Constructing shared understanding of complex interdisciplinary problems: Epistemic games in interdisciplinary teamwork

ABSTRACT

Background

Constructing shared understanding of complex interdisciplinary problems is one of the most challenging aspects of interdisciplinary teamwork. How this process unfolds is under-researched, making it esoteric and difficult to scaffold. This paper aims to provide an articulated and nuanced account of what is involved in student teams’ construction of shared understanding of complex interdisciplinary problems.

Methods

The study combines the theoretical lens of epistemic games with an ecological analytical perspective. Drawing on the analysis of ethnographic cases, it explores how four graduate student teams construct shared understanding of complex problems during interdisciplinary projects.

Findings

Construction of shared understanding is a multifaceted and dynamic process that extends over the entire problem-solving activity. It relies on epistemic moves that explicate and juxtapose different perspectives and connect abstract ideas with knowledge grounded in students’ experiences and contexts. The pursuits students engage in, when creating shared understanding, reveal epistemic differences related to formulation of the interdisciplinary problem and what constitutes trustworthy knowledge.

Contribution

The paper contributes to the literature on the nature of students’ construction of shared understanding of complex interdisciplinary problems revealing critical recurrent moves. It also extends earlier conceptualizations of epistemic games as primarily discourse games and demonstrates that they are profoundly multimodal and distributed.

Introduction

Humanity is increasingly faced with complex problems that cannot be fully understood or addressed without considering different perspectives and drawing on diverse areas of knowledge and expertise. The causes and consequences of climate change, pandemics, and social exclusion offer multiple examples of such problems, which often require professionals and other knowledgeable people to work across and beyond disciplines when addressing them.

The changing nature of problems and professional work is reflected in higher education, where students engage in interdisciplinary project-based learning and teamwork as a part of their curriculum (Lyall et al., 2015; Warr & West, 2020). Despite an increased focus on interdisciplinary teamwork, learning to work across disciplines has proven challenging (Ashby & Exter, 2019). Teamwork requires constructing not only a shared solution, but also a shared understanding of the problem to enable team members’ coordinated actions in developing an effective solution. In interdisciplinary settings, this work needs to be done with people who do not share the same foundational knowledge and do not use the same concepts, problem-solving methods or disciplinary standards. Further, complex interdisciplinary problems are often called “ill-structured” or “wicked,” suggesting that they do not have a single solution and can be framed in different ways, making construction of shared understanding particularly challenging, but essential (Crowley & Head, 2017; Jonassen, 2011; Markauskaite & Goodyear, 2017; Rittel & Webber, 1973).

Research on interdisciplinary learning and teamwork highlights the significance of dialogue (Choi & Richards, 2017; Nikitina, 2005; Woods, 2007), cognitive alignment (O’Donnell & Derry, 2005; Pennington, 2016), joint focus (Goodwin, 2018; Nicolini et al., 2012) and other linguistic, cognitive, and epistemic aspects (Gorman, 2010; MacLeod, 2018; Thagard, 2005) in constructing shared understanding. Concrete methods and tools intended to help teams work across disciplines have also been described in the literature (Bammer, 2013; Hubbs et al., 2021). However, there is a significant gap between the former research findings and the latter practitioner-oriented guidance. Our current research helps build bridges by closely examining how interdisciplinary teams of graduate university students, whose project-based coursework involves them in solving ill-structured problems, which we refer to as complex problems, jointly construct shared understandings of these problems in order to make progress on their tasks. What is needed is an integrative, nuanced, and articulated account of how such interdisciplinary teams construct shared understanding, so that important parts of the process, and the architecture of their relations, can be mapped, named, and discussed explicitly in guidance for teachers and others. Our main research question is: How do student teams construct a shared understanding of complex interdisciplinary problems?

We address this question by combining the theoretical framework of epistemic games (Collins & Ferguson, 1993; Markauskaite & Goodyear, 2017) with an ecological analytical perspective (Goodwin, 2018; Hutchins, 2010; Markauskaite et al., 2023). Epistemic games can be described as “generative patterns of inquiry that underpin knowledgeable actions” (Markauskaite & Goodyear, 2017, p. 399). This concept has origins in Wittgenstein’s (1958) notion of “language games” and suggests that construction of shared situated meaning has recognizable game-like qualities, such as rules that all parties embrace and an overall “family resemblance” of the game. Each game is enacted in unique ways by particular people in a particular context, enabling games, while having shared qualities, to be generative and played in an infinite number of situation-specific ways. This framework allows us to focus on identifying shared qualities and patterns (rules, moves, constraints, etc.) as well as target knowledge forms that guide joint knowledge work, thus helping us create an articulated account of what is involved in constructing shared understanding.

It is essential to note here that “knowledge” and “understanding” are intertwined but distinct concepts. “Knowledge” refers to the facts, skills, and comprehension that participants bring to the game, forming the building blocks of their actions. “Understanding” depends on but is not reduced to knowledge. Rather, it involves thinking and the flexibility to act with what one knows (Perkins, 1998). Shared understanding emerges as the collective interpretation and shared meaning crafted through interaction within the game’s framework.

Previous studies of epistemic games primarily conceptualize them as discourse games, focusing mainly on language (Morrison & Collins, 1995; Odden & Russ, 2018). However, studies of joint teamwork in complex settings show that coordinated meaning-making activities draw heavily on the environment, embodied interaction, and identities (Goodwin, 2018; Hutchins, 1995; Nersessian, 2005). Therefore, to answer our research question, we adopt an ecological analytical perspective (Goodwin, 2018; Hutchins, 2010; Markauskaite et al., 2023) and analyze distributed multimodal interaction, including how it is shaped by disciplinary and other experiences. We investigate what kinds of epistemic moves student teams make and how they construct shared understanding of complex problems.

This paper is divided into five sections. First, we provide a conceptual elaboration of interdisciplinarity and shared understanding, and present the ecological perspective and theoretical lens of epistemic games. Next, we provide an overview of the ethnographic cases and ecological analytical approach used. Following this, we share our findings by highlighting and describing key features of the epistemic game students played while constructing shared understanding. We reflect on the implications of our findings for interdisciplinary teaching and learning and conclude by suggesting areas for future research.

Literature review and theoretical framework

Interdisciplinary knowledge work

There is no agreed view of what interdisciplinary knowledge work is, beyond the fact that such work must include knowledge and methods from more than a single disciplinary field (Barry & Born, 2013). Literature often uses the term “interdisciplinarity” in either a specialized sense to describe a particular kind of knowledge work or in a broad sense as a general term (Huutoniemi et al., 2010).

