Transdisciplinarity in practice: aims, collaboration and integration in a Swedish research programme

Abstract

This article discusses the practice of crossdisciplinarity in the context of future-oriented sustainability studies. Much research into crossdisciplinarity has concentrated on programmatic and epistemological questions. In this study, we focus on research practice and efforts to realize transdisciplinary aims across a research programme. We use the Swedish Future Forests programme as a case study and explore its aims, forms of collaboration and level of conceptual integration. The study demonstrates that efficient integration requires organizational settings able to support the development of a common conceptual framework. To achieve this, the aims and forms of collaboration and the means of integration ought to be consistent. Far-reaching integration and short-term instrumental objectives may be difficult to combine because integration requires intellectual space, specific boundary settings and time. Short-term instrumental objectives may also hamper open and reflexive discussion of alternative pathways to sustainability and of how participating actors shape the research process. These insights may help researchers and participating actors to design research programmes that enable a realization of their transdisciplinary ambitions.

Keywords::

• crossdisciplinarity
• transdisciplinarity
• sustainability
• research practice
• integration
• forests

1. Introduction

Over the last few decades, discussions of new ways to conduct scientific research have proliferated around the world. Long-standing concerns about the increasing specialization of science and debate about the relationship between the empirical sciences and societal values have driven criticism of dominant scientific ideals (Hirsch Hadorn et al. 2008; Weingart 2010). Inter- or transdisciplinary research approaches have been increasingly motivated by demands for applicability and social relevance, for example, in environment-, development- and industry-related fields (Schild and Sörlin 2002). An additional driver is the recognition of complex environmental problems with high levels of uncertainty about outcomes and impacts (Scoones et al. 2007).

In response to both financing requirements and challenges related to sustainability, the research community is increasingly exploring ways to cross traditional disciplinary boundaries (see, e.g. Pohl 2005; de Boer et al. 2006; Maasen et al. 2006; Hirsch Hadorn et al. 2008; Frodeman et al. 2010). This advance of cross- and transdisciplinary approaches can now be considered a global trend (Mobjörk 2009).

Much research into cross- and transdisciplinarity has concentrated on epistemological and conceptual questions (Maasen et al. 2006) while knowledge of its practice and benefits is still limited (Podestá et al. 2013). To fill this gap, we investigate one particular research programme by asking the following questions: What are the stated motives and aims of the research programme? What forms of collaboration are used? What is the level of conceptual integration 2 years after programme initiation? Empirically, we use the Swedish Future Forests programme as a case study, a programme in which we are also participating as researchers. We focus on the programme level, as large programmes, more than individual projects or small-scale local programmes, tend to address a plethora of goals involving not only scientists from diverse fields but policy-makers, businesspeople and other stakeholders (Maasen et al. 2006).

2. Theoretical framework

2.1 Exploring crossdisciplinarity

The literature uses many concepts to denote various kinds of research integration across boundaries (Mobjörk 2004, 2010; Thompson Klein 2004, 2010; Pohl 2005; Pohl and Hirsch Hadorn 2007; Hirsch Hadorn et al. 2008). Terms such as crossdisciplinarity, interdisciplinarity, transdisciplinarity and multidisciplinarity are used in shifting ways. All these terms concern efforts to cross disciplinary boundaries. Many authors refer to disciplines as delimited by differences in theory, method, conceptual framework and institutional setting with various organizational, social and cultural dimensions (Thompson Klein 1996; Mobjörk 2009; Weingart 2010).

Following Mobjörk (2010), we use the term crossdisciplinarity as an umbrella term to denote all research approaches crossing disciplinary boundaries of any kind. Three main approaches are identified: multidisciplinarity, interdisciplinarity and transdisciplinarity. To properly differentiate these approaches from each other, three dimensions are emphasized: the motive underlying the research, the scope of collaboration and the integrative approach. These dimensions emerge from the view that integration alone is insufficient for differentiating the various forms of crossdisciplinarity. A key question is what ought to be integrated, how and by whom. A distinction between ‘critical’ and ‘instrumental’ approaches is accordingly suggested (Thompson Klein 1996; Mobjörk 2009, 2010). In instrumental research across disciplines, the crossdisciplinary activity is considered a means to achieve certain goals, for example, solving a particular problem. In ‘critical’ research approaches, epistemological issues and motives are in the foreground.

