Curiosity to Question: Tracing productive engagement in an interdisciplinary course-based research experience

ABSTRACT Background Course-based research experiences (CBREs) are highly valued for science learning and research. Most are discipline-centered, but there is a great deal of interest in developing them to promote interdisciplinarity. Yet, we have much to learn about how CBREs work, and even more to learn about how disciplinary diversity operates as an element of design to promote learning and research. Method This mixed methods case study triangulates data on student experience, research networks, research artifacts and fieldnotes from participant observation to understand how disciplinary diversity factors into productive engagement for learning and research in an interdisciplinary course-based research experience (I-CBRE). Findings A common boundary object (scientific paper) mediated developmental interactions among individuals both within and beyond the course. The boundary object also facilitated the productive engagement of students as instrumental actors in broader interdisciplinary research networks. Contribution This study shows how an “object tracing” approach can be used to examine productivity in disciplinarily diverse scientific contexts, and reveals some of the distinctly syncretic moves that participants deploy to make progress as learners and researchers. Extending work by Engel and Conant, the study emphasizes the importance of designing for “productive syncretic engagement” in research experiences characterized disciplinary diversity.


Introduction
Education researchers have demonstrated that course-based research experiences (CBREs) can be highly engaging and inclusive modes of instruction at the university level (Bangera & Brownell, 2014), capable of contributing not just to gains in conceptual knowledge and skills, but also enhanced science identity and diversity in STEM learning pathways (Auchincloss et al., 2014).At the same time, researchers have noted that we know much less about the design-relevant processes of learning in CBREs (Linn et al., 2015).Furthermore, the vast majority of CBREs are discipline-specific, designed and carried out in disciplined academic programs in pursuit of questions and evidence of disciplinary priority.Despite the broadly recognized need for scientists who can work, talk and pursue questions across and outside of disciplinary boundaries, there have been few systematic studies of the how a CBRE might be approached to create ecologies that mix different disciplines to promote students' development as researchers.This study aims to make such a contribution by examining how students are productively engaged as learners and researchers in a university-level research methods course called Curiosity to Question, a CBRE that was specifically designed to leverage disciplinary diversity.

Literature review
Scientists and science educators have long worked to develop researchfocused learning experiences that reflect and leverage our university-and industry-based research infrastructures and practices.The US National Science Foundation (NSF), for instance, has funded the development of "real-world" student research experiences since 1958, with the NSF Research Experiences for Undergraduates (REU) program investing $1.12 billion in support for student research experiences from 2002 to 2017 (McDevitt et al., 2020).Many studies have highlighted the utility of student research experiences in preparing students for STEM careers (Kuh, 2008).Ahmad and Al-Thani (2022) categorize undergraduate research into traditional, course-based, and hybrid models.Traditional experiences, often summer internships, offer practical lab work under mentorship.Coursebased experiences, increasingly common since 1956 (Fromm, 1956 as cited in Ahmad & Al-Thani, 2022), typically introduce students to research within their disciplines during early undergraduate years.Hybrid models blend these approaches to overcome time, mentorship, and inclusion limitations in academic programs.
Considering this longstanding interest, large investment and apparent value in student research experiences, and considering the way that these environments link and mediate students' movement from academic to professional science work, such research experiences would seem to be critical environments for refining our understanding of learning goals and processes.

Learning, research and disciplinarity
Disciplines are fundamental to how science and scientific research is organized as a professional and academic pursuit (Latour, 1987;Pickering, 1995), and they fundamentally pattern the funding, goals, design and implementation of university research experiences for students.By couching academic science education in disciplinary communities, domains and practices, students are prepared to think, see and operate in ways that conform to the disciplinary environments they aspire to join as professionals.Immersion in authentic laboratory and field contexts supports affective, embodied learning and skills development (Mogk & Goodwin, 2012), disciplined and professional ways of seeing the world (Goodwin, 1994;Stevens & Hall, 1998) and social relations and capital that facilitate transfer to post-academic careers (Aikens et al., 2016).
But not all science learning or research is patterned by disciplines in the same way.In describing scientists, their work and the environments in which they grapple with and work across boundaries and disciplinary "microfractures" (Pickering, 1995), the literature has deployed a number of terms: multidisciplinary, interdisciplinary, transdisciplinary, and antidisciplinary (Alvargonzález, 2011;Choi & Pak, 2006;Finch et al., 2021;Klein, 2017;Pickering, 1993Pickering, , 2008)).Finch et al. (2021) place these notions on a spectrum, with multidisciplinarity characterized by minimal integration, interdisciplinarity more integrated, and transdisciplinarity characterized by deep integration.Pickering characterizes antidisciplinary science as a transdisciplinary process of "eruption" in response to novel and compelling "ontological theatre," playing out "across the disciplinary map" to describe and engage more and more of the world in terms of the new ontology (2013, pp. 209, 211).Learning to engage with such disciplinary complexity and diversity at whatever point along this spectrum is seen as essential in promoting science that is innovative, complex and generative in its pluralism (Bangera & Brownell, 2014;Stein et al., 2008) and that values, engages and responds to a diversity of student and stakeholder voices, experiences and agencies (Bang, 2015;Finch et al., 2021).
We define "interdisciplinary" activity in a research experience as activity that articulates people, knowledge, practices, and ways of knowing and learning from two or more disciplines or established fields of study (Boix Mansilla, 2006).We are interested in studying and designing for interdisciplinary activity that is (1) disciplined, that is, embodying practices, concepts, tools, methods, theories, findings, and forms of communication developed by disciplinarians; (2) integrative, where individuals and collectives make progress through "productive articulation" (Boix Mansilla, 2006, p. 19) of disciplinary practices and understandings; and (3) purposeful, where interdisciplinary engagement is sought not as an end in itself, but a means to other valued ends (e.g.student learning and scientific productivity).

Disciplinary diversity in research experiences
The role of disciplinary diversity in fostering effective interdisciplinary research and learning is a growing focus in course-based research experiences (CBREs).We define a CBRE roughly along the lines of Auchincloss et al. (2014), as an educational experience where students actively engage in scientific inquiry that matters to the wider scientific community, iteratively solve problems, make progress, and share findings that contribute to collective knowledge.Building on Boix Mansilla (2006), an interdisciplinary CBRE (I-CBRE) also purposefully promotes progress in learning and research through the engagement of people, practices, concepts, tools, methods, theories, findings, and/or forms of communication from different disciplines.While I-CBREs may vary significantly in the degree to which they integrate disciplinary practices and perspectives, they all engage some degree of disciplinary diversity, so we focus on this as a basic element in their design.
Many CBREs are documented, but only a few are interdisciplinary by design.Disciplinary diversity presents unique challenges, particularly in managing and adapting to complexity within the limited duration of a typical semester.Pedwell et al. (2018) and Lau et al. (2019) illustrate that while interdisciplinary CBREs offer rich learning opportunities, students often face hurdles in teamwork, requiring adjustments in faculty mentorship and support to accommodate diverse research projects and backgrounds.
Within the broader realm of interdisciplinary research experiences organized outside of courses (e.g.Canaria et al., 2012;Danowitz et al., 2016;Maaz et al., 2022), Maaz et al. (2022) reported success in promoting interdisciplinary research and learning by extending the timeframe of the research experience and taking a cohort-building approach.This enabled students to engage more deeply with their projects over the course of a five-year study, suggesting that extended durations may better support interdisciplinary learning.Majka et al. (2023) found that a CBRE designed around principles of collaboration, discovery/relevance and iteration could lead to psychosocial gains among STEM students.While the authors presented promising associations between each key element and psychosocial outcomes like belonging to school and science self-efficacy, interdisciplinarity was not explicitly assessed as a factor in these outcomes, highlighting an area for further research.
These studies collectively underscore the importance of thoughtful design in CBREs, recognizing the difficulties of aligning diverse disciplinary perspectives within the constraints of academic schedules.Furthermore, while the value of interdisciplinary and transdisciplinary approaches is widely recognized, there is a notable gap in operationalizing and assessing disciplinary diversity in relation to specific learning outcomes.Overall, we have found that current scholarship illuminates significant challenges that remain in conceptualizing and designing research experiences that (1) engage disciplinary diversity and ( 2) that can be carried out in a course-based mode.

Goals and questions
This study takes up the case of an I-CBRE that is disciplinarily diverse by design, examining how disciplinary diversity factors into how students are engaged as learners and authentically productive researchers.Specifically, we focus on a set of three interrelated research questions (RQs): RQ1: What does learning and research engagement look like in the I-CBRE?