In a specialized sense, interdisciplinarity often refers to knowledge work that connects several disciplinary fields to create a coordinated whole that is more than the simple sum of its parts (Huutoniemi et al., 2010). Such knowledge work frequently involves the integration of disciplinary ideas and tools, such as concepts, theories, and methods (Boix Mansilla, 2017). In this case, interdisciplinarity is different from crossdisciplinarity, which involves the exploration of the same topics across disciplines without making connections, or multidisciplinarity, which involves solutions to problems using knowledge from multiple disciplines without integrating them (Klein, 2010). It is also different from transdisciplinarity, which involves the emergence of new perspectives that transcend the boundaries of academic knowledge fields, for example, by including the perspectives of the clients or the local community (Nowotny et al., 2001). To illustrate these concepts, consider the COVID-19 pandemic. In a crossdisciplinary inquiry, the pandemic’s impact may be examined from various disciplinary perspectives (epidemiological, psychological, economic, etc.) without making connections between them. In a multidisciplinary inquiry, experts from different disciplines, such as epidemiologists and economists, may work on specific aspects of a pandemic response plan, but the integration of their expertise may be limited. In an interdisciplinary investigation, experts from various fields, such as geneticists, data modelers, and epidemiologists, may collaborate to model the virus’s behavior and develop an integrated vaccination strategy. In a transdisciplinary inquiry, a more holistic approach may be taken by involving epidemiologists, psychologists, media specialists and local community members, working together to create a culturally appropriate vaccination promotion campaign that transcends disciplinary boundaries. That said, a number of scholars question the usefulness of making hard distinctions between inter-, cross-, multi-, or trans- disciplinary inquiry, as in practice the relationships between disciplines are more diverse and complex and continually evolve, restricting the utility of these terms when explaining complex problem-solving practices (Barry & Born, 2013; Graff, 2016).

In a broad sense, interdisciplinarity is used as a generic “umbrella” term to refer to knowledge work that includes different kinds of engagement with knowledge beyond a single discipline, including multidisciplinarity, crossdisciplinarity, and transdisciplinarity (Barry & Born, 2013; Huutoniemi et al., 2010). For example, Huutoniemi et al. (2010) describe the term interdisciplinary as a “characterization of all collaboration across epistemological boundaries.” (p. 83).

In our study, we use the term “interdisciplinarity” in a broad sense to encompass all forms of knowledge work that involve engagement and collaboration across disciplines and with non-academic fields (such as the private, community and industry sectors). Our objective is to explore the mechanisms underlying the construction of shared understanding of complex real-world interdisciplinary problems, placing particular emphasis on the utilization of diverse ecological resources (knowledge, skills, methods, tools, etc.): hence a broad view of interdisciplinarity is most appropriate.

Constructing shared understanding

There is a lack of consensus on what shared understanding is. A wide range of terminology is used to describe closely related concepts. Examples of these terms include “joint understanding” (Puntambekar, 2006), “common ground” (Clark & Brennan, 1991), “team mental models” (Klimoski & Mohammed, 1994), “shared mental models and transactive memory” (Pennington, 2016), “sensemaking” (Renger et al., 2008), “schema and mental model alignment” (DuRussel, 2005), “cognitive convergence” (Teasley et al., 2008) and “convergent diversity” (Goodwin, 2018). Dominant notions differ in their focus, ranging from those that describe shared understanding as overlapping mental models or conceptual core (e.g., Mulder et al., 2002; Renger et al., 2008; Thagard, 2005), to those that emphasize the exchange, coordination and integration of mental models and perspectives when people work together on specific problems (e.g., DuRussel, 2005; Pennington, 2016; Puntambekar, 2006; Stein et al., 2007), to those that focus on the distributed and interactional nature of jointly constructed understanding in specific situated settings (e.g., Goodwin, 2018; Hall et al., 2002; Nersessian, 2012).

Shared understanding is often described in terms of overlapping mental schemas or conceptual core (Thagard, 2005). For example, Thagard (2005) uses the metaphor of “trading zones” to describe a “shared cultural arena” between several dissimilar groups (e.g., disciplines) in which these groups can find a “common ground.” According to this view, “Trading zones do not require complete agreement or a universal vocabulary, but they do require an overlapping conceptual core among the cultures or disciplines that participate in them” (p. 328, italic added). Such “common ground” and “conceptual core” are usually developed as a part of joint socialization (DuRussel, 2005; Pennington, 2016). In cognitive literature, this conceptual core is often described as “mental schemas” (DuRussel, 2005, p. 188), “mental models” (Pennington, 2016, p. 302) and other reusable mental representations of prior knowledge that form the basis for understanding newly encountered circumstances. In disciplinary contexts, this core usually involves knowledge of key concepts, theories and methods. Such shared understanding is relatively enduring and can be developed by the members of the team prior to, or outside of, joint teamwork. For example, team members may have shared knowledge of “relativity theory” or “open data” from learning about these topics at different times and places outside of teamwork. This shared understanding also includes kinds of knowledge that disciplinary teams take for granted, such as epistemic standards, values, interpretative schemas, and other epistemological features of different domains. They play a critical role in interdisciplinary collaboration (Dougherty, 1992; MacLeod, 2018).

Other notions of shared understanding emphasize the exchange, coordination and integration of mental models and perspectives (DuRussel, 2005; Pennington, 2016). They suggest that while the construction of a team’s shared understanding may require the existence of the overlapping conceptual core or other prior knowledge, this is not sufficient, and a team’s shared understanding is necessarily constructed through situated interaction. For example, Pennington (2016) makes a distinction between an individual’s mental model, which is an “individual, internal, cognitive simplification of reality that is formulated through experience and is relatively stable” (p. 302) and a group’s shared mental model: “emergent, distributed cognitive states that are generated in real-time in a specific context, reflect shared understanding of a problem (e.g., shared problem model), and evolve as group understanding evolves” (p. 302). According to this view, shared understanding extends beyond the conceptual core and includes aspects specific to the context, team, and problem, such as knowledge of the task to be done (Agredo-Delgado et al., 2022; O’Donnell & Derry, 2005; Pennington, 2016), what the team, the individuals within it, and others know and do not know (Cannon-Bowers et al., 1993; Pennington, 2016; Salas, et al., 2017), constructed representations and objects (O’Donnell & Derry, 2005; Pennington, 2016; Pennington et al., 2016, 2021), technologies and procedures (Mulder et al., 2002). In interdisciplinary teams, the challenge of knowledge integration (Pennington, 2016) reflects the complex process of merging different disciplinary insights, emphasizing the need for shared understanding of shared goals and methods. From this perspective, shared understanding is primarily a matter of adequate alignment between individual work-related schemas (DuRussel, 2005) or coordination between individual learning and group processes (Pennington, 2016). This view makes a clear distinction between internal mental processes and interactions in the external environment. A team’s shared understanding is a mental construct, even if it is constructed through dialogue and extensive use of external representations.

Some other notions of shared understanding focus on the distributed and interactional nature of jointly constructed understanding (Goodwin, 2018; Hall et al., 2002; Jornet & Steier, 2015; Nersessian, 2012). They move away from investigating mental structures and instead look at the structures that underpin distributed interactions. From this perspective, shared understanding is not so much a mental structure or a mental process but a coordinated action around a common focus in a shared environment. This perspective focuses on how people use language and incorporate external objects and representations within their joint activity. For example, Goodwin (2018), drawing on Wittgenstein (1958), argues that “the meaning of a representation is not its bearer … but rather the grammatical processes used to articulate the representation within a relevant language game” (p. 283, italic added). This view rejects equivalence between internal (mental) and external representations and processes, suggesting that the coordinated use and construction of external representations and interactions around them are themselves a part of shared understanding. For example, mock-ups and architectural drawings are often incorporated in the construction of joint understanding in architectural work without expecting that all parties have a similar “internal mental model” or understanding of their meaning (Kasali & Nersessian, 2015). From this perspective, shared understanding can be understood by looking at common interactional structures—“language games” (Wittgenstein, 1958)—in which this understanding is constructed and articulated.