According to Mobjörk (2010), multidisciplinarity is described as collaboration between researchers from various disciplines investigating a specific problem from their own perspectives using each discipline's conventional methods. The degree of integration is normally addressed in a subsequent synthesis phase and does not affect the individual approaches shaping the research. The motive underlying the research is mainly instrumental. Interdisciplinarity is characterized by collaboration between researchers from various disciplines, but with the distinction that the research process is jointly established to develop an integrated shared methodological framework and a shared problem formulation. The motives underlying the research may be both instrumental and critical. In transdisciplinarity, this process is extended to include participants from various academic disciplines and from outside academia. The degree of integration is essential, but the forms of collaboration are placed at the centre. As in interdisciplinarity, the motives underlying the research may be both instrumental and critical. Mobjörk (2010) stresses the role of non-academic actors and differentiates transdisciplinary research including all kinds of actors on equal terms (participatory transdisciplinarity) from merely having actors from outside academia respond to the research conducted (consulting transdisciplinarity).

Integration, which is a more-or-less developed element of all forms of crossdisciplinarity, may occur at various stages and levels of the research process, and may result from various forms of interaction between the researchers and practitioners involved. However, most scholars tend to agree that integration across research disciplines involves a degree of sharing, or blending, of concepts and methods, as well as the development of some sort of common conceptual framework (Haugaard Jakobsen et al. 2004; Mobjörk 2004; Smelser 2004; Pohl 2005). Such processes involve the crossing of established boundaries. The construction of boundaries in science and in science–society relationships is extensively analysed in the social studies of science literature (Gieryn 1983). Boundary crossing does not happen automatically but requires boundary work. Several ways to analyse how objects, concepts and institutional conditions may facilitate this process exist in the literature. Star and Griesmer (1989) focus on the role of ‘boundary objects’, that is scientific ‘objects’ which inhabit several social worlds and satisfy the informational requirements of each of them. They show how such objects can be used as means of translation across diverse groups of actors in a museum setting. Mollinga (2010) takes this discussion further and suggests that boundary work in inter- and transdisciplinary settings requires three types of effort: the development of boundary concepts, the configuration of boundary objects and the shaping of conducive boundary settings. Boundary concepts refer to words that function as concepts in various disciplines and perspectives, refer to the same object, phenomenon, process or quality of these, but carry (sometimes very) different meanings in those different disciplines or perspectives. Such ‘loose concepts’ may enable members of distinct professional groups to work together and develop areas of efficient collaboration (‘trading zones’ or ‘pidgin zones’) without obliging them to give up the advantages conferred by their group identities (Löwy 1992). Boundary objects, according to Mollinga, apply to devices and methods that allow action in situations of incomplete knowledge, nonlinearity and divergent interests. Boundary settings refer to the organizational and institutional conditions that foster the development of boundary concepts and objects (Mollinga 2010). Consequently, integration at the conceptual, methodological and epistemological levels cannot be seen in isolation from the organizational structure and forms of collaboration, as these represent different but interrelated forms of boundary work.

In this study, conceptual integration will first be investigated at the level of the individual low-level concepts used by the various research teams. Low-level concepts refer to concepts taken from the disciplines and worldviews to be integrated. In the next step, we group the listed low-level concepts into so-called conceptual clusters (Newell et al. 2005), which are groupings of broader and more general higher-level concepts that capture essential aspects of the human environmental problems at stake.

2.2 Crossdisciplinarity in practice

Whereas the benefits of successful disciplinary integration are quite obvious, the barriers to its realization are not always as apparent. Kloet et al. (2013) argue that large-scale interdisciplinary research programmes funded to boost scientific, economic and societal advancement often have difficulty realizing their goals. Reasons include the monodisciplinary nature of academic research, insufficient interaction with societal actors and funding conditions. Wiesmann et al. (2008) identify a number of stumbling blocks, including questions of participation, power, integration, education, leadership and career building. Some argue that the call for ‘interdisciplinarity’ itself manifests ambivalence: interdisciplinarity is simultaneously recommended and not rewarded by discipline-based career structures (Schild and Sörlin 2002; Smelser 2004). Differences in the use of terminology, methods and research design as well as perceived power hierarchies of disciplines are other barriers (Haugaard Jakobsen et al. 2004). Several authors discuss the barriers to interdisciplinarity from an institutional perspective (Broto et al. 2009; Buanes and Jentoft 2009). Rules, Buanes and Jentoft (2009) argue, are relatively easily changed. Culture and values, in contrast, are not easily altered and doing so requires time. In transdisciplinary research contexts involving actors outside of academia, these challenges are amplified. Sufficient time to build trust and personal relationships is essential (Höchtl et al. 2006; Harris and Lyon 2013; Podestá et al. 2013) and institutional changes in the knowledge-production process may be required (Cornell et al. 2013).

