify and engage their societal contexts (Calvert 2013; Darling et al. 2014; Fisher 2007; Flipse, van der Sanden, and Osseweijer 2013; Krabbenborg 2013; Schuurbiers 2011; Shilton 2013; Taebi et al. 2014; Valve and McNally 2013; Wickson, Carew, and Russell 2006). Successful collaborative integration inflects one or more integral components of the engaged expert practices, whether by effecting new “inputs” to them (Barben et al. 2008; Cho et al. 2008) or through their “modulation” (Fisher, Mahajan, and Mitcham 2006; Rip 2002). Thus, a crucial difference between collaborative socio-technical integration and other forms of multi-disciplinary research is that the interactions include an explicit recognition by the domain experts of the societal contexts in which their work is conducted and in which it will be applied (Rodriguez, Fisher, and Schuurbiers 2013; Schuurbiers and Fisher 2009; Shilton 2013). Successful integration demonstrably changes technical practice, modifying it both formally and tacitly (Fisher and Maricle 2014).

These three characteristics of collaborative approaches to socio-technical integration are consonant with position, long maintained by the field of Science and Technology Studies (STS) that, while experts are “indispensible to the politics of knowledge societies” (Jasanoff 2005), they are nevertheless not the only ones “competent to make decisions about science and its trajectory” (Kleinman 1998). Here, we treat expertise as something that is acquired by individuals and social groups through various processes of socialization and that is embodied in specific sets of cultural practices. Accordingly, the collaborative engagement of expert practices tend to see expertise as fundamentally limited in important ways due to its specialized focus, but calls for integration differ in regards to how this limitation is constituted and how it ought to be addressed.

2.2. Calls for socio-technical integration

Even with the restrictions imposed by these three characteristics, contemporary calls for socio-technical integration are diverse in terms of their source and scope. Some highlight scientific agency and emphasize “collaborative” (Calvert 2013; O'Rourke and Crowley 2013; Rabinow and Bennett 2009a) and “embedded” (Felt 2014; Fisher 2006) interactions with humanities scholars and social scientists. Others emphasize the need for deliberation among scientific experts, either with citizens and stakeholders (Funtowicz and Ravetz 1993, 2001; Kleinman 2000; Scolve 1995) or with a specific focus on developing the reflexive capacities of scientists (Mejlgaard et al. 2012; Gibbons et al. 1994). Of course, these two areas of emphasis can overlap (Barben et al. 2008; Macnaghten, Kearnes, and Wynne 2005). Calls for integration can resemble calls for building capacities such as effective communication within scientific teams (National Academy of Sciences, Committee on Facilitating Interdisciplinary Research and Committee on Science Engineering and Public Policy 2004, 2005)² or responsiveness of techno-scientific programs to public values (Fisher, Mahajan, and Mitcham 2006; Stilgoe, Owen, and Macnaghten 2013; Taebi et al. 2014). It should also be noted that while some calls target applied science and engineering design, in order to shape technological and innovation outcomes, others emphasize the importance of engaging the reflexive capacities of expert practitioners themselves, including within socialization and training sites such as laboratories that are focused solely on basic or fundamental science (Gjefsen and Fisher 2014).

A groundswell of science and technology policies throughout the industrialized world, including high-level prescriptions in the USA and Europe, have recently mandated socio-technical integration in an attempt to address societal concerns associated with public investments in science and innovation (Fisher 2011; Macnaghten, Kearnes, and Wynne 2005; Rodriguez, Fisher, and Schuurbiers 2013). This shift in policy rhetoric has been accompanied by a shift in program funding that moves away from the parallel research exemplified by the ELSI program of the Human Genome Project toward a more integrated approach that attempts to bring ELSI research to bear on technical practice (Fisher 2005; cf. Balmer and Bulpin 2013; McCain 2002; Rabinow and Bennett 2009b; Rodriguez, Fisher, and Schuurbiers 2013).

This purported shift toward integration is conspicuous in the US National Nanotechnology Initiative's requirement for “integrating research on societal, ethical, and environmental concerns with nanotechnology research and development” (US Congress 2003; cf. Bennett and Sarewitz 2006; Fisher and Mahajan 2006). This law led to the explicit incorporation of societal research into nanotechnology network and center training programs (Viseu and Maguire 2012; Patra 2011) and into laboratory research (Fisher 2007; Tuma 2013); it also authorized the funding of two large societal research centers, one of which explicitly adopted socio-technical integration as part of its strategic vision (Guston 2014). European policy prescriptions for the “harmonious societal integration of new scientific and technological knowledge” (European Commission 2007) ostensibly extend integration to a wider range of research and closely involve the social sciences in numerous ways (Hackett 2014). Indeed, references to socio-technical integration within formal European solicitations have increased in both absolute and relative numbers over a recent 12-year period, with social scientific and humanistic research increasingly operating as integral components of core science and engineering activities (Rodriguez, Fisher, and Schuurbiers 2013). Alongside the USA and Europe, comparable policies have also been issued in the UK, Canada, Norway, Belgium, and the Netherlands, and have been proposed for a variety of technological areas, including genetics, synthetic biology, neurotechnology, converging technologies, and geo-engineering (Fisher 2014; Fisher and Rip 2013; Ommer et al. 2011; Owen 2014; Roco et al. 2013).