RQ2:
In what ways can engagement be understood as productive?RQ3: How does disciplinary diversity factor into productive engagement?

Theoretical framework
As all CBREs are designed to support student learning through the production of authentic research, we need a framework and analytical approach that attends to engagement for learning, scientific productivity, and-within our more specific context of an I-CBRE-the role of disciplinary diversity.Our general approach is to put "developmental" perspectives on learning and "instrumental" perspectives on scientific research into dialogue, examining how an important boundary object (scientific paper) mediates productive engagement among disciplines at different "moments" in the course.Engle and Conant (2002) put forward the well-known goal of productive disciplinary engagement (PDE) as a way of understanding and designing pedagogy for disciplinary science learning.Courses that promote PDE provide learners with opportunities to engage productively in day-to-day activities, roles and practices that reflect those of the disciplines prioritized in our school curricula (i.e.STEM)."Productive disciplinary engagement can be fostered by designing learning environments that support (a) problematizing subject matter, (b) giving students authority to address such problems, (c) holding students accountable to others and to shared disciplinary norms, and (d) providing students with relevant resources" (Engle & Conant, 2002, p. 399).From a traditional PDE frame, the norms and practices with which students are engaged and evaluated as "productive" learners extend from the specific disciplinary centers of authority and communities of practice that students aspire to join.Productive engagement involves disciplining (e.g.Stevens & Hall, 1998) students in ways that are likely to promote, for instance, their near-or long-term movement from the periphery to the center of a scientific community of practice.

Productive engagement
However, many STEM work and learning environments are characterized by multiple disciplinary domains, and they may not operate as a communities or share a common set of routine practices (Engeström, 2007).In an I-CBRE that invites a diversity of potentially competing disciplinary norms and perspectives from a variety of social worlds, what does it mean to "hold students accountable" for their learning and knowledgeable performance?What counts as "productive engagement" and how do we design for it in a disciplinarily "impure" learning ecosystem?To understand learning and productivity in an I-CBRE we need to move outside of the canonical sociocultural analytical boundary of the community of practice, and look at ways of theorizing learning and change in complex, heterogeneous, decentralized and multicentered ecosystems or networks.For this we turn to two different-but-related sociomaterial perspectives: cultural historical activity theory and actor-network theory.

Developmental and instrumental productivity
Cultural-historical activity theory (CHAT) and actor-network theory (ANT) have both been used to examine interactions that produce complex networks (Spinuzzi, 2008) for learning and research.Both of these sociomaterial perspectives theorize-albeit in different ways-the disciplinary landscapes and infrastructures within which science is undertaken as complex, heterogenous, human-material networks, and both attend to how issues of contradiction and noncoherence are managed productively in such contexts.
From a CHAT standpoint, where learning plays out in the context of object-oriented activity systems, a complex, disciplinarily diverse learning and research network can be construed as a network of "chained" or overlapping activity systems, or "activity networks" (Spinuzzi, 2008).Such networks are inevitably multivoiced, and the shared objects that orient joint work are polycontextual, orienting and organizing network activity in dialogic tension (Engeström, 2009;Kajamaa & Lahtinen, 2016;Spinuzzi, 2008).The contradictions that arise within and between activity systems serve as focal-points and catalysts for learning and change, driving cyclical, collaborative, and developmental activity over time (Engeström, 1987;Foot, 2001;Foot & Groleau, 2011;Miettinen, 1999).By focusing on contradictions, CHAT-informed analyses help us gain insight into where students and researchers must engage in mutual learning and change, and how they work together to journey across a "zone of proximal development" to collectively achieve shared objects.From a CHAT perspective, productive work develops students and expands the objects of their research activity over time in ways that are irreversible.Students fundamentally change themselves, the activity systems within which they operate, and the objects of their activity through their developmental work.Such processes of networked learning and research are developmentally productive.
ANT, on the other hand, theorizes network assembly, enactment and transformation as a process of translation, a term standing for the way that diverse human and non-human entities-actants-are brought together in ways that translate them into actors in a network of coordinated things and actions (Latour, 1987).According to Callon (1984), translation involves four "moments:" (1) Problematization is the moment when an initial actor or set of actors specifies an issue or problem in a particular way, and offers itself as an indispensable part of the solution, interposing itself as an "obligatory passage point," spokesperson, or gatekeeper.(2) Interessment is the entangling and knotting together of various actors as potential stakeholders in the issue as problematized, defining the ways they could be linked and the kind of roles and relations they may negotiate.
(3) Enrollment is achieved when interessment is successful, that is, when actors have negotiated and defined their roles and relationships in the problem space.(4) Mobilization describes how stakeholders ultimately achieve their collective solution by agreeing to a collective representation of their interests and enforcing the negotiated roles and relations.
Translation produces a settlement, that is, a more-or-less stable association of actors that achieves competence in assemblage (Callon, 1990).
A particularly stable, well-coordinated, recognizable and useful association of actors may itself be enrolled as an actor and mobilized in different and broader networked contexts for other ends.In such cases, the complex sets of alliances and histories of negotiation that produced the settlement are often ignored or treated as invisible by the broader network; that is, the association is treated as a black box (Latour, 1999).Such black-boxed actors can always in theory be unboxed, the actors unpacked and their internal associations reversed or re-assembled in other ways or with other actors.ANT's instrumental perspective is useful for understanding how students may be productively engaged by being enrolled and mobilized in broader research networks.That is, thier productivity can be understood in terms of their relational position in competent sociotechnical research processes, routines and infrastructures.This differs from developmental understandings of productivity as human learning, as a competent network of researchers and research infrastructure can be disassembled in theory, erasing competence and leaving the individual actors unchanged.
In conceptualizing actor-networks as disciplinarily diverse or "impure" (Law et al., 2014) assemblages, ANT invites a focus on the variety of situated tactics that actors deploy to cope with noncoherence.Borrowing from religious studies, Law et al. (2014) have argued for the importance of attending to "modes of syncretism" (Table 1) when examining how heterogeneous knowledge systems and networks assemble, coexist, compete, and transform within societies.Such modes of syncretism include, for instance, "separation" of competing or noncoherent perspectives across space or time to avoid direct conflict.A syncretic mode of "care," meanwhile, involves dealing with the stress and potential conflict related to noncoherence by committing to iterative, empathetic tinkering in search of provisional fixes.Such modes of syncretism can help us understand how disciplinary noncoherence is managed by student-scientists to make progress amidst disciplinary diversity.Historical differences and confrontations between CHAT and ANT scholars are well-documented (see Kaptelinin & Nardi, 2006;Latour, 1996;Miettinen, 1999;Spinuzzi, 2008).ANT scholars have troubled the dialectics that underpin CHAT's approach to contradiction (Latour, 1996), while CHAT scholars have critiqued ANT for what they see as a fundamental contradiction between the principle of general symmetry (which would seem to exclude human needs, motives and intentions from analysis) and a persistent tendency toward analyses that emphasize the political and technocratic agendas and machinations of "Machiavellian" actors (Miettinen, 1999).Recognizing such tensions, Spinuzzi (2008) points out that CHAT and ANT do different, but related and potentially complimentary things: While ANT is useful for describing network learning or transformation as the instrumental assembly of power and competency in a network, CHAT focuses on networked learning and transformation as a cyclical, developmental, human-centered process.Furthermore, CHAT and ANT both attend to how objects mediate and pattern activity across social worlds, a point of conceptual overlap that presents itself as useful for carrying out joint analysis of networked learning and change.