Many of these challenges have been identified in previous studies of Swedish crossdisciplinary research programmes (Brewer 1999a, 1999b; Mobjörk 2004; Pohl 2005). However, the international literature includes more systematic attempts to explore crossdisciplinary research practice in order to define the state of the art (Maasen et al. 2006; Hirsch Hadorn et al. 2008; Frodeman et al. 2010). Hirsch Hadorn et al. (2008) investigate 18 transdisciplinary projects/programmes with respect to seven ‘cross-cutting issues’. Participation and integration are generally identified as major challenges (Elzinga 2008). Pohl et al. (2008) suggest that the challenges of integration often concern creating or restructuring the meaning of theoretical terms to capture what is regarded as relevant in a particular transdisciplinary research context.

Previous studies that explore practice and outcomes at a research programme level emphasize the need for time. Messerli and Messerli (2008) demonstrate that knowledge production has changed from a dominant disciplinary mode to include transdisciplinary components in transdisciplinary human–environment research programmes, but that this process required almost 30 years. Pregernig (2006) argues that even the intra-scientific side of transdisciplinarity requires time to bear fruit, as interdisciplinary integration is difficult to plan in advance and often comes about as a result of long-term processes of communication and negotiation. According to Höchtl et al. (2006), as much as 30% of the working time in their transdisciplinary project was used for internal and external communication. Accordingly, Roux et al. (2010) argue that without a 5–10-year time frame, the considerable upfront costs of transdisciplinary research are unlikely to pay off.

2.3 An analytical framework

Many of the studies cited above discuss general notions of science and knowledge and focus on research foundations or the programme management level. Several address the motives driving research programmes, collaborative efforts or means of integration, but few offer a comprehensive analysis of how these aspects relate to each other in one practical research setting. Drawing on Mobjörk's (2010) dimensions of crossdisciplinarity (aims, scope of collaboration and integrative approach), we explore, through a three-step analysis, how the crossdisciplinary aims of the Swedish Future Forests programme are realized:

• What are the stated motives and aims of the research programme?

• What forms of research collaboration are used in it?

• What are the levels of conceptual integration 2 years after programme initiation?

Table 1 shows how the research questions were operationalized and the methods and empirical sources used in the analysis. The study is based on a content analysis of selected documents and websites and a survey directed to Future Forests research leaders (see Section 3.3). The research questions have guided the selection of documents and the analysis. Text outlining motives and aims, forms of research collaboration as well as the question of integration has been analysed. The survey was used to explore research question number 3 because existing documentation was sparse and did not reflect the level of integration in research practice. We have chosen to rely on written documentation and surveys rather than participatory observation, although we took part in the programme as researchers. This study was initiated after the programme was started and our observations were not systematically documented from the outset. Nevertheless, our experiences have shaped our discussion and conclusions. Assets and problems associated with a situated research position are extensively discussed for example by anthropologists doing research ‘at home’ (see, e.g. Hastrup 1987). The strategies adopted for minimizing the risk for prejudgement and blindfolding are continuous reflexivity and dialogue with peers having an outside perspective. This text and our conclusions have accordingly been discussed with peers outside the research programme.

Table 1 Research questions, methods and empirical sources used in the study.

	Research question	Operationalized research question	Methods and empirical sources
Step 1	What are the stated motives and aims of the research programme?	How are programme aims, visions and contributions described?	Analysis of Future Forests programme plan and Future Forests web site
Step 2	What forms for collaboration are used?	How is Future Forests organized? What opportunities for collaboration exist? What experiences are documented?	Analysis of Future Forests programme plan, MISTRA start-up review and MISTRA final evaluation
Step 3	What are the levels of conceptual integration?	What concepts are used? How are these concepts understood across the programme? What ‘boundary concepts’ exist?	Two-step survey of involved research groups

Note: MISTRA, Swedish Foundation for Strategic Environmental Research; ForSA, Future Forests Centre for Forest Systems Analysis and Synthesis.

3. The case of the Future Forests programme

3.1 Motives and aims

The Future Forests programme is a joint initiative of the Swedish University of Agricultural Sciences, Umeå University and the Forest Research Institute of Sweden. The programme is financed by the Swedish Foundation for Strategic Environmental Research (MISTRA), the Swedish forestry industry and the universities involved. Future Forests phase 1 lasted from 2009 to 2012 and is followed by a phase 2 to be completed in 2016. This study focuses on phase 1. When we refer to ‘Future Forests’ or ‘the programme’ we henceforth mean Future Forests phase 1.

Future Forests addresses a number of complex challenges, including responding to climate change and growing demand for energy and raw materials. As presented on the Internet website, the stated aims of phase 1 were as follows.

• Develop and evaluate new forest management methods and strategies to meet increased, unpredictable and conflicting demands on forest production and ecosystem services by current and future societies.

• Develop an understanding to support strategies and methods for improved and more effective governance within the forestry sector.

• Create a world-leading and long-lasting Centre for Forest Systems Analysis and Synthesis (ForSA) where interdisciplinary research questions can be successfully analyzed and synthesized (Future Forests 2011).