Empirical studies that draw on the above perspectives reveal various mechanisms that play a role in the construction of shared understanding. Specifically, studies show the importance of shared experiences and preexisting knowledge among team members (DuRussel, 2005; Pennington et al., 2021). Further they reveal that interaction, negotiation, and alignment of diverse perspectives play a role in the construction process (DuRussel, 2005; Pennington et al., 2016, 2021). Other studies show that shared understanding emerges from the dynamic interactions and coordinated actions focused on shared goals and environments, highlighting the distributed and interactional nature of this construction (Goodwin, 2018; Hall et al., 2002; Nersessian, 2012). The use of external representations and artifacts plays a crucial role in this process, as these tools facilitate the collective construction of understanding (Goodwin, 2018; Kasali & Nersessian, 2015; Pennington et al., 2016, 2021). This construction of shared understanding is a real-time process that involves the evolving understanding of a team, underscoring the fluid and emergent nature of shared understanding in collaborative problem solving.

In our work, we have extended the distributed and interactional view of shared understanding by adopting an ecological analytical perspective, to identify common interaction patterns associated with students’ situated pursuit of shared understanding of complex interdisciplinary problems.

An ecological perspective

The ecological analytical perspective, as we employ it, foregrounds the inherent interconnectedness and relations among various elements within the cognitive ecosystem (Goodwin, 2018; Hutchins, 2010, 2014; Markauskaite et al., 2023). According to Hutchins et al. (2009) “the meaning of the event is established by the mutually constitutive relations among task, gesture, and local space” (p. 452). Thus, we focus on the distribution of interactions as a central element in an empirical investigation of how shared understanding is constructed, taking a relations-focused perspective that zooms in on the dynamic interplay of gestures, language, and other modalities in interaction with the environment in teams’ collaborative efforts to construct a shared understanding of the problem (Hutchins & Palen, 1997; Markauskaite et al., 2023).

Following Goodwin (2018) and Hutchins (1995), our approach is grounded in the observation of tangible, real-world interactions. Similar to others (Nersessian, 2019), we do not make ontological claims about the nature or functioning of the cognitive system, employing the ecological perspective as a lens to analyze the interconnected network of influences that form part of the broader cognitive ecosystem within which these multimodal interactions are situated.

Epistemic games

The term “epistemic games” refers to the different, culturally patterned ways of constructing knowledge¹ (Collins & Ferguson, 1993; Markauskaite & Goodyear, 2017; Morrison & Collins, 1995). Creating an ordered list, making a comparison, and analyzing a trend are examples of basic epistemic games that people often play within, across and beyond different disciplines. “Epistemic” connotes knowledge and knowing. Epistemic games consist of “sets of moves, constraints, and strategies that guide the construction of knowledge around a particular epistemic form” (Morrison & Collins, 1995, p. 40). Collins and Ferguson (1993) conceptualize epistemic forms as “the target structures that guide inquiry” (p. 25). An ordered list, a compare-and-contrast table or a trend graph are examples of epistemic forms that need to be “completed” when one plays an epistemic game in which they feature.

Markauskaite and Goodyear (2017), extending this notion to complex knowledge work, consider epistemic games as generative patterns of inquiry, used for a range of epistemic purposes, within and across many domains of human activity (including disciplines and professions), to produce actionable knowledge. For example, diagnosing a patient and creating a treatment plan, in medicine; assessing students’ learning and creating a lesson plan, in teaching; doing market research and producing a business plan, in commerce, are common epistemic games that doctors, teachers and entrepreneurs play when creating situated, actionable knowledge as a part of their professional activities. Some of these patterns are well articulated and enacted deliberatively, but some are spontaneous and intuitive. They are not hard-wired to a specific context; rather, they are flexible enough to be deployed across different situations and contexts and often enacted collaboratively with other professionals or clients. Each domain of human activity has a range of such epistemic games. Further, different epistemic games are interrelated: sometimes played one after the other, sometimes woven together and sometimes played as a part of a larger game. For example, conducting an experiment is an epistemic game that could include various smaller games, such as identifying samples, recording measurements, creating graphs, and writing a report.

While epistemic games are generally considered in relation to strategies used by experts (Collins & Ferguson, 1993; Markauskaite & Goodyear, 2017), others have adopted this concept to examine knowledge practices and actions of non-experts. For example, Tuminaro and Redish (2007) examined students’ (non-experts’) problem-solving strategies, while Odden and Russ (2018) examined students’ sensemaking processes in introductory physics education, showing that these processes also have features of epistemic games. Epistemic games, therefore, provide a flexible analytical lens to examine how interdisciplinary student teams construct shared understanding. This lens allows us to identify and articulate the patterns (e.g., rules, moves, constraints, epistemic forms) involved in constructing shared understanding.

Methods

Research context and teams

To understand how interdisciplinary student teams construct shared understanding of complex problems, and the epistemic games they play while doing so, we draw on data collected as part of a larger ethnographic case study conducted across two graduate interdisciplinary courses (see for details Arthars, 2021). We studied four teams, each composed of 3–5 students (18 in total), engaged in interdisciplinary projects focused on complex interdisciplinary problems that lasted 5–13 weeks. Ethics approval was received from the University of Sydney Human Research Ethics Committee (approval number 2018/582). All students provided their informed consent.

The first course was an elective course that provided students with several broad problem areas (e.g., trust in digital environments and air pollution) to choose from. Teams were required to develop a commercially viable solution and pitch it to a panel of business experts who played the role of potential investors. In this course, team members were undertaking graduate studies in Business, Design, Science, and Information Technology (IT). The second course was a required course for Business students and an elective for Design students. A corporate client posed the problem “How can the placement of reinsurance² be improved,” and teams were required to take the perspective of external consultants to develop potential solutions. The problem contexts presented to the teams varied, with two teams facing a novel domain (i.e., reinsurance), while the other two teams dealt with broader themes that were outside team members’ core expertise and needed to be narrowed down. This allowed us to explore how shared understanding is constructed from different starting points and for different complex interdisciplinary problems.

In both courses, students came from diverse professional backgrounds and had different professional experiences. For example, the students in the most homogeneous team, composed of students studying Business, had backgrounds in advertising, marketing, business development, insurance, finance and exercise physiology. They worked in financial services, transportation and logistics, performing arts, sports and governmental organizations. Therefore, the interdisciplinary nature of the teams extended beyond their current academic disciplines, to include diverse professional experiences, and backgrounds. Their collaborative efforts leveraged this wide array of backgrounds and experiences, when engaging with complex real-world problems.