Objects in productive engagement
Boundary objects, initially theorized by Star and Griesemer (1989), are things (structures, constructs, toolkits) that exist in multiple social domains to organize information and support practice.They are enacted objects, that is, things that people "act toward and with" (Star, 2010, p. 603), and their characteristic attributes of broad recognizability, local flexibility and immediate relevance to diverse practices make them ideal mediators of interactions between disciplines.Boundary objects have enjoyed a "vigorous academic career" (Trompette & Vinck, 2009, p. a), having been employed to examine learning and work in a wide variety of interdisciplinary contexts (Akkerman & Bakker, 2011;Trompette & Vinck, 2009), including science learning (e.g.Polman & Hope, 2014).
Starr and Griesemer's initial work on boundary objects was fundamentally informed by ANT's translational perspective on science productivity, but emphasized a more ecological rather than ego-centered analysis.Rather than focusing on Machiavellian actors, Star and Griesemer wanted to examine how ecologies of "n-dimensional" objects (Star, 2010, p. 603) mediate "n-way" translations (Star & Griesemer, 1989, p. 412) that hold complex scientific networks together.From an instrumental (ANT) perspective, a boundary object can thus be understood as a nodal actor in an actornetwork, and in particular one that performs a translational role linking multiple other actors in a more-or-less stable alliance.Boundary objects can be understood as instruments that simply splice actors together, enrolling and mobilizing them as a networked association.They are in this sense translational mechanisms through which purposes can be communicated and aligned, and noncoherence among actors can be managed and disguised in ways that facilitate growth and expansion of the network.A boundary object may mediate actor relations in any moment of translation, from problematization to mobilization, or may structure the whole process, helping to "stabilize" networks in the face of controversy, disagreement, and noncoherence.
From a more human-centered, developmental (CHAT) perspective, a boundary object may be taken up in human activity in several different ways.Most fundamentally, it may be the object of activity that motivates the work of a complex activity system.As such, it operates as a "sense-maker" (Kaptelinin, 2005) to orient and impel the transformation of the boundary object itself through the developmental work of one or several interlinked activity systems.Polman and Hope (2014), for instance, have described how science news stories authored by youth functioned as developmental boundary objects at the intersection of multiple social worlds.
A second way that a boundary object may function in developmental activity is as a mediator in an activity system or activity network oriented toward the transformation of a different object.Critically, such a boundary object may function simultaneously as a mediational object in one activity system and the motivating object of another activity system that produced it.In this case, from an ANT perspective, such a mediating object may be understood as a black-boxed activity system, the object of which is taken up instrumentally in other systems (perhaps with little-to-no understanding of its cultural-historical origins).
Third, a boundary object may be understood as what Engeström (2007) calls a "runaway object."In later generations of activity theory, which shift the level of analysis from individual activity systems to activity networks, this term is used by Engeström and others to describe broadly compelling objects that demand and mobilize the "swarming" (Engeström, 2007) of individuals across heterogeneous networks, and which invite developmental "knotworking" among seemingly disparate threads of activity (Engeström, 2008).Due to their perceived urgency and broad relevance, runaway objects are often seen as overwhelming, out-of-control, or potentially disruptive.They have a tendency to take on a life of their own.Engeström (2007) offers the example of the Linux open source operating system and urgent issues like climate change.We argue that in contemporary university research contexts that promote "fast science" (Stengers, 2016, p. 4) and incentivize research collaboration via "publish or perish" cultures and systems, the scientific paper emerges as a distinctly urgent, demanding, all-consuming object of joint work, and as such, can be understood as a runaway object.
From an ANT perspective, a runaway object may be understood as a boundary object that manages to compel the assembly of many actants from diverse areas of a large and heterogeneous network.Assuming ANT's fundamental perspective on human-nonhuman symmetry, runaway objects might be seen as highly agential in sociomaterial context, "demanding" a swarming of diverse actants and many interrelated and competing problematizations.In the context of contemporary research systems and infrastructures that prioritize scientific publication as the key product of research collaboration, scientists may find their collaborative practices and interactions with other scientists to be heavily patterned (some might say distorted) by the overwhelming demands and norms of the runaway object of the scientific paper.

Tracing objects to examine engagement over time
Object tracing is one way of putting CHAT and ANT into a "dialogue" (Spinuzzi, 2008) to theorize an I-CBRE as a sociomaterial patchwork of (1) interconnected human-centered developmental activity systems (e.g.Engeström et al., 1995) and ( 2) assemblages of human and non-human competence (e.g.Callon, 1990).By identifying boundary objects in the network and using them as a locus of integrated analysis, we can describe both developmental and instrumental processes that play out as science learners and researchers grapple with issues of contradiction and noncoherence in disciplinarily "impure" research networks.By moving back and forth between these two different-but-related perspectives (Burke, 1966;Wertsch, 1998) to analyze scientific papers as key boundary objects in an I-CBRE, we can better reveal a variety of important ways that students are productively engaged through disciplinary diversity, highlight different mechanisms of learning and change, and draw together implications for the design of I-CBREs.

Design case
We take a mixed methods (Teddlie & Tashakkori, 2009) triangulation approach to examining the Curiosity to Question (CtQ) course as an intrinsic case (Stake, 1995) of a disciplinarily diverse course-based research experience (I-CBRE), that is, a unique and uniquely valuable case to understand in its own right.The CtQ course was initially piloted by the second author, a natural scientist, in 2017 as a way to engage students in the kind of interdisciplinary research she does in her own lab.In a design-based manner (Barab, 2014), the course was iteratively developed over six semester-length cycles.Students in each of these cohorts developed their own research questions and carried out their own semester-length research projects, culminating in the production of a complete scientific paper.
From inception, the design of the course was guided by the general conjecture that innovative science requires inclusion, engagement and interaction across disciplinary and sociocultural boundaries.The CBRE specifically recruited a mix of graduate and undergraduate students for disciplinary and sociocultural diversity, integrated their learning activities around a set of common boundary objects (Star & Griesemer, 1989), and facilitated feedback and interactions between them and across disciplinary boundaries (Akkerman & Bakker, 2011) in order to produce authentic research for use by others (Figure 1).Specifically, the current iteration of the CtQ model describes a CBRE that leverages the following design elements: and language for collaborative work across perceived and practice relevant boundaries (Akkerman & Bakker, 2011).( 4) Iterative workflows: Iterative, low-stakes, feedback-rich writing and representational workflows that help students construct knowledge, evaluate how it performs and links with others in context.(5) Distributed mentorship: Systematic development of research mentorship and peer support skills and relations within and outside of class.
Each semester, students completed 13 modules (Digital Supplement) sequentially, though activities like peer review often overlapped and were often revisited.(See Clarke et al. (2021) for full module descriptions.)Students collaborated to enhance their understanding of science and scientific research, formulating meaningful inquiry questions, conducting hypothesis-driven investigations, analyzing data visually and quantitatively, and articulating their methods, results, and significance in a scientific paper and presentation.Through group discussions and personal reflections, they contemplated their research experiences, the process of becoming scientists, and their specific scientific communities and disciplines.

Context, population and sample
The CtQ course was offered through a geoscience program at a large southern university in the U.S. A total of 96 students participated in the course over six iterations from 2017 to 2022 (Table 2).Based on student self-reports (n = 57), roughly 47% of the students in the course identified as female, and 33% with groups designated by the NSF as underrepresented in STEM.This was a higher proportion of underrepresented students than the university as a whole, and a higher proportion of female students than the school, but lower than the university.The faculty instructor (second author) was joined by three post-docs working at different times in her lab who provided instructional support for the R modules on visualization of quantitative data.The instructor and post-docs were all sampled for interviews.An education researcher (first author) was a participant observer in the course, assessing learning as it unfolded over the course of the semester and providing writing guidance and feedback to students.

Data sources
We assembled and analyzed (1) a standard survey of student perceptions of learning in research experiences, (2) a name generator survey of student support networks, (3) a purposive sample of post-course semistructured interviews, (4) student generated artifacts of learning (including written student reflections, research papers, peer feedback), and (5) fieldnotes on learning activity taken by the first author over the course of participant observation.

Surveys
To roughly evaluate course effectiveness and guide iterative design cycles, we used a modified version of the Undergraduate Research Student Self-Assessment (URSSA) (Weston & Laursen, 2015) at the end of the course.This measured student satisfaction and learning gains related to knowledge, confidence, skills, and attitudes/behaviors.Responses to open-ended questions about student reaction to the course were downloaded and added to a project database for analysis using the NVivo 14 qualitative data analysis software (Lumivero, 2023).
Anticipating that a wide variety of disciplinary interactions and influences would guide and shape students' work over the course of each semester, a name generator survey (Carolan, 2014) was used to identify important relationships related to mentorship, writing and research support, and in Y4 through Y6 the instrument was adjusted to also capture users of student research.Respondents were asked to identify up to five people for each type of relationship and rate them in terms of degree of importance.Named individuals were then identified and described by the researchers in a post hoc fashion through interviews and public directories, adding attribute data including whether they were formally a part of the course or not.Responses were structured as a sociomatrix for each type of relation in order to visualize course relations as a set of research, writing and mentorship networks.

Interviews
Interviews were carried out in a purposive fashion using an interview guide (Merriam, 2009).They were generally conducted immediately after the course conclusion, prior to the start of the new semester.Initial sampling was focused on understanding student experience at a diversity of levels (graduate and undergraduate students), disciplinary identifications (as determined through field observations) and varieties of performance in the course (in terms of perceived struggles, successes, and final grades).In later iterations, sampling was also focused on understanding emergent issues related to course design and questions about types of learning interactions and relations, based in part on visualization of support networks.Semistructured interviews were carried out in-person, over the phone and via Zoom.Interviews were conducted and recorded by the first author, electronically transcribed and edited for accuracy.Interview sampling was carried out simultaneously with analysis, informing sampling in an ongoing manner, and allowing for member-checking of emergent questions and findings.Transcribed interviews were added to the project database.