Future Forests was a future study expected to identify alternative ways to manage Swedish forests and the likely effects of these alternatives. One aim was to serve as a flexible and effective interdisciplinary competence platform enabling researchers from various scientific disciplines to meet and collaborate with each other and with practitioners from various sectors of society (Future Forests 2011). According to the programme plan, ‘Future Forests will create knowledge and tools to enable sustainable decisions for the future of one of our most important resources – our forests’ (Future Forests 2009, p. 3). In addition, the programme was expected to generate new knowledge that would strengthen Sweden's international competitiveness (Future Forests 2011).

Most research of Future Forests was organized in 10 different component projects, each addressing a particular theme or perspective. The component projects employed perspectives ranging from political science and history to forest ecology and silvicultural management (see Figure 1), and were expected to produce ‘detailed, high-quality scientific results that can feed into a crossdisciplinary scenario analysis and be directly relevant to our stakeholders’ (Future Forests 2009, p. 3). While much multidisciplinary research was expected to be carried out in the component projects, ForSA was given the main responsibility for integration, synthesis and interaction with external actors. It was expected to become a ‘unique institution with an explicit mission to foster synthesis, turn information into understanding and through effective collaboration alter how science and forest management are conducted’ (Future Forests 2009, p. 3). The programme was described as ‘a creative and intellectually stimulating environment, where different scientific disciplines will merge to address problem areas in a multifaceted setting’ (Future Forests 2009, p. 25). Systems and scenario analyses were seen as means to explore possible futures in an integrated way (Future Forests 2009). The programme was expected to develop ‘the theoretical basis necessary to bridge natural and social sciences’ (Future Forests 2009, p. 16). However, it was also assumed that it would generate novel insights applicable to management and governance in an uncertain future. It was anticipated that the scenarios would act as decision-support tools for stakeholders (Future Forests 2009).

Figure 1 Schematic overview of the programme structure.

Obviously, Future Forests was intended to ‘bridge’ or ‘merge’ various scientific disciplines, which means integration to a level at which disciplinary boundaries are transcended. It was expected to alter how science and forest management are conducted, to develop new forms of knowledge production – in other words, to address epistemological issues. The programme focused on problem solving and had a clear instrumental approach. The aspirations were thus both instrumental and critical. The programme plan expressed the intention to involve practitioners in the research process and aimed to enhance collaboration between discipline-based researchers and between society and academia. Consequently, Future Forests included multi- as well as transdisciplinary objectives.

3.2 Forms of collaboration

Figure 1 shows the organizational structure of Future Forests as presented in the programme plan. A core team of scientists representing the various component projects served as a coordinating body and a panel of practitioners was set up to facilitate participation of non-academic actors. Hence, the organizational set-up offered a multitude of boundary settings in which collaboration and integration could thrive. How were these opportunities realized in research practice?

According to a start-up review carried out by MISTRA in 2010, collaboration across component projects and disciplinary boundaries was slow to take off, but was on the rise at the time of the review. Several component projects included researchers from more than one discipline and explored questions of an obvious interdisciplinary character. New ways to collaborate and address common research problems were noted. However, the efforts of parts of the consortium were described as manifesting ‘business as usual’, i.e. existing potentials for collaboration were not fully exploited (Bradshaw 2010). The final evaluation panel noted a promising interdisciplinary potential of some component projects, but argued that the programme as a whole should apply more of an ‘interdisciplinary approach’, meaning that all pillars of sustainability should be part of the basis of the programme (MISTRA 2012).

Realization of Future Forests' collaboration with actors outside academia faced challenges. In accordance with the programme plan, a panel of practitioners was set up to: (1) contribute personal/practical knowledge; (2) discuss research results generated by the researchers with other researchers and fellow panel members and (3) put scientific research into practice (Future Forests 2009). The panel was involved in a scenario-production process, in some individual component projects and gathered on a number of occasions when research results were presented. However, some stakeholders stated that they could have been more fully integrated in Future Forests, that the researchers dominated the agenda and that they were treated as passive recipients of research over which they had little influence (Bradshaw 2010). MISTRA's start-up review recommended that Future Forests find new ways to involve relevant stakeholder groups (Bradshaw 2010), and the panel was dissolved. MISTRA's final evaluation panel concluded that the programme had tried to create participatory structures but not met its objectives (MISTRA 2012).