Data collection

Our data collection strategy was informed by our core interest: teams engaging with complex, interdisciplinary problems. Data collection was structured to capture the dynamism of this phenomenon in real-time, focusing on the interactive, collaborative processes of complex problem solving and how students drew upon diverse knowledge and other resources in this process. Our methods included video-recorded observations of teams throughout the problem-solving activity, culminating in over 90 hours of video footage, supplemented with field notes, a reflective journal, interviews, as well as various digital artifacts such as photographs of whiteboards, flip charts, post-it notes, and online team-created resources. In addition, students were interviewed about their disciplinary backgrounds and professional experiences as these elements were important for understanding the unique contributions and perspectives they brought to the team.

Data analysis

Our analysis draws primarily upon video-recorded group meetings, with other data used for triangulation. Video data provided a detailed record of the teams’ multimodal interactions, including language, gestures, and artifacts, crucial to examining the process of constructing shared understanding.

First, all 90+ hours of video-recorded observations were reviewed and annotated by the first author to exclude episodes that did not feature interaction between team members or where teams were not working on the task. The remaining episodes included a range of interactions, including interpreting tasks, agreeing on meeting schedules, and interpreting feedback from mentors. Second, using intensity sampling (Creswell & Poth, 2016), all episodes that illuminate the phenomena of interest—the construction of shared understanding of the problem – were selected for further analysis. These episodes involved situations directly related to problem exploration, such as discussing a brief for the project, narrowing down the focus, and creating shared problem statements (12.1 hours in total). This corpus was analyzed in detail.

The ecological perspective was operationalized by focusing on the embodied and environmentally situated organization of activities (Müller et al., 2013; Streeck et al., 2011) at a relational level (Markauskaite et al., 2023). This includes analysis of the shared focus and the dynamic interplay among team members’ language, bodies, actions, gestures, gaze, and use of artifacts during their interactions (Goodwin, 2018; Markauskaite et al., 2023; Salas et al., 2017). In order to capture the complexity of these multimodal interactions, we combined Multimodal Interaction Analysis (Norris, 2004) with Jeffersonian notations (Jefferson, 2004) (see Appendix). This combination enabled us to capture paralinguistic features-such as pauses, overlaps, and intonation, verbal and nonverbal modalities, artifacts, and other physical elements in the environment. This approach aligns with our aim of capturing the rich interactional dynamics, central to analyzing the construction of shared understanding in team problem solving from an ecological perspective.

We initially approached the construction of shared understanding as a collection of multi-player epistemic games (Bielaczyc et al., 2013) recognizing that it is through collective contributions and interactions between team members and their environment that the games are played. We followed a multi-step process adapted from Odden and Russ (2018) and coded the episodes against the following characteristics: 1) entry conditions, 2) epistemic goal, 3) target epistemic forms, 4) constraints, 5) moves, 6) transfers, and 7) exit conditions (Table 1).

Table 1. Features of epistemic games.

Feature	Description
Entry conditions	The conditions that must be met for a game to begin.
Epistemic goal	The purpose or intended outcome of the game.
Epistemic form	The knowledge form or epistemic focus that guides construction of shared understanding.
Constraints	“Rules” of the game that should be followed.
Moves	Main actions that participants take to reach the target epistemic form.
Transfers	Ways in which participants temporarily move from one unfinished epistemic game to another.
Exit conditions	The conditions in which the epistemic game ends.

We analyzed the key features of team interactions and identified recurring patterns. Turns of talk and actions (e.g., writing, expressing confusion through bodily cues) were used as the main unit of coding. A number of game-like patterns were initially identified across the episodes by the first and second author.

The patterns, often recurring in the teams, were iteratively refined until a coherent set of key epistemic moves, constraints and other characteristics of epistemic games was agreed upon by the first and second author. This set effectively captured how shared understanding of the problems was constructed across all teams. Moves were closely interwoven; and defined not by particular turns and actions but a distinct epistemic purpose (e.g., knowledge sharing), with the problem statement representing the broader epistemic focus across these moves. Further all moves had some shared characteristic features (e.g., “silent players”). This led us to reconceptualize moves as recognizable composites of turns of talk and actions with a distinct epistemic purpose within a broader shared understanding game. In answering the research question, we also focused on depicting characteristics of the game (e.g., entry conditions, moves) and on features that characterized game-play (e.g., critical role of “silent team members”). While the analytical process was carried out by authors one and two, author three acted as a sounding board throughout to enhance trustworthiness.

The results presented are drawn from data analyzed across all four teams; however, in this paper the results are illustrated using episodes from one team exclusively. Our rationale for this choice was twofold. Firstly, the selected episodes well exemplify gameplay and construction of shared understanding across all teams. Secondly, by focusing on a single team, we were able to delve deeper into illustrating the nuances of the gameplay, providing a richer and more detailed examination within the available space. Our presented episodes came from the Trust team, featuring three international students: Robert, a graduate Design student with professional experience in user experience (UX) design; Paul a PhD Business student whose professional background spans product and process management, business analysis, and project management in finance and other sectors, along with prior studies in science, engineering, finance and marketing; and Andrew, a graduate IT student, with a professional background in IT, logistics, artificial intelligence and startup businesses.

Results

Our results showed that constructing shared understanding of the problem can be characterized as an epistemic game that is played repeatedly by teams throughout the project. This epistemic game, played in pursuit of shared understanding of the problem, focused on understanding the issue (i.e., what is the issue here), framing the problem (i.e., what kind of problem are we trying to solve) and articulating it (i.e., how can we express our framing of the problem in a sufficiently concise statement). However, it also included reaching agreement on the problem’s significance (i.e., is this the problem that we, as a team, want to solve) and the level of flexibility attached to their decision to frame the problem in a particular way (i.e., how open we are to changing the problem later). This process was dominated by five key recurring moves that had clearly identifiable patterns: 1) knowledge sharing, 2) perspective-taking, 3) simulation, 4) knowledge validation, and 5) negotiation. These moves focused on explicating and juxtaposing different perspectives, including students’ epistemic understanding, and connecting abstract ideas that students were discussing with knowledge grounded in students’ experiences and concrete contexts.

Overall, students’ construction of shared understanding contained important common qualities. First, their attempts to create a shared understanding of the problem revealed epistemic differences, which often mirrored aspects of their prior epistemic experiences, such as their disciplinary backgrounds. However, these differences were rarely acknowledged or discussed. Second, grounding the problem and specialized knowledge in concrete (real and imagined) contexts and experiences was central to students’ joint work. This approach mitigated a lack of formal disciplinary or professional knowledge but sometimes presented unexpected epistemic challenges. Third, students’ interactions constructing shared understanding were profoundly multimodal, entwining language with shared representations and students’ bodies. Below, we present the shared understanding epistemic game (summarized in Table 2), including the five main moves and we illustrate key qualities of students’ gameplay.

Table 2. Features of the shared understanding game.