Artifacts
To directly assess student learning, we collected artifacts of their work, including their research papers, the set of slides used by each student to orally present their work, and three reflections on their learning activity done at the beginning, middle and end of the course.These artifacts were downloaded from the online learning management system used in the course, and added to the project database.

Fieldnotes
To ground self-reported learning experiences and gains in an account of learning interactions, the first author took fieldnotes as a participant observer in the course.The first author's role varied over the course of the design cycles, initially involving the provision of instructional support for scientific writing, and eventually evolving into a general support role, which included participating in whole class discussions and breakout group settings.Fieldnotes (Emerson et al., 2011) were taken to document learning and research activity as it unfolded in classroom, small group, and online Zoom contexts (see Table 2 for an overview of observation contexts).Handwritten fieldnotes were taken by the first author while in the classroom, and then elaborated as typed fieldnotes after class.Handwritten fieldnotes generally captured descriptions of situations and events as they unfolded in the course, including descriptions of group-and whole-class interactions and notes on classroom discourse that included quotes and paraphrased discussion.Typed fieldnotes also involved "analytical writing" (Emerson et al., 2001) to develop initial insights, interpretations and analyses beyond the written descriptions.All notes were ultimately added to the project database as primary observation data for analysis.

Data analysis
The specific analytical approach presented here-rather than reflecting strict fidelity to an analytical method used across all design cycles-is the result of a careful process of "tinkering" with data across six annual cycles for both design and theory-building, a process intended to surface and describe which course design goals were important to participants, to develop ways of articulating and examining tensions related to those goals, and to link the analysis to useful theories of learning and change.Through this process, researchers arrived at the specific "object tracing" approach described here as a way of conducting a temporal analysis of what became the primary research focus in the interdisciplinary course: productive engagement.The approach was inspired by ANT-based object-following strategies (e.g.Czarniawska, 2014) and CHAT-based investigations into the circulation of communication genres within organizations (Spinuzzi, 2003).
Qualitative and quantitative data were analyzed in an ongoing and recursive fashion over the course of the study.For each design cycle, fieldnotes were the first data to be collected and analyzed.This involved reading and typing fieldnotes, writing analytical memos, and constructing narrative accounts of observed activity.Memoing (Glaser, 1978) involved noticing and connecting themes in the data, and working back and forth between theory and observations to develop tentative interpretations.Narrative analysis involved attending to tensions (contradictions, issues of noncoherence) in course-related activity, and describing "the passage from one equilibrium to another" (Czarniawska, 1998, p. 19) as participants worked to manage and resolve different kinds of tensions in the context of different kinds of activities.Memos and narratives were themselves coded in NVivo for emerging themes and connections to theory.In the first design cycle, a key analytical goal was to articulate the practice-informed course design more explicitly in terms of learning theory, and coding was primarily descriptive and informed by a CHAT perspective on object-oriented learning activity.Initial coding therefore emphasized "tools," "shared objects," "tensions" in learning, and the "resolution" of tensions.
At the conclusion of each semester, we used the R statistical computing environment (R Core Team, 2018) to generate descriptive summaries of quantitative survey data.Additionally, network visualizations were created using the igraph package (Csardi & Nepusz, 2006) in R.These summaries and visualizations informed purposive sampling and the construction of interview questions, providing insights into self-reported learning outcomes, affective responses to the course experience, and network dynamics.
Simultaneously, we analyzed open-ended survey responses and reflections, coding them according to evolving design-salient categories.Initially, our focus was on understanding the diversity of student learning and research goals ("objects" of activity), the roles and division of labor in distributed "mentorship," and "tensions" in research and writing interactions stemming from disciplinary diversity.
Student artifacts (papers and reflections) were coded based on their highlevel disciplinary orientation, technical subdisciplines, the total number of coauthors, and the academic disciplinary affiliation of the coauthors.This coding provided insight into the extent to which student writing contributed to collaborative research efforts and engaged various disciplines.Over time, the set of high-level disciplinary codes expanded to include geoscience, physics, math, biology, environmental science, education, kinesiology, public health, and urban planning.
Through network visualizations and student interviews, it became clear that research activity was mediated by helpful support networks that expanded well-beyond the boundaries of the classroom, and that this also created specific challenges for students.We began to see the utility of actornetwork theory as a way of understanding the construction of complex research networks, and what it might mean for students to be engaged "productively" in such networks.Having identified productive engagement in the disciplinarily diverse network as a particularly useful analytical focus, coding and memoing in the last three design cycles attended more specifically to the role of "boundary objects" in linking diverse laboratory and classroom practices, to specific points of "noncoherence" introduced by differences in disciplinary perspectives and practices, and to translational processes and concepts that helped describe productive engagement, including "gatekeeping," "inscription," and "black-boxing."

Object tracing
The importance of boundary objects in mediating activity was evident in both earlier CHAT-based and later ANT-inflected analysis, and it became clear that such objects could be used as focal points for integrated analysis of productive engagement.While multiple boundary objects had been coded in the data (e.g. the R platform), the boundary object of the scientific paper was ultimately chosen as the most useful analytical focal point because (1) it was clearly important to most students and therefore salient from a designrelevance perspective, and (2) it was discussed and engaged from the very beginning to the very end of the course and therefore was convenient for longitudinal analysis of learning and change.Individual narratives of participant activity related to the boundary object of the scientific paper at different times in the course were linked to produce a semester-length "object tracing" narrative that examined how the object was manifest and deployed in different ways at different moments to deal with contradiction and noncoherence amidst disciplinary diversity.

Trustworthiness and limitations
Creswell and Poth (2018) describe a number of strategies employed by researchers to enhance the credibility and trustworthiness of their findings, and they recommend that studies employ at least two in their designs.In this study, trustworthiness was enhanced through: (1) Member checking: Emerging narrative analyses were shared with participants during interviews and over the course of participant observation for feedback, helping to integrate participants' insider understandings of how the course operated with the researcher's outsider perspective.( 2) Triangulation: By examining student-produced artifacts (research products), researcher fieldnotes developed through participant observation, and student accounts of their experience drawn from openended survey questions and interviews, evidence of how students were engaged productively was corroborated across multiple sources and methods.
(3) Prolonged engagement: By engaging persistently with the course over six annual design cycles, the research team was able to gradually identify and refine the focus of the study, develop a number of appropriate analytical approaches, and ensure that the study was relevant to participants (students and designers) and that findings would likely serve their interests.
At the same time, this study has important limitations, notably in its partial reliance on self-reported and narrative accounts as data.While such accounts can surface things that pattern and organize learning and research practice that may otherwise be "invisible" to researchers (Czarniawska, 1998, p. 28), stories of practice are not the same as practice itself, and claims about what is experientially true and meaningful for participants should be not be treated as truths about an objective, universally shared reality when interpreting the findings of this study.

Results
Here we trace the boundary object of the scientific paper over the course of the semester, examining how people "act toward and with" it (Star, 2010, p. 603) at four different moments in the course, and looking at how students are engaged in different ways at different times for learning and research.
The narrative of classroom activity, unless otherwise noted, is drawn from fieldnotes taken by the first author over the course of participant observation.