Justifications for engaging expertise in an integrative manner can echo those for technology assessment, such as working toward “better technology in a better society” (Schot and Rip 1997) or “enhancing linkages between innovation and societal action in ways that can add to the value and capability of each” (Guston and Sarewitz 2002). Often, calls for expert engagement represent science and technology as always inflected by a particular set of social interests, values, and assumptions (cf. Bijker 1995), arguing that these values are typically those of elite social groups (Irwin 1995) or are shaped by cultural and institutional arrangements that are taken for granted (Fountain 2004; Jasanoff 2005). In these cases, various explicit themes emerge, such as power (van Oudheusden 2014; Stirling 2008), ethics (Boenink 2013; Khushf 2007; Rabinow and Bennett 2009b), social justice (Fehr 2011), environmental sustainability (Sweet et al. 2014; Voss, Bauknecht, and Kemp 2006; Wiek et al. 2014), and democratic values (Guston 2014; Macnaghten, Kearnes, and Wynne 2005).

As stated, despite the diversity of values that motivate calls for integration, most implicitly or explicitly acknowledge limits of specialization. Thus, technical expertise comes across as bounded in one way or another by its specialist focus. To make up for its limitations, expert practices either require additional specialist expertise or broader awareness of contextual complexity. In both cases, calls for integration suggest that the power of the technical gaze to frame problem and solution spaces simultaneously gives rise to blind spots that can reduce, even undermine, its legitimacy and effectiveness. Responding to these blind spots requires either building the reflexive capacity or supplementing the disciplinary capacity of expert practitioners. Calls for enhancing reflexive capacity – or the reflexivity rationale – seek to productively alter, even disrupt, scientific practices from within by introducing new contextual elements or awareness (Fisher, Mahajan, and Mitcham 2006; Wynne 2011). For example, in responsible innovation, socio-technical integration is seen as a mechanism for embedding societal reflection within existing research and innovation processes (Owen, Macnaghten, and Stilgoe 2012). Likewise, in anticipatory governance, integration is aimed at “long-term reflective capacity building” (Guston 2014). By contrast, calls for supplementing disciplinary capacity – or the interdisciplinary rationale – seek to support existing practices by helping scientific and other expert teams achieve their own stated goals. These calls tend to focus more on aptitudes thought to be relevant for effecting socio-technical integration, including studies of collaborative practice (Salazar, Lant, and Kane 2011), interdisciplinary team facilitation (Morse et al. 2007, Salazar et al. 2012), and effective technical communication (O'Rourke et al. 2013).

2.3. Theoretical developments

Socio-technical integration calls and approaches are sympathetic to constructivist epistemology and to work within the field of STS more generally. Thus, many are concerned with the nature and distribution of expertise; the distance between expertise and citizens; the problems that disciplinary paradigms create for conducting, applying, coordinating, and communicating research; and the problem of ensuring societal responsiveness and mutual assurance. In each case, the problem can be seen as one of recognizing that the social context of specialized knowledge production both enables and constrains the life worlds of actors. In theorizing the interactions that collaborative integration creates, John Dewey's theory of inquiry is influential as are two more recent approaches within STS: Studies of Expertise and Experience (SEE) and trading zones.

Mutual learning is important in the case of reflexive integration approaches, which refer to theories of inquiry and learning to both describe and prescribe their practices. Several communities of integration draw explicitly upon Dewey (Fisher and Schuurbiers 2013; Krabbenborg 2013; Rabinow and Bennett 2012) and on reflective learning more generally (van de Poel and Doorn 2013). For instance, midstream modulation posits analytically distinct stages of human agency in relation to its social material interactions, emphasizing reflexive awareness as the focal point for intervention in decision processes (Fisher and Mahajan 2010; Fisher, Mahajan, and Mitcham 2006). Here, collaborative humanistic and social science engagement can intensify the experience of what Dewey called a “problematic situation” by technical experts, triggering “inquiry” and “encouraging both first- and second-order reflective learning” (Schuurbiers 2011; cf. Fisher and Schuurbiers 2013).