Feature	Description
Entry conditions	The team lacks a cohesive understanding despite individual knowledge. Agreement on the need to understand and formulate the problem.
Epistemic goal	A singular, integrated vision of the complex problem the team are attempting to solve.
Epistemic form	A statement of the problem
Constraints	Requires at least two active participants. At least one participant must have knowledge of the problem domain. Teams must be open to revising their understanding of the problem.
Moves^*	Knowledge sharing Perspective taking Simulation Knowledge validation Negotiation
Transfers	Individual work Problem-solving External consultation
Exit conditions	Consensual problem statement Impasse Running out of time

*N.B. This list is not exhaustive, it is representative of moves identified across all interdisciplinary teams.

Entry conditions, epistemic goal and epistemic form

The realization that a problem could be framed in multiple ways and interpreted differently by individual team members, coupled with agreement regarding the need to understand, agree on and formulate the problem to be solved, act as entry conditions into the shared understanding game. The epistemic goal of the game is to construct a singular, integrated vision of the complex problem the team are attempting to solve. This vision is articulated as a problem statement, which serves as the epistemic form.

Across all the teams in our study we found the shared understanding game was preceded by some form of discussion related to actions to be taken (i.e., what the group should do next). This indicates that the shared understanding game is played in order to enable further problem-solving activity.

The following episode illustrates this realization and its articulation. It comes from one of the earliest meetings of the Trust team, following a discussion about what to do next. Robert has suggested the team agree on “personas” (the term comes from a design-thinking technique and is used to represent the type of person, the user, they are solving the problem for). In response to Robert’s suggestion, Andrew signals the need to construct a shared understanding of the problem prior to understanding the intended user (Figure 1).

Figure 1. The Trust Team entering the shared understanding game.

Andrew indicates his lack of understanding of “exactly what the problem is,” which is acknowledged verbally by Paul (“mmm”) and non-verbally by Robert (nodding his head). Andrew’s use of the metaphor “on the same page” underscores the need for shared understanding. When Andrew turns to look at Robert while gesturing, he is creating a point of shared focus, on the verbal message and on the embodied action, anchoring the abstract concept in the physical space. Andrew’s gesture—a straight line motion with his hand—visually echoes the verbal metaphor, demonstrating how team members use multimodal means to enhance communication.

Excerpts containing similar entry conditions were identified across team meetings, indicating that the game was played by all teams on multiple occasions.

Constraints (rules) of the game

All team members must be open to discussing and aligning on the problem’s definition, suggesting an underlying openness to revising their understanding as part of the shared understanding game. This openness is illustrated in the follow-up turns in the Trust team (Figure 2).

Figure 2. Discussing openness to constructing shared understanding.

This episode shows that all three members agreed to engage in reaching a shared understanding of the problem. It also illustrates differences between team members that often mirrored their disciplinary backgrounds and professional experiences. Andrew, representing an IT perspective, seeks to construct a shared understanding of “what exactly the problem is” from the outset, signaling a discipline that values clarity of the problem before proceeding to creating a solution. On the other hand, Robert, through his design lens, is comfortable with a much less precise and tentative understanding of the problem, suggesting a preliminary agreement on the problem that can be further iteratively refined through user interviews, reflective of a discipline that integrates user feedback into problem formulation.

This difference was also evident in later meetings where Robert attempted to review the formulated problem and enter the game again, but due to Andrew’s unwillingness to revise the problem statement, the team did not proceed (Figure 3).

Figure 3. Failed entry to the shared understanding game.

These episodes also illustrate one of the inherent challenges faced by all interdisciplinary teams: the varying willingness of students to play the shared understanding game throughout problem-solving. While some students were less willing to play the game in earlier meetings because they did not have enough knowledge about the problem (e.g., Robert), some were less willing in later meetings because they viewed it as regressing “back to the drawing board” (e.g., Andrew).

The shared understanding game is a multiplayer game, requiring at least one team member to have some knowledge of the problem domain. But active participation is also required from other team members. Verbal acknowledgments, such as “mmm,” “yeah” and gestural affirmations, such as head nods, are integral to the game’s dynamics as they signal attention, active listening and participation within the game.

Further, the game involves a dynamic distribution of interactions with team members shifting between speaking, writing, drawing, listening, gesturing, and observing. This shift is illustrated in the extract below where Paul has begun providing a verbal explanation of the problems he had identified and Robert interrupts him (Figure 4).

Figure 4. Shifting from verbal explanations to visual representations.

This extract illustrates two key features of this epistemic game. First, construction of shared understanding often involves shared external representations (e.g., written inscriptions or drawings). Second, this game sometimes involves explicit epistemic actions, (i.e., actions related to how an epistemic game should be played and why). In this case Robert makes this explicit epistemic action by asking Paul to draw and explaining why.

Key moves within the game

Five key recurring moves are presented below—knowledge sharing, perspective-taking, simulation, knowledge validation and negotiation.

Knowledge sharing

Knowledge sharing aims to make explicit the knowledge held by individual team members, including conceptual, procedural, contextual, and experiential knowledge.

For example, in the Trust team, understanding of the problem involved discussion of open banking, a system that enables customers to securely share financial data with approved third-party providers (e.g., banks, insurance providers). To understand how open banking works, the team needs shared knowledge of key related concepts, such as “read access” and “write access.” Paul introduces these concepts while visually illustrating their relationship on the whiteboard wall (Figure 5).

Figure 5. Paul shares knowledge of open banking.

Using verbal explanation and external representation, Paul is attempting to build a foundation of shared knowledge among team members for understanding of the complex problem. Robert positions his body toward the whiteboard wall, with his hand on his chin and his gaze directed at what Paul is writing, indicating that he is participating. Andrew responds to each of Paul’s turns positively (Mmm↑, Ye:ah, Mmhmm), encouraging Paul to continue, before seeking clarification by asking “who authorizes to take” while gesturing toward what Paul has written on the whiteboard wall (read access—“authorize to take info”). Andrew’s pointing gesture directs attention while invoking the shared history of the conversation that the whiteboard represents.

This interactive exchange illustrates that knowledge sharing encompasses verbal communication, non-verbal elements such as gestures and body language, as well as external representations such as whiteboard inscriptions. Inscriptions create a shared visible space that facilitates joint attention, supports verbal explanations and demonstrates relationships between concepts, while simultaneously serving as enduring reference points that enable ongoing discussion and continual development.

Perspective-taking

The aim of perspective-taking is to construct an understanding of the possible experiences, including feelings, thoughts, or actions of a person or entity (e.g., a customer or Facebook). Constructing shared understanding of complex interdisciplinary problems requires constructing shared understanding of various stakeholder perspectives across different contexts. The dynamic nature of these problems is such that perspectives change and therefore perspective-taking may involve projecting a perspective onto an imagined past, present or future scenario.

An example of this is illustrated in the episode below where Paul directs the team to shift their perspective from the present to a situation five years in the future. The discussion focuses on future changes to banking and financial services (Figure 6).

Figure 6. Paul illustrates future open banking scenario.

Here, playing the shared understanding game involves all members of the team perspective-taking as a customer of a bank, five years in the future.

Teams often shifted between perspectives and they had to reconcile differences in the perspectives taken when they resulted in conflicting understandings of the problem. In the following episode Robert challenges Paul to reconsider the problem from a different perspective (Figure 7).