Moment 1: A compelling object that engages researchers across disciplines
The CtQ course began as an effort by Julia, the instructor, to engage more students in the kind of research she undertakes in her own lab, a lab which publishes in high impact scientific journals with broad and diverse readership, and which is designed to engage students with different sociocultural backgrounds, disciplinary affiliations and experience levels in peer-supported scientific inquiry driven by personal curiosity.But courses in Julia's home academic program are not necessarily diverse by default; she needed to make a diverse course.In order to create a research experience characterized by disciplinary diversity, Julia advertises the course broadly in the large R1 public university in which she works and teaches, sharing information with other faculty, disseminating it via mailing lists in other departments, and visiting student groups that assemble individuals with the diversity of sociocultural and disciplinary backgrounds she seeks.Julia also ensures that the course carries a curricular "writing flag," a formal and highly visible designation in the university's student-facing course catalog that indicates that completing the course fulfills a writing requirement for undergraduate study.The "flag" in this published index of courses signals to undergraduate students who are outside of Julia's home department that their participation will clearly support writing requirements for their own academic advancement, regardless of their discipline.Julia also recruits students from her own lab and the labs of her colleagues, mostly in the various subdisciplines of geological and biological sciences.
The course assembles a relative diversity of students on the first day of class (Figure 2).Most students report being drawn to the opportunity to learn to write well (a learning goal) or to working on a specific manuscript that they feel the need to produce, though many also report that they have never written a scientific paper before, that they have not done research, and that they have not worked in such a disciplinarily diverse context (personal reflections, fieldnotes).The course requires students to write, code in the R programming language, and do statistical and quantitative analyses, all potentially challenging new activities for many of the students.In the face of these challenges-and other options for meeting writing goals and requirements-Julia must deal with the fact that students may not decide to stay in the course.She must link the course to a diversity of student interests and agendas.She does this on the first day by allowing them to learn about each other, taking initial steps to build familiarity and community.Julia gives a brief history of the course and relates it to her own history of becoming a scientist who is "driven more by questions and curiosity than by disciplines."She distinguishes the kind of science she is interested in, "curiosity-driven" science, from "discipline-driven" science.She also reveals her own history and work in broader scientific contexts, positioning herself as a reliable authority on the construction of scientific papers, noting her success in publishing papers in "high impact" journals that publish work from a variety of disciplines, and describing her history and her role as an editor in these journals.She talks about the importance of the common quantitative toolkit, R, for instance, but most salient to our narrative is how she emphasizes the importance of learning to write a scientific paper, highlighting this recognizable and meaningful object as the focus and goal of the course, as well as the thing which, in her words, "drives the research process."

Student perspective
Masud is an international student who is pursuing a master's degree in urban planning (interview, initial reflection).He writes in his initial reflection exercise that his long-term goal is joining an academic faculty and teaching and doing research.Masud's immediate goal, however, is to do thesis research, and this course appealed to him as a way "to learn the nuances of writing a scientific paper" (initial reflection).He writes that "good research work often doesn't turn into a good scientific paper.I want to overcome this issue.I also want to develop my research idea through systematic process."In his reflection at the end of the course, Masud revealed having some initial reservations: "The diversity was a lot" (interview)."When I started this course I was very unsure about continuing the course after seeing that almost all the other students are from other backgrounds than mine" (final reflection).
Overall, student interviews and reflections emphasized the compelling nature of the scientific paper as an object of activity.Analysis of student reflections revealed that most students cited the opportunity of learning to write scientific papers as a main reason for taking the course, stating this sometimes as a developmental need to "improve writing skill" or "learn proper research paper etiquette and formatting," and at other times in terms of scientific productivity: "I would love to . . .publish my research" (student reflections).The developmental and instrumental importance of this goal of writing was reiterated by students in the self-assessment survey at the end of the course, wherein they identified particularly high learning gains related to scientific writing skills and reported high levels of satisfaction related to progress they made writing about their research.

Analytical synthesis
At this point, the scientific paper is something of a nebulous boundary object, a broadly appealing, motivating and compelling object that is meaningful to students from a diversity of pathways and disciplines due to its general visibility, urgency and utility in a variety in their disciplined academicdomains.
From a developmental perspective, which emphasizes cyclical, humancentered learning processes coordinated with and among diverse and interconnected activity systems, the object appears to attract and compel broad engagement, operating as what Engeström (2007) calls a "runaway object."It is not (yet) the uniformly recognized and understood focus of a wellorganized activity system.Rather, it invites "swarming" among individuals from a broader network of disciplined activity systems and communities.It is urgent and useful, means different things to different students in different academic programs and stages of their careers, and colocates them in Julia's classroom as a potential site for the development of collaborative work on their own contextual instantiations of this runaway object.From a developmental perspective, the boundary object of the scientific paperat this moment acting as a runaway object-engages students productively by compelling them to "swarm" from a diversity of disciplinary activity systems to Julia's classroom to consider joint work on paper production.
From an instrumental perspective on the translational production of research networks, the scientific paper appears to mediate an initial moment of problematization (Callon, 1984).In this moment, students from diverse backgrounds and disciplinary interests are presented with a particular formulation of a common issue: how to become a productive studentresearcher.Julia's course frames the research process as an interest-and curiosity-oriented activity that is driven by the writing process, illuminates the prior success she has had in publishing scientific papers in broadly recognizable journals, and makes the argument to the newly assembled class that producing a scientific paper over the course of the semester in her class will help them advance their academic and professional interests in becoming researchers.From Masud's perspective, we see that while he is amenable to Julia's initial problematization, he is also disconcerted by the disciplinary heterogeneity of the course.Critically, through this process, students like Masud who are facing problems of thesis-writing and academic advancement, are presented with the option of ceding at least local, temporary control over the way they pursue and carry out their research goals, positioning Julia's course as what Callon (1984) calls an "obligatory passage point" with power over their progress: she will be their instructor and will grade their performance in the class.
We also note that Julia offers a process, timeline and toolkit to support paper production, but does not strongly dictate what students will focus on in their research, what questions they will ask, how they will assemble their data, who they will work with to analyze data, or what they will do with their papers outside the context of the course.At this early phase, the scientific paper operates as a boundary object by providing a broadly compelling rationale for assembling and undertaking collective work and points toward the possibility that that the disciplinarily diverse classroom can be stabilized as a working research network and course infrastructure.
Overall, how does disciplinary diversity factor into productive engagement at this moment?From both a developmental and instrumental perspective, productive engagement at this moment looks like a physical and virtual movement of individuals from the divided disciplinary spaces and infrastructures that organize research activity within the university to a shared context that is disciplinarily diverse: Julia's classroom.Here, from a developmental perspective, this movement is productive in the sense that it positions students around a potentially shared object of joint developmental work: they are considering entering into cyclical learning and mentorship interactions that will allow them to get better at writing papers and make progress on their academic degrees and research agendas.From an instrumental perspective, the students are engaged productively in the sense that they have assembled in a common problem space framed by Julia, wherein they may consider the benefits of forming an association that will help them meet their own academic and research goals.The design decision to recruit across a diversity of disciplines at this point contributes to a productive first meeting by drawing together common interests, but also presents challenges and tensions that impede progress: some students are suspicious that they don't belong in the course, that they won't be able to collaborate with others, or that others won't be able to support their discipline-specific needs.Julia deals with these tensions by building familiarity and community, emphasizing her own authority, emphasizing the common utility of the scientific paper in diverse research contexts, and downplaying differences in genre.
From an instrumental perspective, the association is not yet stable.From a developmental perspective, the motivating object that drives and brings sense to joint work is not yet clearly visible.The disciplinarily diverse network is at risk of fracture and disintegration due to lack of familiarity among participants in the class, and concerns about the feasibility and value of interactions between disciplinary actors: "the diversity [is] a lot" (Masud).

Moment 2: A concrete object that contextualizes research across disciplines
Per the course schedule, the students and instructors begin to convene twice a week.In these early meetings, the object of the scientific paper is immediately centered through a set of activities that are designed by the instructor to help students begin to understand what a scientific paper consists of, why they are constructed the way they are, and how they work as part of the scientific research processes.Notably, Julia does not simply tell the students how research works or what a scientific paper is, but rather invites them to examine what papers look like in their own contexts and bring artifacts to class which allow them to compare what counts as research in different domains.Julia asks students to bring a scientific paper that they like or use to class so that they can examine them together and figure out what they are and what they do.The students are instructed to bring empirical work as opposed to a conceptual or review-style paper, as they will be doing empirical work in class and will write empirical papers.By the third day of class, each student has submitted an example paper for consideration.These papers almost always represent work in their discipline and are often directly related to specific research interests and questions they bring to class.The students tend to select papers that they know from prior coursework, that are related to their new and often idiosyncratic science interests, or that are a part of the literature behind thesis work or research they are already undertaking with a lab or collaborator outside of the class.When the students reassemble in person in the classroom, they bring paper copies of these papers with them.

Student perspective
Masud is interested in researching how neighborhoods cluster in terms of their characteristics, such as municipal services, land uses, and educational facilities.He wants to potentially tie this work into broader thesis research that he hopes will help find him a permanent faculty position.
Sean, meanwhile, is interested in making progress on his undergraduate thesis, examining how obsidian pyroclasts can be used to understand volcanic degassing processes.He is carrying out this work in the context of a wellestablished volcanology lab led by an experienced member of the geosciences faculty external to the course.He hopes that this research will help him on his pathway to graduate school to study volcanology.