SEE and trading zones highlight capabilities they deem essential for integration. In the case of SEE, Collins and Evans (2002, 2007) develop a typology of expert knowledge in which they make a distinction between “contributory expertise” and “interactional expertise”. Contributory expertise is acquired by successful socialization into a given community of practice, with contributory experts defined by the mastery of both the language and practice of that domain. Interactional expertise is also acquired by socialization, but only in the language of the domain. The importance of this distinction for collaborative socio-technical integration is that it provides a way of understanding how different groups of experts can come to a shared understanding of each other's domains of practice without necessarily being able to perform each other's work (Collins, Evans, and Gorman 2007; Gorman 2002; Plaisance and Kennedy 2014; Selinger, Dreyfus, and Collins 2007). It explains, for example, how a social scientist or humanities scholar can understand the work of the researchers in the laboratory, how those researchers can understand the work of the other specialists, and how managers of large projects can interact successfully across a diverse range of expertises (Collins 2004, 2011; Collins and Sanders 2007; Selinger and Mix 2004). The current test for the acquisition of interactional expertise is through an “imitation game” that takes the form of a modified Turing Test in which an expert judge sets questions that are answered by another expert practitioner and the person be tested. By examining what judges ask, how different players respond to these questions, and how these answers are judged by contributory experts, it is possible to explore the content, level, and distribution of interactional expertise across social groups (Collins et al. 2006; Collins and Evans 2014; Evans and Crocker 2013). In general, players with higher levels of interactional expertise would be better able to contribute to socio-technical integration as they would be able to move more easily between the different expert communities.

Theoretical work around trading zones examines the shared arenas formed when individuals belonging to different fields with incommensurable languages and practices negotiate terms of exchange. Such exchanges are thought to be necessary when disciplinary cultures have to share knowledge, resources, and time to accomplish a goal such as inventing radar or new particle detectors (Galison 1997). Through in-depth cases studies, trading zones researchers illuminate how cultures, such as different academic or industrial disciplines, trade around a jointly created boundary system, with interactional expertise being one of several possibilities (Collins, Evans, and Gorman 2007; Gorman 2002, 2010). By examining how different kinds of collaboration succeed or fail, trading zones research can be used to overcome the incommensurability that impedes effective integration. It is important to note that not all kinds of trading zones can be classified as examples of collaborative socio-technical integration, nor are all forms of collaboration accurately described as trading zones.

2.4. Criticism

The justifications, practices, and outcomes of integration have their own sets of limitations to contend with. Above all, criticism of collaborative approaches to integration points to their potential loss of critical distance (Guston 2014; Nordmann and Schwarz 2010; Valve and McNally 2013). Closely related to this are concerns over the capture of social science and humanities scholars and their loss of agency (Doubleday and Viseu 2010; Felt 2014; Jasanoff 2011; Rodriguez, Fisher, and Schuurbiers 2013). For instance, Valve and McNally suggest that collaborative relationships come with the cost of social scientific “loyalty” to the scientific or technological trajectories envisioned by experts. Integration can also run the risk of overlooking historical and theoretical perspectives (Jasanoff 2011) and of failing to appreciate the already normative and reflexive dispositions of scientific experts (Caudill 2009; Thoreau 2011). Finally, there are scholars who question the outcomes of integration. Some find them forthcoming but limited to a narrow set of actors or sites (Krabbenborg 2013; Wynne 2011); others find few or no pertinent integration outcomes (Rip 2009; Taebi et al. 2014)³; and still others find that established power dynamics and asymmetries preclude or undermine meaningful integration in the first place (Doubleday and Viseu 2010; van Oudheusden 2014; Rabinow and Stavrianakis 2013).

3. Communities of integration

In this section, we highlight the practice of collaborative socio-technical integration by focusing on four communities of practice chosen to illustrate the different dimensions of the framework we propose: Human Practices, Socio-Technical Integration Research (STIR), the Toolbox Project, and Value Sensitive Design (VSD). We briefly describe each community in terms of their contextual and collaborative stance (forms), main methods (means), and motivating rationales (ends), summarizing these descriptions in Table 1.⁴

• Human Practices pursues collaboration – including the mutual definition and adaptation of problems, modes of inquiry, and problem-solving approaches – among anthropologists, biologists, and ethicists midstream during innovation processes; as such, it is a “post-ELSI” approach that moves beyond downstream societal implications, documenting the “ramifications of research as they unfold” (Rabinow and Bennett 2009b). Human Practices “poses and reposes” (Rabinow and Bennett 2009b) questions of how natural scientific research is ramifying and contributing to the development of the near future, often by resorting to “frank speech” (Rabinow and Bennett 2012). As an intervention, it designs and sustains interactions between multiple disciplines that consider the possible futures enabled by contemporary equipment (such as emergence and remediation) and their consequences for the metric of collaboration, mutual flourishing. Application and development of the approach has primarily taken place in and around two synthetic biology laboratories (Rabinow and Stavrianakis 2013).

• STIR is a platform for situated and ongoing dialogue between an “embedded humanist” and scientific experts. Aimed at midstream modulation, dialogue is guided by a “decision protocol” (Fisher 2007) and ideally moves from shared descriptions of social processes to collaborative inquiry. STIR seeks to probe the societal responsiveness of expert practices through productively disrupting their routines and expectations. STIR “laboratory engagement studies” conducted in university and industrial labs across a dozen nations correlate these probing activities to the transformations of scientific practices by practitioners themselves (Fisher 2007; Fisher et al. 2010; Fisher and Schuurbiers 2013; Flipse, van der Sanden, and Osseweijer 2013, 2014; Schuurbiers 2011). Thus, as a research program, STIR 1.0 probes the conditions and capacities for broadening socio-technical integration while, as an intervention, STIR 2.0 attempts to exercise these capacities deliberatively and in light of multiple normative commitments (including ethical reflection, sustainability, and democratic governance).