Figure 7. Robert challenges Paul to consider another perspective.

As illustrated in the discussion between Robert and Paul, team members view problems through different perspectives. Robert recognizes that Paul views the problem through the perspective of a policy maker, and Robert urges him to consider it from an alternative viewpoint, such as “the company that come in and solve the problem.”

These episodes also illustrate two general features: the need to 1) shift flexibly between perspectives and 2) recognize these perspectives. Deliberately and explicitly shifting between multiple perspectives assists teams in developing shared understanding of the experiences, thoughts, and actions of diverse stakeholders across different contexts.

Simulation

Interdisciplinary problems are inherently complex, characterized by multiple interrelated causes, where the effects of the problem (symptoms) are often mistaken for problems. Therefore, being able to explore interrelationships between causes and effects and understand the consequences of possible changes holistically is important. Further, complex problems often require reaching a shared understanding of the possible futures; thus, experiences in which such understanding can be grounded are not readily available. For this purpose, teams often created and explored different scenarios. The aim of the simulation move is to elicit and explore multiple possible cause and effect relationships together with their consequences.

The episodes below illustrate the simulation move. To construct a shared understanding of the problem, the team has to construct a shared understanding of the current context for customers purchasing or using services that involve disclosure of their data. Paul makes three different attempts at this. First, Paul suggests using a concrete scenario where Andrew takes the perspective of a bank and Paul takes the perspective of a customer wanting to get a loan from the bank right now. Paul uses this to highlight that he would need to disclose “all the personal information” to Andrew (the bank). Paul then switches to an imagined future where he already has a relationship with the bank (Andrew) (Figure 8).

Figure 8. Paul attempts to create a simulated future scenario.

Paul attempts to convey multiple points: 1) due to his existing relationship with the bank there is already “some guarantee,” and 2) because of this guarantee, in a future where open banking exists, Paul does not need to give Robert “incrementally more data” and his data “only stays with” Andrew. Paul is attempting to create a simulation where his purchase of services from a third party (Robert) in a world where open banking is mandated, would not require disclosure of data to that third party. Verbal interactions alone erase Robert’s role in this simulation. Paul shifts his visual attention toward Robert when explaining that he doesn’t need to “give you (Robert) incrementally more data.”

On his second attempt, Paul turns to the whiteboard wall to give a different example: purchasing travel insurance. However, he abandons this example. On his third attempt, Paul positions himself as a customer using the website Skyscanner to find and book a cheap flight and uses the whiteboard to illustrate the changes over time (Figure 9).

Figure 9. Paul attempts to illustrate changes over time.

Paul continues the simulation by articulating how using the Skyscanner website right now creates challenges (Figure 10).

Figure 10. Paul uses the simulation move to demonstrate challenges.

These episodes collectively illustrate one general feature: construction of shared understanding of the problem requires grounding in multiple concrete experiences familiar to all team members, irrespective of their disciplinary backgrounds. Perspective-taking and simulation games allow teams to ground abstract issues (e.g., trust in digital services) in concrete experiences, which may involve knowledge as well as feelings and emotions (e.g., “feeling betrayed”).

Knowledge validation

Knowledge sharing, perspective-taking, and simulation moves were found to be non-judgmental. During these moves, there were no major communication breakdowns and differences in opinions were reconciled by transferring to the negotiation game (see below). These three moves contrast sharply with knowledge validation. The main area where team members were more likely to have conflicting perspectives, when constructing their shared understanding of a problem, was related to what they considered to be feasible, valid, reliable, trustworthy, or good enough knowledge.

Differences in what team members consider to be trustworthy knowledge and ways of constructing knowledge (e.g., framing problems based on thematic analysis of qualitative data or on analysis of industry reports) surfaced in relation to various kinds of knowledge (e.g., from research reports, client briefs, possible future scenarios). This move is illustrated in the episode below.

The team is creating a presentation for their class assignment which will include an explanation of their problem. Andrew has questioned whether the problem they are trying to solve actually exists. Paul (the PhD student in Business) often shared different kinds of published consultancy reports with the team. In this episode he introduces survey data from such reports to support the existence of the problem, explaining that the report contains a survey methodology. He explains that he has incorporated the survey results into the presentation.

The report containing these survey results was created by professionals within the finance industry (Paul’s professional field). However, Andrew questions the survey results, explaining that he wants to conduct surveys of his own (Figure 11). In this example, he gives reasons in support explaining that “if I don’t see it I don’t believe it,” adding “That’s the only thing that’s true.”

Figure 11. Andrew expresses his view on valid and good enough knowledge.

This divergence in views on what constitutes valid and good enough knowledge is further emphasized by Robert in a later discussion when he questions whether Paul’s strategy for constructing a knowledge base for formulating a problem is the best, pointing instead to the suitability of knowledge derived by less formal means (Figure 12).

Figure 12. Robert suggests the use of less formal approaches.

Here, Robert points to Paul’s inclination toward formal methodologies, attributing this to Paul’s academic pursuits as a PhD student, while contrasting this with an inclination toward more informal approaches. This could also hint at Robert’s own values, shaped by his background in Design, where informal approaches (e.g., conversations with, or informal observations of, users) are often used to understand problems.

The validation move is not always straightforward and in multiple cases, team members would support their reasons or grounds with expert opinions, such as from textbooks, documentation or recommendations provided by the teacher. In cases where teams were unable to reach an acceptable outcome (i.e., reached an impasse), they would defer to authority opinion, which was primarily the teacher.

These episodes also illustrate a second inherent challenge faced by interdisciplinary teams: differences in what team members consider trustworthy knowledge and ways of constructing knowledge. This challenge emerges alongside the previously identified challenge of the varying willingness of students to play the shared understanding game throughout problem-solving.

Negotiation

Central to the negotiation move is the team’s initial acknowledgment that individual members may have different views and that an explicit discussion of their formulation(s) of the problem may help reach consensus—the most desirable exit condition of the game. This can take different forms: from formulating one statement and refining it through joint dialogue, to integration of different formulations or voting on the “best” or “favorite” statement.

The negotiation move is illustrated through extracts from one episode. In the first extract, Andrew synthesizes the perspectives into a refined problem statement that considers the use of consumer data by third parties with trust and consent (Figure 13).

Figure 13. Andrew synthesizes perspectives into a refined problem statement.

Discussion ensues after Paul asks Andrew for clarification of what he means by big organizations and the team shifts their focus to discussing solutions (transferring to a different epistemic game), before returning to the problem: Andrew presents another iteration of the problem statement while writing it on the whiteboard wall (Figure 14).

Figure 14. Andrew presents another iteration of the problem statement.

Robert emphasizes the concepts of “trust,” “transparency” and “confidence” while writing them on the whiteboard wall. Next, he turns to Paul and asks for his input (Figure 15).

Figure 15. The Trust Team collaboratively refine the problem statement.

The discussion, primarily between Andrew and Paul, with Robert’s interjections, leads to a collaborative refinement of the problem statement, reflecting a synthesis of their different perspectives on the problem.