ForSA had a key role in fostering collaboration, as it was coordinating a range of cross-cutting activities, such as scenario analyses, thematic working groups and integration projects (Future Forests 2009). In 2010 and 2011, all component projects and practitioners were engaged in a scenario-construction process (Moen et al. 2012). Driving forces and ‘critical uncertainties’, defining the contingent parameters of the scenarios, were jointly identified by the transdisciplinary research group but explored and described primarily within the individual component projects. The production of the actual scenarios and the synthesis were mainly carried out by ForSA. A range of non-academic actors were intensely involved at the start, but their involvement in the process declined over time. According to MISTRA's start-up review, the scenario-building process provided a useful forum in which project members could meet and for the integration of scientists from traditional social and natural science disciplines. However, several stakeholders found that they were presented with already-developed scenarios or questionnaires that they had not helped design and that in some cases did not represent their main concerns (Bradshaw 2010).

ForSA also had the capacity to initiate ‘integration projects’ for the specific purpose of fostering crossdisciplinary integration within the programme (Future Forests 2009). Such projects included researchers from several disciplines (i.e. humanities scholars as well as natural and social scientists) and some non-academic actors. A third kind of activity coordinated by ForSA was thematic working groups. Future Forests programme management and researchers invited researchers, experts and stakeholders from outside the programme to help analyse and synthesize particularly complex research questions. Thematic working groups included both natural and social scientists and sometimes stakeholders and practitioners. MISTRA's final evaluation concluded that ForSA had good intentions, the potential to do excellent work, but that the scientific quality of the outputs and the collaboration with external actors could be improved (MISTRA 2012).

Evidently, a range of boundary settings enhancing collaboration evolved within the Future Forests framework. Collaboration involving all component projects started with instrumental aims and with activities initiated at the programme management level, such as the scenario analyses. These activities were tightly structured, had a common methodological framework and involved practitioners. Inter- and possibly transdisciplinary intentions and activities can be observed in the early phases. As the scenario analysis project proceeded, it shifted towards a multidisciplinary mode in which collaboration took the form of division of labour. Much integration took place in the final synthesis, which did not involve the practitioners. Most integration projects and thematic working groups were later initiatives that emerged from a more flexible approach. Some were initiated by ForSA and other were products of the researchers' own interests. The forms of collaboration in these initiatives ranged from multi- and interdisciplinary approaches to various kinds of practitioner involvement (i.e. transdisciplinarity) depending on the research problem in question.

3.3 Conceptual integration

This part of the study is based on a two-step survey exploring the level of conceptual integration between the various research groups (component projects and ForSA) involved in the programme. How were concepts used and shared in order to cross established disciplinary boundaries? What kind of boundary work was taking place?

In the survey, leaders of the various research groups were asked to list the key concepts used in their Future Forests research. In a next step, the listed concepts were returned to the research leaders who were asked to indicate their understanding of them. Four alternatives ranging from ‘unfamiliar’ to ‘good scientific understanding’ were offered, and respondents were also asked to identify concepts with a potentially integrative capacity. The survey was carried out in October/November 2010; all 10 component projects and researchers in ForSA responded.

In total, 79 key concepts were listed by the research leaders. According to the respondents, 38 concepts were used primarily in research into natural systems and 41 in research into social or socioecological, systems. Nine concepts (see Table 2) were used in more than one component project.

Table 2 Concepts used by more than one component project.

Concept	No. of projects
Governance	4
Adaptive co-management	2
Globalization	2
Policy	2
Institutions	2
Sustainability	2
Productivity	2
Socioecological systems	2
Ecosystems/ecosystem services	2

As a next step, the research leaders' understanding of the listed key concepts was explored. The concepts listed in Table 3 scored the highest in terms of self-estimated understanding across the component projects, i.e. the leaders of at least nine projects rated their understanding as ‘limited scientific’ or ‘good scientific’. All these concepts also scored high in the assessment of potentially integrative concepts (indicated by more than seven project leaders). Many are broad and likely to be interpreted differently by the various researchers involved. These are potential boundary concepts.

Table 3 Concepts well understood across the component projects.

Concept	No. of projects
Adaptation/adaptive capacity (social systems)	10
Multiple forest use	10
Forest sector	11
Ecosystems/ecosystems services/ecosystem management	11
Sustainable development	11
Futures studies/trends	9
Climate change	10
Adaptive management (natural systems)	10
Resilience (natural systems)	9
Sustainability (natural systems)	10

Table 4 shows those concepts that scored the lowest in terms of understanding across the various component projects, i.e. the researchers typified them as ‘unfamiliar’ or their understanding of them as ‘common sense’ based, in six or more component groups. This indicates they were less well understood both by researchers with a natural science background and by those coming from the social sciences. Most Future Forests research teams identified themselves as primarily natural or social science groups depending on their disciplinary backgrounds. Accordingly, they are referred to as natural science and social science groups in the discussion below.

Table 4 Concepts less understood across the component projects.