• The Toolbox Project studies how epistemic reflection and sharing can encourage mutual consideration of underlying research assumptions and worldviews (Eigenbrode et al. 2007). The approach is based on the use of a modular survey instrument to structure dialogue among cross-disciplinary collaborators in a facilitated workshop. Each module opens with a “core question” that announces the focal issue of that module followed by several “probing statements” that highlight aspects of the focal issue. The workshop setting is designed to facilitate a structured dialogue in which participants articulate their own research worldviews and exchange disciplinary perspectives with one another. The goal is to increase self- and mutual understanding, inducing improved team cohesion and scientific communication. The project has conducted workshops in academic and non-academic contexts with various groups, including transdisciplinary teams involving stakeholders (Crowley et al. 2010; O'Rourke and Crowley 2013; and Looney et al. 2013).

• VSD is “a theoretically grounded approach to the design of technology that accounts for human values in a principled and comprehensive manner throughout the design process” (Friedman et al. 2013). VSD explicitly broadens the term “value” beyond financial or efficiency values, engaging with those things that “people consider important in life” (Friedman, Kahn, and Borning 2008). Three types of investigation are used to consider the values at hand: conceptual investigation considers the values, stakeholders, trade-offs, and relative weightings of values involved, as well as key terms, definitions, and issues. Empirical investigation then examines differences between values articulated and demonstrated in practice. Finally, technical investigations examine the underlying systems and embedded values within those systems. These three modes of investigation are ideally used proactively to influence design of technology, iteratively explore a holistic set of values, and consider a wide range of stakeholders. VSD has been applied primarily in the context of computer and information technology.

Table 1. Forms, means, and ends of collaborative socio-technical integration exemplars.

Community	Forms	Means	Ends
Human Practices	Collaborative study, inquiry, and conversation on contemporary technological developments	Ongoing, participant-observation between biologists, ethicists, and anthropologists	Design, develop, and sustain collaboration following principles of emergence, flourishing, and remediation
STIR	Collaborative description and inquiry embedded within expert practices	Regular, situated semi-structured dialogue between embedded humanist and lab practitioners	Heighten reflexive awareness in order to exercise deliberative and responsive capacities
Toolbox Project	Workshop-based approach that aims to elucidate native socio-ethical dimensions	Structured dialogue method that reveals epistemic, metaphysical, and axiological assumptions	Facilitate improvements in scientific communication and collaboration
Value Sensitive Design	Additional considerations and methods during a design process to address value-based design factors	Conceptual, empirical, and technological modes of “investigation”	Construct technologies that better serve values and stakeholders by including concerns in design process

4. Mapping the integrative field

Socio-technical integration involves a host of conceptual and normative issues. We attempt to map this diversity among approaches to socio-technical integration using the exemplars described above.⁵ We start by considering each approach in terms of three characteristics: the forms of collaborative integration, meaning how the approach positions itself in relation to the expert practices it engages, its means, understood in terms of the nature of its own practice, and its ends, represented here by the goals and values that motivate the integration approach and justify its pursuit.⁶ We then collapse these different features into two orthogonal dimensions – values and capacities – upon which each approach is positioned in order to illustrate and facilitate more general comparisons.

4.1. Forms

Integration can assume a variety of forms, depending on which ideas or explanatory frameworks are privileged and which participatory relations are put in play by a particular community of practice. Here, we take the form of an integration approach to be revealed by how it is related to its domain of inquiry along two dimensions: (1) how it conceptualizes the relationship between the societal contexts it focuses on and the practices it engages, and (2) how it organizes the relationship between its own practices and those it engages. With respect to the first, we ask to what extent the approach represents the societal dimensions as latent within and native to the engaged practices, in which case it presents itself as engaged in clarifying (STIR and Toolbox) already intertwined dimensions; and to what extent it represents them as absent or otherwise missing from the practices, in which case the approach may tend to present itself as introducing (Human Practices and VSD) something new. The second dimension concerns whether the approach is positioned as somehow embedded within the site of the focal practices (STIR, Human Practices, and VSD) or whether it organizes its own separate venue or parallel process (Toolbox). Positionality also concerns whether an approach is independent of (STIR) or dependent upon (Toolbox, VSD, and Human Practices) the focal experts in terms of financial support and/or accountability.