This iterative process exemplifies the negotiation move, which often involved transferring to other epistemic games, focused on solutions. Shared external representations played an important role in this move as the team recorded each iteration of the problem statement in writing on the whiteboard wall. This supported the negotiation and integration of diverse viewpoints, while also making the evolution of the problem statement explicit. Through this negotiation, diverse viewpoints are integrated into a consensual problem statement, before reaching the end of the game. The problem statement negotiated by the Trust team is “how can consumers trust and give consent to third parties to use their data.”

The problem statement represents a synthesis of insights influenced by different disciplinary perspectives, each contributing to the understanding of the problem. Ensuring data privacy and safeguarding against unauthorized access, which is discussed but not directly reflected in the agreed upon statement, resonates with the discipline of information technology. The focus on ensuring customers trust organizations and that those organizations secure consent aligns with ethical business practices. The insights into trust-building also resonate with the design perspective. Together, these perspectives converge in the problem statement.

Transfers

The most commonly identified type of transfer seen across teams was to constructing a shared understanding of the solution, highlighting the interconnectedness of those two aspects. The following episode illustrates the team, having transferred to developing shared understanding of the solution, transferring back to the problem (Figure 16).

Figure 16. Transferring to constructing shared understanding of the problem.

Exit conditions

The primary exit condition for the shared understanding game was reaching a consensual problem statement, indicating an agreement amongst team members. If the teams successfully cultivated an understanding that culminated in a consensual problem statement, the game was considered (at least temporarily) concluded. In the Trust team, the negotiation move ended with the problem statement that was solidified through verbal consensus and a whiteboard wall inscription, as illustrated in the previous episode where Andrew recorded the agreed-upon problem statement on the whiteboard wall. Andrew also drew a large tick next to the problem statement, indicating this is what they have agreed upon (Figure 17).

Figure 17. Whiteboard wall inscription of agreed upon (ticked) problem statement.

This action is followed by a short period of silence in the team, before they go on to discuss whether they should focus next on solutions, use cases, or personas.

However, the complex dynamics of team problem-solving often meant that a once-agreed-upon problem statement was continuously revised. As teams engage in the construction of the solution, they find the need to re-visit and adapt their shared understanding of the problem, even as solutions are in the process of development and testing. Despite the ideal exit condition being an agreed-upon problem statement, teams found themselves ending the construction of shared understanding game prematurely. Often, this was attributable to logistical constraints such as running out of allocated meeting time or reaching an impasse in discussions. Pursuing a shared understanding of the problem, thus, remained a recurrent goal.

Discussion

Our study highlights the richly dynamic nature of student teams’ constantly evolving shared understanding. Nersessian (2012), writing about concept formation in research labs, described a similar process metaphorically as “building the plane while it is flying—and with only a vague idea of what a flying vehicle might look like.” (p. 229). The same metaphor can be applied to describe the dynamic relationship between solving complex interdisciplinary problems and constructing a shared understanding of these problems in students’ teamwork (Dorst & Cross, 2001).

Overall, the shared understanding of a complex problem is a temporary state reached in a situated moment and serves to guide the team’s subsequent problem-solving action. Our findings demonstrate that shared understanding of the problem rarely persists, it is questioned and reconstructed as new knowledge emerges; however, this process is not without its own internal “grammar.”

Key moves and epistemic functions

Expanding existing studies that examined common interactional structures in interdisciplinary teamwork (Goodwin, 2018; Hall et al., 2002), we focused on how graduate student teams construct shared understanding of interdisciplinary problems. By combining an epistemic games theoretical lens with an ecological analytical perspective, we gain insight into the characteristic moves within the shared understanding game, and how they are enacted through multimodal distributed interaction during students’ joint activity.

Our analysis shows that while the shared understanding game played by students during their joint attempts to understand and articulate complex problems is itself complex, it aligns with the patterns observed in epistemic games within professional and other collaborative learning contexts where people engage in situated meaning-making (Markauskaite & Goodyear, 2017; Odden & Russ, 2018). Such games are characterized by moves with distinct epistemic purpose and game-play with shared characteristic features.

The epistemic games theoretical lens enabled us to examine and articulate the five main moves that shaped students’ construction of shared understanding (Table 2). These moves served three main epistemic functions: explication, grounding and juxtaposition. Knowledge sharing and knowledge validation games enable explication and sharing of various kinds of knowledge and personal understanding related to the problem. This includes students’ epistemic understanding of what counts as trustworthy or “good enough” knowledge for making their decisions and framing problems. Perspective-taking moves enable grounding this knowledge in concrete, easier-to-grasp experiences and contexts. Simulation and negotiation moves supplement this by enabling juxtaposition and connecting these perspectives together. The former moves allow exploration of the problem from different insider perspectives (service providers, users, etc.); the latter moves enable exploring how the problem could be framed from different problem-solvers’ perspectives (i.e., team members). These moves are critical in interdisciplinary problem settings, where the exploration of problems from multiple perspectives and integration of different viewpoints is necessary (Boix Mansilla, 2017).

Our analysis of how this epistemic game is played reveals that student gameplay is not without challenges. These challenges relate to navigating epistemic differences and navigating disciplinary, spatial, and temporal boundaries, which we discuss below.

Navigating epistemic differences

Students’ pursuit of shared understanding reveals epistemic differences regarding when the problem should be formulated and how (e.g., at the start and by the team vs. later and after doing initial research or consulting users). These differences were often left unnoticed by the students, resulting in unproductive tensions and disagreements. Overall, we observed across all teams that students from some disciplinary fields (e.g., design and business) sought to continually revise and refine the problem statement throughout problem-solving, whereas others (e.g., IT and science) saw this as going “back to the drawing board.”

Students’ pursuit of shared understanding also highlighted epistemic differences related to the trustworthiness of knowledge—these differences arose across all four teams. While the sources of these differences were not explicitly identified, there was a notable alignment between students’ expressed views and their respective disciplinary and professional backgrounds and experiences. For example, students from business were more likely to look for and trust industry reports and design students were more likely to trust direct empirical evidence, or informal ways of knowing, including personal experiences. In some cases, they were presented as rigidly held positions, while in others they served as starting points for negotiation and dialogue within the team.

These findings reflect epistemic differences found in interdisciplinary research teams, regarding what constitutes good knowledge and how it should be produced (Freeth & Caniglia, 2020). Research has pointed to the role boundary objects can play in addressing these epistemic challenges (Freeth & Caniglia, 2020) and creating shared understanding (Pennington et al., 2016). The lack of awareness regarding disciplinary differences has been reported in both interdisciplinary student and research teams (Freeth & Caniglia, 2020; Horn et al., 2022) as an issue that needs to be addressed. Our findings provide concrete examples of how these epistemic differences manifest when students attempt to construct a shared understanding of interdisciplinary problems, including when and how shared understanding is constructed.