Concepts	No. of projects
Institutions	6
Co-management/adaptive co-management/collaborative adaptive management	7
Collective action	6
Social construction/perceptions/frames/discourses/fundamental ideas	6
Ecological modernization	7
Resilience and concepts used in resilience theory (socioecological systems)	6
Extinction	6
Species turnover	6
Genetics	7
Evolution/co-evolution	6
Dispersal/pathway analysis	6
Colonization (natural systems)	7
Invasiveness (natural systems)	7
Resistance (natural systems)	6
Sedimentation	7
Weathering	6
Forest management intervention	6
Silvicultural systems/stand establishment/thinning/rotation length/nutrient management	7

Despite low scores in terms of understanding, resilience relating to coupled socio–ecological systems was highly valued as a potentially integrative concept (see concepts valued for their integrative potential in Table 3). Six research groups (i.e. three natural and three social science groups) indicated that their knowledge of the concept resilience was ‘limited’, but most still saw the concept as worth considering due to its integrative capacity, i.e. it was identified as a potential boundary concept. No social science concept was deemed unfamiliar/familiar in only a common sense way by most social science research groups, whereas the absolute majority of natural science concepts were. Accordingly, none of the natural science concepts appeared to be ‘unfamiliar’, or only known in a common sense way, to most natural science groups, whereas many of the ‘social-science concepts’ were. The only natural science concepts that were not scientifically well understood by the natural science groups were risk (relating to natural systems) and forest management interventions. The natural science–social science divide is thus reflected by a significant conceptual boundary.

Along with resilience, risk was suggested to be a highly integrative concept by both social and natural science groups, although they rated their understanding of it as moderate. In some cases, the natural and social science research groups saw different concepts as potential boundary concepts. In addition to the above concepts, governance and socio–ecological systems were suggested as potentially integrative concepts by a clear majority of the social science research groups. Both are concepts that the social science groups considered relatively familiar, while the natural scientists rated their knowledge more in terms of common sense. The natural science research groups, on the other hand, saw sustainable forest management and silvicultural systems as potential boundary concepts. Obviously, the understanding of these concepts varies between researchers with social and natural science backgrounds. However, they are broad concepts that could allow members of different disciplinary traditions to engage in conceptual development and trading.

An alternative way to explore this material is to go beyond concept definitions and look at the kind of phenomena being described. People from a wide range of cultural and educational backgrounds may share many experience-based, fundamental concepts, but as they enter more specialized domains, they may develop divergent language and understandings. Newell et al. (2005) recommend that crossdisciplinary research teams produce a checklist of relevant ‘high-level concepts’, conceptual clusters that capture essential aspects of the human environmental problem at stake, along with the associated, more specialized ‘low-level concepts’. Inspired by Newell et al. (2005), we use the low-level concepts listed by the Future Forests research teams as a point from which to explore the potential for integration and boundary crossing at a higher conceptual level. By grouping synonyms and thematically related low-level concepts, such as adaptation, ecosystem services, institutions, etc., we have constructed five conceptual clusters that assemble concepts relating to: (1) rules and management, (2) meaning and cognition, (3) practice/action and learning, (4) systems and (5) temporal and spatial developments/scales. Ideally, this exercise is carried out by the research teams. This was not done in this case, but our purpose of analysing the listed low-level concepts in terms of conceptual clusters is to demonstrate that despite the obvious disparity between the concepts and methods used, there is significant potential for boundary work. This potential is more obvious, and perhaps more easily realized, when moving to a higher conceptual level. However, a realization of this potential requires much more active boundary work than what was the case in the Future Forests programme.

The Future Forests programme is kept together by a defined research agenda. Problem definitions and research questions are therefore harmonized, or to some extent even integrated, to fit the common research frame. However, the level of conceptual integration across component projects was, at the time of investigation, relatively modest. Much research was carried out with little conceptual integration, or common methodology, and can be characterized as multidisciplinary. Only nine concepts were used in more than one component project. This likely reflects a lack of integration at the programme planning stage, i.e. most component projects were based on concepts and methodologies already familiar to research leaders. As the programme started, little active efforts were made to facilitate a process of mutual learning and systematic integration of concepts and methodologies. When asked to identify potentially integrative concepts, team leaders identified ‘loose’ concepts such as socioecological systems, ecosystems, resilience, risk, adaptation, multiple forest use and sustainable development/forest management. These concepts could potentially become productive boundary concepts and are plastic enough to apply to several conceptual clusters.

4. Discussion

In the following, we use our study of the Future Forests programme to address the questions posed in the introduction: (1) What are the stated motives and aims of the research programme? (2) What forms of collaboration are used? (3) What is the level of conceptual integration 2 years after programme initiation? We conclude by reflecting on the relationship between these factors and the research practice.