4.2. Means

Consideration of means focuses on the methods and standards an integration approach employs as it operates. Decisions about means highlight what an approach takes to be the most rewarding opportunities, pressing challenges, and perplexing trade-offs associated with socio-technical integration. Attention to socio-technical means will underline best practices and raise practical questions about opportunities and challenges that are amenable to empirical analysis. Given the variety across the communities of practice we have chosen as exemplars, we should expect the means to constitute a contextualized repertoire of different approaches, practical insights, and design considerations. Our analysis of means is also two-dimensional, addressing (1) the degree to which a community is structured in its engagement with its focal practices and (2) the degree to which its techniques are fixed in advance. An approach is structured to the extent that it engages with a practice in a relatively consistent way each time (Toolbox and VSD), and is more unstructured as the type of engagement varies depending on the particular engagement venue or group (Human Practices). Approaches can also combine elements of fixedness and flexibility, having a semi-structured method that is topically oriented and therefore locally adapted on each occasion (STIR).

4.3. Ends

The ends of collaborative integration are the values and norms that motivate and justify an approach's activities. These include helping scientific teams become more efficient and aligned, enhancing the reflexive and socially responsive processes of scientific research and technology development, and working to make outcomes more equitable, democratic, or sustainable. For analytical purposes, we map the ends of collaborative integration along two dimensions: (1) instrumental-normative and (2) facilitative-substantive. The instrumental-normative dimension tracks the extent to which the integrative effort is conducted for the sake of ends that are valuable as a route to other ends (e.g. economic security) or in themselves (e.g. human dignity). The facilitative-substantive dimension traces the degree to which an approach emphasizes aiding the focal practice in achieving the ends it has set for itself or contributing new and better ends to the focal practice. Some integrative approaches are primarily instrumental in that they aim to enhance communication as a means to achieving broader project goals (Toolbox); others attempt to engage in co-labor around specific ethical issues and are more explicitly normative in seeking to advance goals, ends, or principles taken to be ends in themselves (Human Practices). We situate commitments to broadening values in technical practice midway between these extremes (STIR and VSD). Along the second dimension, some approaches are more substantive insofar as they seek to build general capacities for reflection and deliberation on both ends and means (STIR and Human Practices), while others are focused on facilitating boundary-crossing activities (Toolbox) or advocating for specific value-adjustments (VSD).

5. A comparative framework

To represent several of the most salient differences, we offer a two-dimensional matrix of values and capacities (Figure 1). The values dimension represents a range of possible ends/forms combinations, from those that seek to enhance the existing goals and commitments of the focal expert practices (“native values”) to those that emphasize introducing new goals and commitments for expert consideration (“alternative values”). The capacities dimension represents a range of possible means/forms combinations, from those that seek to introduce additional knowledge, content, and resources (“alternative capacities”) to those that seek to enhance latent resources and capacities (“native capacities”).

Figure 1. Comparing communities of socio-technical integration by forms, means, and ends.

To illustrate the features suggested by the four quadrants, we highlight aspects of the exemplars. Thus, Human Practices appears in the upper left because of its emphasis on bringing an ethical disposition that is otherwise seen as foreign or lacking into the natural science research environment, and because this is attempted largely through its own conceptual expertise by way of its own posited ideal (mutual flourishing). VSD is similar to Human Practices in that it emphasizes the introduction of its own expertise, and of additional considerations and constraints; however, since this is often explicitly done to support existing project goals, we locate it closer to native values than we do Human Practices. In the lower right, and also close to native values is the Toolbox Project, which attempts both to enhance local communication capacities and to support local goals, without explicitly calling attention to its own content and expertise. Finally, we locate STIR in the upper right, since it seeks (similar to Toolbox) to enlarge the experience of local experts, but does so (similar to Human Practices) with respect to its own posited ideal (deliberation), hence simultaneously tapping into and broadening native values.

The exemplars are meant to illustrate the various norms and postures within the integrative field more generally. Accordingly, the various quadrants suggest modes of collaborative integration – where “mode” is expressive both of the way integration is pursued and of the envisioned outcome – that can illuminate other approaches not specifically analyzed here (Figure 2). Thus, the upper left quadrant suggests a commitment to reforming expert technical practices through the introduction of alternative values, voices, and methods. While we have chosen Human Practices to represent this mode of integration, it is also characteristic of much work in action research. By contrast, the lower left focuses on augmenting local expert practices and decision-making with additional expert practices, supplementing skills and considerations while working within local problem framings. We find this to be the case not only with VSD, but also with other integrative approaches that emphasize their own expertise, for instance bioethics consultancy and some SRPoiSE projects. The lower right represents an emphasis on skill-based learning aimed at facilitating teamwork, working with and within local problem framings but in this case without an explicit interest in broadening or supplementing either values or capacities. This is characteristic of the Toolbox Project and of much skill building and aptitude enhancement, such as those seen in approaches to the science of team science. Finally, the upper right suggests a mode of problematizing existing practices and commitments: rather than working toward direct reconfiguration and reform, this mode seeks to enhance second-order learning with the assumption that experts will voluntarily broaden and reframe their own problem definitions and adjust their own practices. We have chosen STIR to represent this mode of integration, which is also evident in approaches such as constructive technology assessment.