Analytical synthesis
At this moment, the nebulous, compelling boundary object of the scientific paper is engaged by the class in a different way, that is, as a concrete, material object made out of actual paper or composed as a digital file.Students bring their own instantiations of the paper, their own highly contextualized embodiments of the boundary object.Here, the scientific paper is a very concrete object.
From a developmental perspective, the urgent runaway object of the scientific paper that initially motivated swarming from across disciplines to the course begins to resolve as a constellation of concrete, material objects that are specifically situated in students' knowledge building and research practices.The polycontextual nature of the runaway object is revealed through this activity, setting the stage for a potentially productive classroom dialogue about what it means to write a good paper, and indeed a recognition that there may be multiple ways.The scientific papers, drawn from diverse contexts of practice and activity systems (e.g.labs, research groups), enter the emerging activity system of the classroom as tools for articulating the salience and role of the paper both in particular practice contexts, and more generally across them.
From an instrumental perspective, the paper engages students in a moment of interessment; it entangles the diverse interests of students as stakeholders in the course by showing how the writing and research to be undertaken in the class relates to their very specific academic and research writing, agendas, labs, infrastructures, theses.By inviting students to bring papers that are meaningful to them and matter to their practice, the course engages students in specifying links between the developmental problem at hand in the classroom as formulated by Julia-learning to write a scientific paper-and the instrumental problem of positioning and advancing themselves as researchers in the specific contexts, activity systems, and broader networks that matter to them.
At this moment, disciplinary diversity factors into productive engagement in a number of different ways.Productive engagement for individual students appears to involve them linking generally compelling concerns of the disciplinarily diverse course with very specific, disciplinary instantiations of them.From a developmental perspective, students are engaged productively as they articulate a nebulous but compelling boundary object that is valued across disciplines in terms of a specific disciplinary paper-a concrete cultural-historical artifact that is immediately salient to their disciplinary goals and work as researchers and students.From an instrumental perspective, students are engaged productively in the sense that they link the boundary object of the scientific paper around which the course has assembled with concrete research that is of interest to them and holds sway and influence in their broader disciplinary networks beyond the course.By assembling research papers, students have begun to knot together actants as potentially relevant to the problem at hand in the course, and linking external research, labs and researchers as potential stakeholders to their own work in the course.The students and Julia are able to avoid issues of noncoherence among example papers by deferring issues of difference to a later point in the course sequence, syncretically separating the work of interessment from the hard work of negotiating shared standards for papers and student roles in enforcing them, that is, the work of enrollment.

Moment 3: A normative object that standardizes research across disciplines
Having assembled a diverse group of students and papers, the course then engages students in a series of activities that are designed to force individuals in the classroom to identify what counts as a "good" scientific paper in their individual fields, and to build consensus understanding of these qualities that will help guide collective paper production in the course.Students begin to "act toward and with" (Star, 2010, p. 603) the boundary object of the scientific paper in yet another way: as a normative object.This instantiation of the course's boundary object is constructed through a series of classroombased activities that involve "deconstruction" and iterative "live blogging" of the structure of a scientific paper (Digital Supplement).
Asking the students to bring their papers to class, and displaying their individual papers on a screen where everyone can see them, Julia advances students' attention step by step through different parts of their scientific papers, from beginning to end, asking them to view their documents as specimens, as if they are "aliens" who don't know what scientific papers are or how they function.She asks questions like: Who has an abstract in their paper?What do you notice about the abstract?Why do you think it is written the way it is?What does the abstract do?When do you think would be the best time to write an abstract, as a researcher?Through questions like these, students discuss how scientific papers function in research processes (like literature searches), who they are intended to engage (e.g. for publication in broad audience "high impact" journals vs narrower technical journals), how they are evaluated and published through different peer review and editorial workflows, and how they support the knowledge work of other scientists.
Where there is confusion or disagreement about what should be in a general model of a good scientific paper, Julia steps in to deal with it in different ways.Sometimes she provides a functional description of why papers are the way they are, that is, how they function in the day-to-day work of research, literature review, editing and publishing (e.g.titles must convey key information in searches, and positioning figures and tables at the end makes it easier to trace the narrative and conforms to preferred editorial and publishing workflows).Sometimes she offers ethical rationales, arguing for writing in certain ways on the grounds of equity: certain colors are more easily discerned and differentiated by readers who are color-blind.Sometimes Julia simply lays claims of personal preference: "I like to write the abstract first" and revise it over the course of the research.
Once the "live blog of the structure of a scientific paper" is complete, it is edited and simplified by Julia, and shared as the "rubric" that she will use to perform her role as an evaluator and grader of their work, and that the students will use to perform their classroom roles as peer reviewers of each other's papers and as near-peer research mentors.

Student perspectives
Sean, working in his volcanology research lab, thought that "it was really useful to have someone articulate what goes into a good paper.I had never ever heard of what an impact factor was before the class" (interview).Sean appreciated "learning about the different categories or levels of what journal you want to submit to, how you want to make it sound, what is worth articulating or what's important to articulate given a certain journal.I wasn't really aware of that going into the course [. ..]I think if I had to say the most useful thing that I got out of the class was again, just the explicit articulation of what goes into a good section" of a paper (interview).
Masud, initially somewhat vexed by the diversity of perspectives in the class, reported finding that paper-writing "is not actually very disciplinespecific" and that reading across disciplines helped him begin to "know how to conduct any kind of research" (interview).

Analytical synthesis
Having traced the boundary object of the scientific paper from Moment 1 when it engages students as a compelling and broadly attractive, if nebulous, runaway object, to Moment 2 where it engages students in articulating concrete objects produced through different practice-relevant disciplinary contexts and activity systems, we can see in this moment how students are engaged in acting toward and with the boundary object as a tool of standardization that they will use to norm and guide productive research and writing activity in their shared classroom context.Here, the assemblage of individual concrete scientific papers mediates the co-construction of the boundary object as a normative object that will, in turn, mediate developmental activity and clarify instrumental roles among students and instructors.
From a developmental perspective, students are using the boundary object to develop themselves and the tools that they will use to mediate and structure their learning and work in the activity system of the classroom.Student engagements with the boundary object of the paper gradually develop a consensus set of rules-the "live blog of a scientific paper"which is in turn positioned as tool for guiding the application of those rules (the "rubric") through basic roles (writer, peer-reviewer and research mentor) that the classroom community will use to co-construct new research papers.In developing these rules, tools and divisions of labor, students also develop themselves, learning why papers are written some ways and not others based on how they operate in other disciplinary contexts.They learn to perceive common attributes that papers share across disciplines, attributes that make them productive in other activity systems and networks.The diversity among disciplinary papers also grounds student's understanding of such scientific norms as situated, practice-linked, and potentially changeable or dispensable depending upon the research context and community.
From an instrumental perspective, this live blogging activity can be read as a moment of enrollment, whereby the scientific paper functions as a site of negotiation about what counts as good quality research in the classroom, and helps articulate the roles and norms through which students interact with each other and each other's work through research mentorship and peer review later in the course.By live-blogging observations and connections among these discipline-specific examples, students produce a normative account of the scientific paper as a technical genre, which is then adopted as a community-generated template and rubric to guide student writing and research over the course of the semester.The specific normative role that students must take on as enrolled actors in a stable classroom research network becomes much clearer.They must examine the disciplinary work of others-even if outside of their discipline-and ensure that it conforms to the consensus "template." In this moment in the course, student engagements around the boundary object are productive from a developmental perspective to the extent that the students are able to work cyclically, iteratively and collectively to construct a normative object-a rubric-as a tool to guide future joint activity among individuals from different disciplines.Engagements are also productive in the sense that students develop themselves, for instance, deepening their understanding of scientific papers and practices more generally as situated, socially constructed, and changeable.From an instrumental perspective, disciplinary diversity factors into productive engagement in the sense that students are exposed to a variety of perspectives on what counts as a "good" and "quality" scientific paper, and they are able to use this diversity to negotiate a set of rules that will satisfy the needs of a disciplinarily diverse set of writers.Through this process of enrollment, students are put into productive relations that straddle disciplines (e.g. as peer mentors and peer reviewers) within the context of the course.Tensions and issues of noncoherence among disciplines that were deferred in Moment 2 are syncretically negotiated, settled, disguised or otherwise dealt with in Moment 3 in a variety of ways.Julia, for instance, sometimes treats writing norms as contingent and changeable, sometimes she weighs in as a dominant authority and dismisses alternative views, and other times qualitative differences between what count as good and bad papers are reduced or domesticated (Table 1) to a relatively trivial issue of a finite number of points on a grading rubric.