4.1 Motives and aims

The Future Forests programme was characterized by an instrumental approach with high expectations for applicability and short-term problem solving. At the same time, there were far-reaching expectations regarding integration between various research disciplines and between researchers and users. Future Forests was expected both to develop new ways of doing research and to generate short-term applications and decision-support tools. It aimed for transdisciplinarity, while being evaluated according to predefined deliverables on a 2-year basis. Realizing inter- and transdisciplinary aims requires more time (Höchtl et al. 2006; Harris and Lyon 2013; Podestá et al. 2013), not least for joint reflection and conceptual development.

Besides time, intellectual space is needed for reflection and multiple perspectives. Future Forests' research approach was characterized by instrumental interdisciplinarity, that is, research aiming to achieve defined objectives in order to address predefined tasks or challenges (Thompson Klein 1996). The research focused on solving ‘problems’ – often defined by users and decision-makers. The problems and their possible solutions were largely defined from the outset: how can we manage forests ‘to meet all our varied needs’, understood as requiring ‘up-scaling of the forestry industry to extract more paper and energy while at the same time safeguarding biodiversity, recreational needs and ecosystem services’ (Future Forests 2011).

Crossdisciplinarity was motivated by a need to tackle the problems in a multifaceted and comprehensive way. Mobjörk (2004) argues that problem-oriented instrumental research, expected to deliver useful results to authorities, industry and other actors, tends to generate research with a relatively ‘closed’ research design. She also suggests that aspirations of ‘wholeness’ may become problematic if they are based on a perception that all relevant factors can be included, that all kinds of knowledge can be synthesized and compared using the same metric and that there is one superior and generally accepted epistemology. Efforts to realize far-reaching integration in a setting with predefined objectives, narrow problem formulations and limited time risk constraining the latitude for reflexivity, which is a critical component of inter- and transdisciplinarity. There is always a risk that conflicting interests are hidden or re-formulated into a general call for more of everything. Consequently, research aims, outputs and the desired level of integration have to be carefully considered as a whole.

4.2 Forms of collaboration

Ideally, the forms of research collaboration should reflect the aims and motives of the research. In the case of Future Forests, both our study and MISTRA's start-up review demonstrate that crossdisciplinary collaboration was slow to take off, but progressed in the programme's second and third years. The first collaborative attempts, such as the scenario construction process, started with inter- and perhaps transdisciplinary intentions, but shifted into a more multidisciplinary mode. Nevertheless, the process offered the involved researchers an important boundary setting. As they got to know each other, more differentiated crossdisciplinary contexts, such as thematic working groups or integration projects, could be initiated by ForSA or proposed by the researchers themselves. The collaborative arrangements of these initiatives varied, but several included new constellations of researchers working across traditional boundaries. Most of them started after the launch of the already-structured programme, so opportunities existed to develop integrated project designs that spanned the entire cycle from problem formulation to synthesis. In terms of boundary settings, these smaller initiatives appear to offer interesting alternatives to broader cross-programme activities. Previous research also demonstrates that integration may be more easily achieved in smaller units, among individuals with shared interests, and that the importance of personal chemistry and trust should not be underestimated (Tress et al. 2003; Harris and Lyon 2013).

One of the challenges for public–private R&D collaborations in emerging scientific fields is to actively include the demand side (Roelofsen et al. 2011). Future Forests strived to develop collaboration between researchers and non-academic actors, but practitioners expressed displeasure with their ‘consultative’ role and the panel of practitioners was dissolved. Hence, a critical question concerns the nature of the intended stakeholder involvement. Were stakeholders expected to play a consultative or participatory role? As described in the programme plan, both roles were desired (Future Forests 2009). Both roles are possible, but for this possibility to be realized in practice, the terms of reference must be clear and jointly agreed to. This, in turn, requires a well-designed and facilitated process. In the case of Future Forests, neither clarity nor process was sufficiently attended to.

Elzinga (2008) highlights the political nature of ‘participation’: Who is empowered by transdisciplinary research and who is marginalized by it? In Future Forests, the forest industry, a particular subset of users, was co-funding the programme and represented on the Board of Directors. Expectations of some type of payback were logical considering the financial investment. Several other forestry sector actors with an economic stake in the research outcomes were also involved, for example, as practitioners. Given the interests at stake, critical questions address how various users and public interests ought to be balanced and who should be given voice and influence. How can users' expectations be acknowledged without compromising the open nature of a scientifically based futures study? Balanced representation is part of the answer, and Future Forests' aim was to set up a well-balanced panel of practitioners (Future Forests 2009) though, for reasons already mentioned, the panel did not function as intended. MISTRA's final evaluation panel accordingly noted that some stakeholders had been deeply involved in the research process (the forest sector), while others (such as NGOs) had been less involved (MISTRA 2012). These difficulties illustrate how transdisciplinary research in the absence of balanced representation risks amplifying the preferences of some users while marginalizing others.