Navigating disciplinary, spatial, and temporal boundaries

Construction of a shared understanding of complex interdisciplinary problems inevitably requires students to use strategies that enable them to gain insight into real-world challenges, and each other’s disciplinary framings of these challenges, without having real-world experiences of these challenges. In newly assembled interdisciplinary teams—which interdisciplinary student teams in educational settings often are—this has to be done without having the same foundational disciplinary, or other, shared knowledge which established interdisciplinary teams may have already developed. Therefore, grounding shared understanding in concrete social and material experiences and joint situated action plays a critical role in this process. It is important to note that while disciplinary boundaries are critical in interdisciplinary problem-solving, other significant contextual boundaries exist and need to be traversed. They include spatial boundaries between the potential solution of a problem and the real-world situations in which the problem is encountered, as well as temporal boundaries (past, present, and future). Some of these boundaries need to be crossed simultaneously, requiring team members to be fluent in switching between and weaving different epistemic moves. Existing literature mainly focuses on discipline-related obstacles in interdisciplinary problem-solving and teamwork (MacLeod, 2018; O’Donnell & Derry, 2005), whereas the simultaneous presence of other boundaries and obstacles is rarely noticed and addressed.

Multimodal and distributed game-play

The shared understanding epistemic game is profoundly multimodal and distributed. The game entwines language, shared representations, material environments, and students’ bodies with just-in-the-moment emerging shared experiences. Studies of epistemic games often focus on language (Morrison & Collins, 1995; Odden & Russ, 2018). However, solely focusing on language can obscure the multifaceted ways in which team members participate and construct shared understanding. We saw an example in the simulation episode, where the use of gaze by Paul, together with language, indicated he was positioning Robert as the third party organization. Gaze and other non-verbal cues such as gestures are critical modes of participation in game-play that would otherwise be overlooked.

Our findings also shed light on the patterns of interaction between students’ shared understanding and creation of shared external representations. As illustrated throughout the knowledge sharing, simulation, and negotiation episodes, shared understanding and representations symbiotically co-evolve. Shared visual representations, such as those on the whiteboard wall used by the Trust team, are not only flexible but integral to the process of constructing shared understanding. They create a shared visible space that facilitates joint attention, supports verbal explanations, demonstrates relationships between concepts, and serves as an enduring reference point that can be built upon by the team. This highlights the symbiotic and iterative relationship between shared external representations and constructing shared understanding of complex interdisciplinary problems.

Concluding comments

Investigating the construction of shared understanding related to interdisciplinary problems is a challenging task due to its complex and dynamically unfolding nature. Our study acknowledges the diverse approaches to conceptualizing and studying shared understanding that have been attempted across different communities, reflecting the rich, multifaceted nature of this concept in various, especially interdisciplinary, contexts. This diversity presents a significant challenge in synthesizing all threads into one cohesive perspective. Our research makes one such attempt by bridging the ongoing work across two fields, advancing the discourse on interdisciplinary teamwork and shared understanding within the learning sciences. We offer a novel perspective to examine the process of constructing shared understanding by integrating the epistemic games theoretical lens with an ecological perspective, expanding the analytical focus to identify broad generative strategies.

The epistemic games theoretical lens enabled us to examine and articulate the game-like patterns that shape this process, while also capturing its defining features. On its own, the epistemic games perspective might overlook the dynamic interplay between language and the broader ecological context, failing to account for how elements within the environment, and multimodal interactions, contribute to the understanding of complex interdisciplinary problems. Conversely, a solely ecological analytical perspective would miss the nuanced, game-like patterns that the epistemic games perspective sheds light on. However, the combination of the epistemic games theoretical framework with the ecological perspective enables us to expand the analytical focus beyond language to multimodal interaction and gain a nuanced insight into the distributed nature of students’ construction of shared understanding, while also identifying recurring moves. These results, observed across teams and interdisciplinary problems, indicate that these moves are not peculiar to a particular interdisciplinary problem or a particular interdisciplinary team composition, but are broad generative strategies used to produce actionable understanding of what the problem is.

The naturalistic nature of the study does not allow us to make definite causal inferences, such as how specific students’ disciplinary and professional backgrounds and knowledge affected their participation and their contributions to the construction of shared understanding of the problems, or how the nature of the complex interdisciplinary problems that students attempted to understand affected their game-play. In our cases, interdisciplinarity was a feature of both the problems and the teams. The given problems were interdisciplinary and the teams consisted of students with diverse backgrounds, due to the disciplinary and professional specializations and fields in which they worked. However, these results reveal that differences in students’ perspectives on when and how they should construct shared understanding of the problem, and on what kinds of knowledge they should draw—differences that are unlikely to be independent of their backgrounds—create challenges when they engage in the construction of shared understanding of complex problems.

As framing and solving complex problems often go hand in hand, future research should focus on student teams’ co-construction of shared understanding of problems and potential solutions to interdisciplinary problems. This could provide insight into the catalysts for transferring between games, while also mapping the key moves involved in constructing shared understanding of potential solutions, their relations and epistemic functions.

There are several practical implications for interdisciplinary teaching and learning. It may be beneficial for teachers to support an interdisciplinary team, when it is constructing a shared understanding of a problem, by giving the team sub-tasks that encourage specific epistemic moves. This could be better than leaving productive moves to emerge spontaneously, which may not necessarily happen. For example, such sub-tasks may include perspective taking and simulation within the shared understanding game for helping students explore a topic from different perspectives and ground abstract problems in concrete experiences. Another example would be sub-tasks designed to help students notice epistemic differences within their interdisciplinary teams: promoting timely identification of differences about what knowledge and ways of knowing are considered valid. Likewise, scaffolding negotiation and articulation of the problem may assist teams to identify and reconcile differences in their individual understandings, including through periodic revision and re-articulation of the problem.

Disclosure statement

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

Additional information

Funding

This research was partly funded by the Australian Research Council through Discovery Project grant DP200100376 (Developing interdisciplinary expertise in universities).

Notes

¹ It is important to note that the term “epistemic games” has been used in several different ways in the learning sciences. As Markauskaite and Goodyear (2017) observe, these notions have ranged from those strongly focused on cognitive strategies and mental schemas (Perkins, 1997) to those focused on forms of discourse (e.g. Collins & Ferguson, 1993) and from those focused on broad schematized strategies and patterns that are used across contexts (e.g. Collins & Ferguson, 1993) to simulated games, rich in authentic detail, used as a pedagogical strategy to develop expert-like capabilities and identity in specialized domains (Shaffer, 2006). In this paper, we draw upon Collins and Ferguson’s (1993) approach, and do not describe other perspectives that are outside our scope.

² Reinsurance is the practice of an insurer reducing risk, particularly from catastrophic events, by purchasing insurance from other insurance companies. Placement is the process of purchasing or renewing insurance.

Appendix

Notations used in the transcripts (based on Jefferson, 2004)

Table

↑	Upward arrow represents a marked rise in pitch.
↓	Downward arrow represents a downward shift in pitch.
:	Prolongation of the immediately prior sound.
(.)	A brief interval (± one tenth of a second) within or between utterances.
(1.0)	Numbers in parentheses indicate pauses time to the nearest tenth of a second.
underlined	Underscoring indicates a word(s) uttered with added emphasis
[]	Extended square brackets mark overlaps between utterances
(())	Transcriber’s descriptions of multimodal actions (e.g., gesturing, writing)