Figure 2. Idealized modes of socio-technical integration.

The framework reflects the diversity of methodological tendencies, modes of operation and commitment, and normative justifications in approaches to collaborative socio-technical integration, illustrated through four exemplar approaches. We recognize that the framework does not necessarily capture all aspects of integration (e.g. it does not indicate whether approaches are more or less embedded or more or less structured); furthermore, individual approaches are likely to be richer and more complex than their assignment to an individual quadrant may indicate (e.g. collaborative integration is rarely static and individual approaches may include a breadth of norms and positions⁷). With these caveats, we find this exercise to be valuable in that it suggests how collaborative engagement of expertise can be guided by very different justifications and can engage different aspects and modes of expertise. For instance, we expect that approaches that invoke the reflexivity rationale tend to appear on the right side of the figure insofar as reflexivity, learning, and communication are ultimately rooted in local and native capacities, whereas those that stress the interdisciplinarity rationale will be more concentrated in the left half of the figure insofar as they place more emphasis on the introduction of new forms of expertise into the focal practices. Meanwhile, approaches in the lower half of the figure focus less on questioning pre-determined goals and more on achieving them effectively and efficiently. By contrast, approaches in the top half of the diagram seek to develop new values within expert practices.

The framework can also help anticipate and explain how different collaborative approaches to socio-technical integration can find themselves in conflict or disagreement with the experts in question (as when alternative values are meant to challenge native values), despite appearing to share similar commitments. For example, and drawing on the SEE concepts identified earlier, we suggest that collaborations in the left half of the diagram, particularly those in the upper left/reform quadrant, are likely to be more challenging than those in the right half, with those in the lower right/facilitate quadrant being the least complex to manage. The explanation is that, depending on the nature of the collaboration, the technical experts in question will have more or less learning to do, without the incentive of linking this additional work to their own prior commitments. Such an explanation can help illuminate the source of potential difficulties and inform integration design choices in light of what is practicable. Furthermore, seeing expertise as part of a culture and acquired through socialization may help explain why efforts to promote more radical forms of reflexivity or interdisciplinarity are difficult; if responding to alternative values is akin to adopting a new paradigm, it may involve “unlearning” as much as “learning” as domain experts are challenged about things they previously took for granted as part of their contributory expertise (cf. Evans and Marvin 2006).

Finally, the normative commitments that characterize different collaborative approaches to integration – all of which potentially may be seen as supporting responsible innovation – may tend more toward disruption or reinforcement of native values, goals, problem framings, and programs. Significantly, our analysis suggests that all integrative approaches seek to address the limitations of technical expertise – and more generally, the limitations of innovation actors and processes – but they proceed from different diagnoses of this limitation and prescribe different remedies. Comparing the quadrants diagonally helps bring certain tensions to the forefront. A central normative tension thus opens up between the reform mode of integration, which assumes a fundamental lack in the program of those experts with whom they seek to collaborate, and the facilitating mode, which is fundamentally aligned with the goals of programs it chooses to engage in. A methodological tension opens up between the augmenting mode of integration, which stresses the importance of other participants and tools, and the problematizing mode, which relies on experts correcting and modulating their own practices based on heightened contextual awareness and second-order reflective learning.

6. Concluding reflections

We have argued that collaborative socio-technical integration is a form of expert engagement that is distinct from related fields such as transdisciplinarity and stakeholder engagement in three ways: it (a) addresses variously conceptualized socio-technical divides pertaining to expertise, (b) operates closely within expert practices, and (c) seeks to meaningfully transform those practices. Justifications for socio-technical collaborations are implicitly, often explicitly, tied to the inherent limitations of expertise, which are implicated in how experts relate to broader societal dimensions of their work. Socio-technical integration is thus the intellectual and practical work of inflecting expert values and capacities in order to more responsibly align science, technology, and innovation with their broader societal contexts.

As we have presented it, the field is populated by a diversity of individual researchers and communities of practice that collectively seek to explore and foster such alignment. In order to compare relational, methodological, and normative divergences among diverse integration approaches we presented a framework and illustrated it with exemplar communities of integration. The framework suggests that while collaborative approaches to socio-technical integration all seek to orient expert practices more explicitly to under-specified societal contexts, they do this in the comparatively different modes of reforming, augmenting, facilitating, and problematizing expertise. These modes can be used to compare approaches and foreground their synergies and their tensions. For instance, methodological approaches that seek more directly to introduce alternative expertise can be distinguished from those that seek more indirectly to cultivate learning within native expertise. Similarly, normative commitments to broadening expert practices contrast with those focused on reinforcing them.