Moment 4: A novel object that expands new research across disciplines
Moving ahead again in the course, we find the boundary object of the scientific paper enacted and deployed in another important way: as new papers written by students.Here the boundary object emerges in the form of newly produced scientific papers salient to both classroom activity as well as research contexts outside of class.These papers are novel objects.Through a series of designed peer review and research mentorship interactions (Digital Supplement), students begin to assemble their own components of a scientific paper, piece by piece.Feedback is delivered in a variety of contexts, including frequent, ungraded, "low stakes" peer review group sessions in class where students trade sections of papers for comment.Students get feedback on their drafts from Julia and the postdoc in the class.They share and respond to iterations of figures in both whole class and in small group settings.They produce and talk about schematic representations of their research via a "draw your research" activity.They verbally present their paper in an exercise called "what is a minute?" that challenges them to articulate the most important and generally salient aspects of their work in under 60 seconds.In this exercise, students stand face-to-face with a peer in class, and are timed as they attempt to deliver a one-minute message about what they are doing as researchers and why it is important.They visibly and audibly struggle to refine their performance over the course of several iterative rounds (fieldnotes).They also communicate their work to their disciplinarily diverse peers in a five-minute lightening presentation on the final day of class, using five accompanying slides.
All students in the class use the rubric, and they structure feedback sessions around three basic questions: What did you like?What did you want to learn more about?What could be improved?Per the rubric, student attention and discourse in interaction is focused on questions like: Is the hypothesis findable and testable?Are references, tables and figures formatted correctly per the rubric?Additionally, graduate students are given readings on good research mentorship practices, meeting as a small group to discuss what they learned and to share their own history and perspective on mentorship.They then meet outside of class with undergraduates in a mentorship role, often providing research guidance during these meetings, and sometimes connecting them with the technical expertise and interests of others outside of the class who they know.Through these activities, students provide information and insights to each other to support data collection and analysis, they communicate their findings to students who may not have the same grasp of their disciplinary interests or terminology, and they assemble scientific papers that conform to the classroom consensus of quality.

Student perspectives
For Masud, what initially appeared to him as an intimidating diversity of disciplinary research in the class was largely resolved in a helpful way as he engaged in the rubric-based reading, peer review and feedback activities.
I found that the templates work really well.[. . .Initially] I really didn't think that there's literally that kind of template! . . .And then I started reading different papers by thinking about that template in my head, and I found that, yeah, I can match [the structure to] the slides.Yeah.Here should be the hypothesis.Yeah.I found that!And there should be the literature review . . .Yeah.They're doing that!(interview) Not only did Masud find that reading and responding to work across disciplines was helpful for learning to understand and evaluate the work of others, it also helped him see commonalities and areas of overlap between in his research and the research of others.
I found other guys [in class] who were using GIS and a similar kind of analysis.I found other guys who were using principal component analysis.I'm also using that kind of analysis, but yeah, we're from different backgrounds.(interview) He reported that such cross disciplinary windows sometimes edified his own scholarship by, for instance, providing external perspective and support for important assumptions held in his field of urban planning: Kevin worked with energy systems, like how renewable energy is going to replace traditional energy.I had a serious interest in that, because the planners are moving with the slogan [electrify the city].So yeah, I really had a close eye on his work.[. ..]I was really interested to know is it really happening?Is it really that renewable sources are actually replacing traditional sources?. (interview) While Masud reported that "At first, I was considering leaving the course if I didn't find it suitable."Later, he "found that research has a lot similarities across the boundaries, like across the boundaries of different backgrounds or disciplines.And, um, in the end, the course served my purpose," and he was able to make progress on work related to his thesis.
Meanwhile, Sean-who reported learning a great deal about the structure of a scientific paper through the class-also found that the approach to writing and research taken in class sometimes stood in tension with writing and research norms in his broader research network: It was always a toggle or a struggle between, okay, I have to perform this [analysis] and I have to do this in this exact order for this class, and there's no other way about it.I have to do it this way.This is what's due.Versus in real life [i.e.carrying out undergraduate thesis work in an external volcanology lab], I wouldn't have written anything until I had gotten all the data.I wouldn't have even started my methods, my intro, my abstract until I had finished data collection.[So, for my advisor], trying to discuss some things with him, he was kind of confused and he was kind of like, well, maybe I would've done this backwards.Some people write papers in real life, they get the data, [they then ask] what does it mean?Try to figure that out, try to write that up somehow and then work backwards [and] the last thing I would write in real life would be my abstract, summing up the whole paper after it's totally done.(interview) Sean found that the strongly hypothesis-driven type of inquiry, the quick pace and highly structured abstract-first approach to writing-as-iterativethinking espoused by Julia in the class made it difficult for him to carry out what he felt was a slower, more exploratory kind of thesis-oriented research his advisor was expecting from him in his volcanology lab.In the end, instead of resolving this difference by convincing his advisor that his thesis should be hypothesis-driven, or convincing Julia and the rest of the class that his science must be slower, more exploratory and descriptive, Sean simply wrote two different papers, one for the class, and later, a different one for his advisor.

Analytical synthesis
In this moment, the boundary object of the scientific paper emerges as a constellation of wholly new papers, engaging students across disciplines in transforming themselves, their research activities and contributing to broader networks.These scientific papers are produced by students as novel objects that are meaningful in the disciplinarily diverse class as well as the discipline-specific contexts in which students operate outside of class.
From a developmental perspective, the boundary object of the paper has become an object that orients classroom work iteratively over time.The students operate as a collective of subjects in the context of a classroom activity system, complete with rules (consensus paper structure), tools (rubrics and "templates"), a community (the people in the class), and specific roles and divisions of labor (writers, mentors, peer reviewers, instructors) mediating cyclical learning interactions.Within this activity system, students cyclically develop themselves through the production of a scientific paper.The boundary object of the paper mediates interactions in iterative peer review, helping students learn by resolving contradictions between their iterative drafts and the normative expectations co-constructed by the classroom.Furthermore, the object of developmental work in the classroom activity system is often linked to developmental activity systems outside of class, namely via thesis writing relationships with other faculty and external research labs.Students develop themselves as part of these broader networks of research by resolving and acting upon tensions among them.From a developmental perspective, students are learning by changing themselves, their peers and their activity systems inside and outside of class through the writing of scientific papers.
From an instrumental perspective, this moment can be seen as productively engaging students in the mobilization of an actor-network.Having problematized an issue of research productivity as an issue of learning to produce a scientific paper over the course of the semester, and interessed students by reinforcing the salience of the course to their particular disciplinary and academic interests, enrolled students by establishing consensus norms by which collective activity will be undertaken, we now see the students, their papers, and even their research relations external to the course assembled as a working association in service of the diverse-but-related problems and interests which drew the students together in the first place.
Critically, many of the perceived differences in what counts as a good scientific paper have been settled in the context of the classroom through this translational process, and a "consensus" structure of a paper has been inscribed as a rubric and circulated in the class to pattern the peer review work and mentorship that is understood as productive.From an instrumental perspective, students have been situated within, taken on a role within, and deployed productively as part of broader research network, generating papers that are instrumentally valuable in their laboratories and advisoradvisee relationships external to the course.In this way, they have been mobilized productively as a research network in a much broader research network.
How does disciplinary diversity factor into productive engagement at this moment?Engagements are productive from a developmental perspective to the extent that the students collectively cross a zone or proximal development, using a locally constructed normative object to mediate their production of a set of novel papers (motivating objects) that are meaningful and valuable in multiple disciplinary contexts, both in class and outside of class.From an instrumental perspective, students have been productively engaged around the boundary object in the sense that it has led to their mobilization as a actor-network that can competently assemble new scientific papers, thereby expanding the overall research network and stabilizing their position in it.The disciplinary diversity of the course enhances the complexity, size and diversity of this actor-network.This complexity inevitably invites noncoherence, but rather than avoiding disciplinary diversity for the sake of purity, noncoherence is dealt with through syncretic modes like separation (Table 1), as seen in the way that Sean separates the paper he produced for the course from the one he produced for his advisor in his lab.