Regardless of the criteria for involvement, the presence of vested interests in a research process will inevitably politicize the research and influence the science–policy interface. Knowledge production, however, is always situated and socially constructed, and the idea that anyone should be able to research an assumed reality in a detached and objective fashion must be rejected. Given these insights, an open, reflexive and continuous discussion of how actors and interests shape the research process should be integral to any transdisciplinary research programme. This discussion was underdeveloped in Future Forests.

4.3 Conceptual integration

Our third research question concerns conceptual, methodological and epistemological integration. Two years after programme initiation, the integration of the concepts and methods used in Future Forests' component projects was relatively modest. However, lack of conceptual integration is common in projects with inter- or transdisciplinary intentions (Brewer 1999b; Haugaard Jakobsen et al. 2004; Musacchio et al. 2005; Newell et al. 2005; Pohl 2005; Pohl et al. 2008). Project participants are often observed to have major problems making integration work, and component projects frequently end up as parallel studies (Tress et al. 2003). Previous research suggests that specific efforts and sufficient time are needed for integration to evolve (Tress et al. 2003; Messerli and Messerli 2008; Podestá et al. 2013).

Several studies emphasize the need to develop conceptual trading zones, boundary concepts, boundary settings, etc., that offer opportunities for those involved to renegotiate, or develop, the meaning of conceptual and theoretical terms (Messerli and Messerli 2008; Podestá et al. 2013). The present study indicates that such trading zones were under development in the Future Forests programme. Although the level of integration at the time of investigation was low, research team leaders identified a number of potentially integrative concepts, i.e. boundary concepts. More overlap and common direction are also discernable if we look at a higher level of generalization, for example, by constructing conceptual clusters. Conceptual clusters and boundary concepts are tools that may help researchers identify common research interests and conceptual trading zones. However, for this potential to develop, much more work is needed: boundary concepts must be actively explored, more efficient boundary settings have to be developed, skilled and experienced peers ought to facilitate a process of mutual learning and more time and intellectual space are required.

5. Conclusion

Interdisciplinary or transdisciplinary research is not always better, or more appropriate, than disciplinary and multidisciplinary research (Tress et al. 2003; Pohl 2008). For many academic purposes, and for many questions addressed in the Future Forests Programme, multidisciplinary studies may be as appropriate and more efficient. However, when an inter- or transdisciplinary research approach is indeed warranted, sufficient resources have to be mobilized to facilitate proper collaboration and integration. Previous research identifies a need for continuous integration, sufficient time, adequate training and intellectual space for critical epistemological reflection (Mobjörk 2004). The present study also pays attention to the relationship between programme objectives, forms of collaboration and the level of integration. We argue that efficient integration requires organizational and institutional settings that foster the development of a common conceptual framework, including boundary concepts. Favourable boundary settings are thus dependent on appropriate forms of collaboration. As the forms of collaboration often reflect the aims of the research programme, it is crucial that the aims, preferred forms of collaboration and means of integration be coherent. Far-reaching integration requires time, intellectual space, learning and boundary settings that are not always compatible with short-term instrumental objectives. Productive boundary work does not appear by itself but calls for proactive organization and continuous facilitation. We also argue that a continuous reflexive discussion of how actors and interests shape the research process is a necessary component of any transdisciplinary research programme. These insights may help researchers and participating actors to design research programmes that enable a realization of their transdisciplinary ambitions. They may also inspire future research on transdisciplinary national research programmes to explore the relationship between program organization and the practice of boundary work in greater depth.

Additional information

Funding

The research was funded through Future Forests, a multidisciplinary research programme supported by the Foundation for Strategic Environmental Research (MISTRA), the Swedish Forest Industry, the Swedish University of Agricultural Sciences (SLU), Umeå University, and the Forestry Research Institute of Sweden.

Notes on contributors

Karin Beland Lindahl

Karin Beland Lindahl (PhD) is a researcher with the unit of Political Science at Luleå Technical University in Sweden. Her research interest is in the politics of natural resource management and she is involved in a number of interdisciplinary research and education programmes. Her work includes analyses of local forest and mining conflicts as well as studies of global forest trends and governance.

Erik Westholm

Erik Westholm is a professor in Nordic and European Rural Development at the Dept for Urban and Rural Development, University of Agricultural Sciences in Sweden (SLU) and in Geography at Högskolan Dalarna, Sweden. At the SLU, he is program director for the strategic platform Future Agriculture. His main research interests are related to rural change and the future use of natural resources in Nordic environments and he is especially concerned with the role of the “future” as a concept and scientific category.

Notes

^1. Email: [email protected]