By extension, the framework suggests that all four quadrants can potentially support responsible innovation. While socio-technical integration is frequently cited as an established mechanism of responsible innovation, the diversity of integrative approaches imply similarly diverse orientations to broader programs of knowledge production, technoscience, and innovation in society. We suggest that the more critical an approach is of the innovation projects it engages, the more it will seek to reform or problematize their fundamental assumptions and values. By comparison, more progressively inclined integration approaches will seek to augment innovation projects, as if it were completing them with missing elements. Finally, approaches that are aligned with the goals and values of projects they have chosen to engage will seek to optimize their operational processes to make them more effective at achieving their own stated goals. The point then is not that one type or mode of integrative approach is closer to responsible innovation, but rather that each one can be relatively strong or weak in terms of its support for responsible innovation. In this respect, certain modes of integration may be more prone to specific shortcomings: modes that emphasize alternative expertise are more likely to reify divisions of moral labor (in the augmentation mode) or run the risk of ejection (in the reformist mode); while those that tend to accommodate native values or capacities are likely more easily co-opted (in the facilitative mode) or resisted (in the problematizing mode).

The extent to which an integrative approach supports responsible innovation may depend in part on how effectively it balances an awareness of the different modes of integration and judiciously incorporates this into its own practices. In particular, it is helpful to consider how a given integrative mode enlists its opposite mode as depicted in the framework. Consider the case of the problematize-augment opposition: A problematizing mode that primarily disorients the experts it engages will be less effective from the standpoint of responsible innovation than one that allows this disorientation to become a resource for meaningful reflection and productive decision-making moving forward. In this case, the integration agent operating in the problematizing mode may be aided by considering how his or her own tools and knowledge might illuminate enabling questions and practical pathways, something that could be aided by operating temporarily in the opposite mode of augmentation. By contrast, operating exclusively in the augmentation mode runs the risk of reinforcing a problematic division of moral labor, in which technical experts are no longer responsible for the societal and ethical dimensions of their work. To temper this, it may be helpful to incorporate aspects of the opposite mode of problematization, which focuses on reflective and responsive capacity building among technical experts. The same could be said for the reformist-facilitative opposition.

We emphasize that the modes of practice we have in mind are not monolithic and allow for nuanced and complex modes of engagement. In other words, a given mode can encompass critically normative attitudes and instrumentally supporting roles at the same time. The important question of how a given integration approach chooses to balance its manifold relations with native values and native capacities has a more general correlate in how the integration approach is itself balanced with additional activities and structural components that are taken to be more broadly indicative of responsible innovation. These include mechanisms meant to advance inclusiveness, anticipation, reflection, deliberation, and responsiveness at institutional levels that are significantly broader than the expert practices themselves (Owen et al. 2013). Ultimately, the choice and evaluation of collaborative socio-technical integrative modes and mechanisms will require a critical assessment of the social, ethical, and democratic legitimacy of a given innovation program or project as well as a clear sense of how integration is meant to relate to broader goals and criteria of responsible innovation.

It would therefore be helpful for future research on socio-technical integration to develop a more fine-grained account of the diverse techniques employed by various approaches. For instance, whether an approach operates in a manner that is institutionally embedded, institutionally fused, or resembles that of a consultant may play an important role in understanding the range of its application and its comparative strengths and weaknesses. Future studies of integration should also be concerned with evaluating the efficacy of its presumed interventions, especially in light of the broader justifications invoked for them in the first place, and will ideally take a comparative view of diverse integrative approaches. There is also a significant need to develop theoretical accounts of socio-technical integration, both those that seek to explain it and those that seek to critically inform its norms, practices, and reception.

The value of framework for collaborative socio-technical integration lies in its ability to accommodate a considerable diversity of motivations, conceptions, methods, and relationships currently developing around the collaborative engagement of expertise. Taking this diversity into account can lead to more transparency in justification, design, and evaluation of collaborative efforts aimed at socio-technical integration for engaged experts, public sponsors, and affected stakeholders. We do not envision a field that aims at theoretical, methodological, or normative consensus, but rather one that nurtures a mutually enriching discussion among engaged scholars from different disciplinary and problem-oriented traditions around issues that characterize the boundaries of the field. Collaborative socio-technical integration will thereby be more likely to develop with due regard for the cultural and institutional structures that it strives to work within, improving its efficacy without over-promising what it can deliver and mindful of potentially incongruous and even counter-productive outcomes. In our view, this requires a reflexive, critically balanced, on-going, and dynamic interplay between the native and alternative values and capacities that are implicated in any given expert site or innovation project.

Acknowledgements

The authors would like to thank Sharon Ku, Stephen Crowley, J. Britt Holbrook, Rune Nydal, Kathryn Plaisance, Jathan Sadowski, Andrew Berardy and Ben Wender for providing valuable feedback and suggestions during early stages of the planning process.

Funding

This material is based on work supported by the National Science Foundation [grant numbers 0823058, 0849101, 0937591 and 1140190]; USDA National Institute of Food and Agriculture [grant number MICL02261]; and the European Research Council [grant number 269463].