Research productivity
Overall, survey, reflection and interview data suggest that students in CtQ by and large took up work that contributed to ongoing, broader and longerterm investigative agendas, and that these agendas spanned disciplines.Artifact analysis revealed that the research papers produced by students reflected work in a variety of very disciplines, including math (data science and statistics), natural sciences (biology, environmental science and geoscientific subdisciplines including geophysics, geochemistry, paleontology, physical geology, climate science), and the social and helping sciences (urban planning, public health, kinesiology, education research).In the 2020 and 2021 cohorts, 12 students identified a total of 32 different people who they felt would "learn from, use, integrate, or build upon" the research activity they carried out in the class.Artifact analysis revealed that 38.9% (37 of 95) of the students who completed a paper listed a coauthor.Of these students, 62.1% (23) were graduate students and 37.9% (14) were undergraduates.Students cited a total of 95 coauthorial relations overall, 11% (10) of which were with other students or the instructors in the course, and 89% (85) were people outside of the course.According to the post-course URSSA survey, nearly a third (31.8%) of responding students reported interacting with scientists outside of their school at least a fair amount.
Beyond these explicit co-authorial indications of the relevance of student research to the research of others, artifact analysis revealed that many soloauthored student papers were clearly linked to broader research agendas unfolding outside of the class.Interviews confirmed that graduate and undergraduate students alike frequently worked with faculty and labs external to the course to assemble data, refine useful questions and develop their papers for different purposes outside of the context of the class.These relations are reflected in networks of distributed research and writing support that students reported via the network survey (Figure 3), networks which extend beyond the boundaries of the course in every cohort.In the case of graduate students, interviews suggest that papers were often written to advance thesis research and manuscript writing at various stages, ranging from early proposal development to preparation for peer review.Undergraduate students developed papers related to undergraduate research organized through honors/scholars programs or their departments based on the post-course URSSA survey, we also know that some students expected to present their work at scholarly conferences.

Discussion
Returning now to our motivating research questions, what can we say about how the course leverages disciplinary diversity to promote productive learning and research engagement?By tracing the boundary object across four "moments," we can see how it is initially engaged (1) as a broadly compelling and quintessentially recognizable object, then (2) as a concrete object with specific utility in practice, then (3) as a normative object that facilitates fluencies and work across social worlds and disciplinary boundaries, and finally (4) as a novel object-a new scientific paper-which in its production transforms and expands scientists, scientific practice and scientific networks.

Boundary objects mediate developmental and instrumental productivity
From a human-centered and developmental (CHAT-based) perspective, we see students acting toward and with the object to identify and resolve contradictions in classroom activity through cyclical, iterative, human-developmental activities like peer review.Initially, students in the course assemble an urgent, if nebulous and general idea of what a scientific paper is-a "runaway object" around which they "swarm" from diverse disciplines (Engeström, 2007).Then, as students engage with concrete examples of scientific papers from various disciplines, they begin to analyze and compare their contextual instantiations of the polycontextual boundary object and gain a more complex understanding of scientific papers as situated cultural artifacts salient to diverse disciplinary research practices.With an enriched understanding, students then collectively develop a rubric for what constitutes a "good" scientific paper in the context of their interdisciplinary course-a normative set of rules and a concrete tool that mediates and guides their joint activity, their own learning, and the production of their novel research papers.
Disciplinary diversity as a designed element of the course is critical in the way that it invites a diversity of contextualized practices, discourses, concrete tools and artifacts into the learning process and engages learners in identifying and resolving contradictions among them, and transforming conceptualizations of science in support of day-to-day scientific practice.In this way, disciplinary diversity in an I-CBRE like CtQ fundamentally supports learning as a process of "ascending from the abstract to the concrete" (Engeström et al., 2012), whereby student-scientists take up and analyze concrete artifacts and practices in ways that allow them to progressively refine their general conceptualizations of scientific goals and practices, giving them shape, detail and practicality in activity.
However, from an instrumental perspective on the boundary object at these same four "moments," scientific productivity and what counts as "progress" in the course looks quite different.It involves enrolling various human and non-human actors (students, instructors, researchers external to the course, the scientific papers themselves) into a network that stabilizes the production of scientific papers and researchers.The course begins by problematizing an issue of scientific productivity as an issue of writing scientific papers, with the instructor positioning herself and the course as an obligatory passage point through which students must proceed to achieve their academic and research goals.Then, students bring concrete examples of scientific papers from their disciplines to the class, enabling them to discuss, compare, and contrast the different-but-related disciplinary understandings of science and scientific productivity, thereby entangling their specific interests in the context of the course.This process also allows them to negotiate the rules and standards that will organize their new association, producing the normalizing object of the paper-as-rubric.In accepting these norms and standards and enrolling themselves as peer-reviewers and paper writers in the course, they are ultimately mobilized as a scientific network, producing new scientific research papers, papers that are themselves linked to and constructed in alliance with a diversity of external labs and disciplinary contexts.Disciplinary diversity, from this instrumental perspective, demands that students "make progress" by negotiating their enrollment as productive actors in competent, but disciplinarily "impure" scientific research networks.

Attend to and design for productive syncretic engagement
Disciplinary diversity as a designed element of the course challenges students with issues of noncoherence and contradiction, and syncretic modes of reasoning and acting (e.g.Table 1) play a critical role as students navigate and reconcile the divergent norms and practices.We saw Sean operating in a mode that Law and colleagues call separation, a purposeful "holding apart" of apparently noncohering classroom-and lab-based views on what a scientific paper is and how it mediates research.Masud, meanwhile, made progress by citing Julia as an authoritative view that "writing is not very discipline-specific" (a view that is not in fact shared by Julia herself), a form of syncretic denial.A particularly valuable mode of syncretism in an I-CBRE may be what Law et al. call care, or an experimental, empathetic, trial-and-error tinkering in search of a fix that works.In the context of CtQ, such an iterative, recursive mode of syncretic tinkering was evident in the provisional construction of the working rubric, and might also be understood as process of design.By transforming varied normative conceptions of what constitutes a "good paper" into quantifiable metrics within a rubricbased grading system, the course participants effectively domesticated divergent conceptualizations of a quality paper.This standardization process facilitated both individual and collective advancement by taming differences within a structured evaluative framework.
We argue that student engagement in such "modes of syncretism" is important to their learning and productivity from both a developmental and instrumental perspective.Such modes help us understand and describe the common ways that contradictions in activity are "resolved" in complex activity systems and disciplined away in communities of practices.They show us what is involved in negotiating, disguising and settling issues of disciplinary difference and noncoherence in complex and heterodox scientific networks through translation.
Discipline-focused CBREs foster productive disciplinary engagement (PDE), which involves students tackling problems based on disciplinary logics, assuming authority, being held accountable to norms, and accessing necessary resources.We propose that the disciplinary diversity in an I-CBRE fosters productive syncretic engagement (PSE), whereby by students make progress amidst disciplinary impurity by exploring different disciplinary problem framings, negotiating different contextual roles and authorities, weighing different social and disciplinary accountabilities, and identifying and integrating resources and support across contexts (Table 3).Learners are encouraged to identify and integrate resources and support across contexts to do the work.
Design principles for PDE are adapted from Engle and Conant (2002).
Engel and Connant noted that focusing on PDE "allows one to trace the moment-by-moment development of new ideas and disciplinary understandings as they unfold in real-life settings" (p.403).We agree, and suggest that future work on interdisciplinary learning could focus on productive modes of syncretism as they unfold in disciplinarily diverse learning and research environments.Such work could help refine new principles of design for PSE.Designers and researchers may find purchase in studies of emerging forms of work in fundamentally heterodox and multi/inter/transdisciplinary environments outside of schools and universities, such as co-working spaces where many people can be found "working alone together" (Spinuzzi, 2012).

Conclusion
In examining the case of CtQ, we find that disciplinary diversity contributes to learning by putting students into interaction around boundary objects that promote productive engagement in very different ways at different moments, sometimes allowing them to progress by learning and changing themselves and each other over time, and at other times by forming more instrumental arrangements, epistemological alliances and scientific détentes.We think science learners make legitimate and important progress in both ways.Going forward, we suggest that those who study and design research experiences pay particular attention to how students can be engaged in the sometimes developmental, sometimes instrumental, and frequently syncretic work of negotiating issues of noncoherence and contradiction that inevitably arise in contemporary scientific networks characterized by disciplinary richness and "impurity."For example, how can student-scientists learn to recognize tensions among disciplines without immediately resolving them or explaining them away in terms of one discipline or another?How can they learn to "hold difference loosely" long enough to make progress in new, useful, good and undisciplined scientific directions?Such modes of syncretism are essential in the construction of scientific networks with the requisite complexity to address complex social and scientific challenges, and with the democratic pluralism to address them in an equitable and just fashion.

Figure 3 .
Figure 3. Writing and research support and mentorship relationships based on network survey.Network diagrams reveal that students report significant research and writing support relations with both peers in class, the instructor and postdoc in the course, and -for most-actors outside of class.Yellow = undergraduate student, Orange = graduate student, Black = instructor, Red = postdoc, White = outside of class

Table 2 .
Student population and sample.

Table 3 .
Design principles for productive disciplinary and syncretic engagement.