Notes on contributor

Erik Fisher is an Assistant Professor in the Consortium for Science, Policy & Outcomes and Associate Director for Integration with the Center for Nanotechnology in Society at Arizona State University. He also leads a research thrust in Real-time Technology Assessment and directs the NSF-sponsored STIR (Socio-Technical Integration Research) project (https://cns.asu.edu/research/stir).

Michael O’Rourke is Professor of Philosophy and faculty in AgBioResearch at Michigan State University (MSU). His research interests include environmental philosophy, the nature of epistemic integration and communication in collaborative, cross-disciplinary research, and the nature of linguistic communication between intelligent agents. He is Director of the Toolbox Project, an NSF-sponsored research initiative that investigates philosophical approaches to facilitating interdisciplinary research (http://toolbox-project.org/).

Robert Evans is a Reader in sociology at the Cardiff School of Social Sciences. Together with Harry Collins he is the co-author of the ‘Third Wave of Science Studies’ paper (Social Studies of Science, 2002) and Rethinking Expertise (University of Chicago Press, 2007). His current work focuses on the nature of expertise and, in particular, the development of the Imitation Game as a method for comparative research.

Eric B. Kennedy is a Ph.D. student at Arizona State University’s Consortium for Science, Policy & Outcomes. His research focuses on cross-sector and cross-community collaboration in complex sociotechnical systems with the goals of advancing a pluralistic and inclusive approach to expertise that includes scientists, industry, policy-makers, and marginalized communities. Kennedy is a 2014 Generation Fellow with the Breakthrough Institute, a 2014 Equinox Fellow with the Waterloo Global Science Initiative, and a graduate of the University of Waterloo’s Bachelor of Knowledge Integration.

Michael E. Gorman is Professor in the Department of Science, Technology & Society at the University of Virginia and a former National Science Foundation Program Director. With NSF support, Gorman authored Ethical and Environmental Challenges to Engineering (Prentice-Hall, 2000; with Patricia Werhane) and edited Scientific and Technological Thinking (Lawrence Erlbaum Associates, 2005) and Trading Zones and Interactional Expertise: Creating New Kinds of Collaboration (MIT Press, 2011).

Thomas P. Seager, an Associate Professor at Arizona State University's School of Sustainable Engineering and the Built Environment, leads a project funded by the National Science Foundation that applies game theory to develop new strategies for teaching ethical reasoning skills relevant to sustainability to science and engineering graduate students. Seager also conducts research related to environmental decision analysis and the life-cycle environmental impacts of alternative energy technologies.

Notes

1. The present paper is an outgrowth of, and to some extent a reflection of, the recently formed Communities of Integration Network (COIN), which has held meetings at Arizona State University (May 2013) and Waterloo University (June 2014). The paper is based on ideas initially conceived in a project proposal (Fisher, Seager, Gorman, O'Rourke, and Julie T. Klein).

2. For instance, the National Academy of Science (NAS 2005, 81) recommends that “Researchers … desiring to work on interdisciplinary research, education, and training projects should immerse themselves in the languages, cultures, and knowledge of their collaborators”.

3. Even in the case of demonstrable outcomes, critics would still be right to question whether the outcomes are worth the effort, especially considering that integration can require a significant investment of time on the part of both the integration agent and the collaborating experts. See work by the STIR project (Fisher 2007; Flipse et al. 2013), which attempts to establish an empirical basis for considering this potential criticism.

4. The landscape of approaches to collaborative socio-technical integration is broad and diverse. There are a number of additional scholarly communities that include diverse methodological approaches to socio-technical integration in their research purview, including but not limited to science of team science (Falk-Krzesinski et al. 2011), constructive technology assessment (Schot and Rip 1997), ethics consultation (Cho et al. 2008; Fletcher and Siegler 1996), integration and implementation sciences (Bammer 2013), socially relevant philosophy of/in science and engineering (SRPoiSE; Fehr and Plaisance 2010), and field philosophy (Frodeman 2010).

5. The choice of these four exemplars to represent the four quadrants of our framework is inevitably imperfect and is offered here and throughout the next section primarily for illustration purposes.

6. Of course, these characteristics are quite fluid and varied, and our association of any specific characteristic with a given approach is somewhat of an oversimplification, especially since a given integrative project may embody multiple instrumental, interpretive, and normative roles (Jasanoff, 2011) that could also shift over time. Similarly, in enhancing communication, an integrative approach may not only be contributing to the instrumental selection of means for achieving pre-existing ends; it may also allow for substantive opportunities to identify and reflect upon conflicting or diverging normative goals (Stirling 2008).

7. We have acknowledged such internal dynamics and the fact that one integration approach is not necessarily confined to a single quadrant using clouds surrounding each of the four exemplars. For instance, in seeking to articulate segregated and otherwise excluded values that are nevertheless latent in local expert practices, STIR works with native (scientific research) values as a route toward more consciously addressing alternative (broader societal) values; this results in both native (first-order reflective learning) and alternative (second-order reflective learning) outcomes (Schuurbiers 2011; Wynne 2011).