Coauthoring collaborative strategy when voices are many and authority is ambiguous

In interorganizational teams, processes are more complex and structures less clear than in intraorganizational settings. Different perspectives come together and authoritative positions are often ambiguous, which makes establishing what to do problematic. We adopt a ventriloquial analytical lens and pose the question: How exactly do interorganizational team members build a collaborative strategy under these conditions, in their situated interactions? Our findings show how many different voices (individual, organizational, team, and other) shape members’ strategy-making and reveal these voices’ performative authoritative effects: Members established their team’s strategy and produced the needed authority to do so through three coauthoring practices, namely, the proposition , appropriation , and expropriation of voices. When members switched between the practices and different voices, these voices were either woven together or moved apart. We sketch a conceptualization of strategy as a relational assemblage and develop a process model of strategy-coauthoring to illuminate these dynamics.

this article, we bring together these two complications and scrutinize their interplay on coauthoring collaborative strategy.
If successful, interorganizational teams provide a unique opportunity to leverage synergies. Members can exchange valuable expertise, access complimentary resources, and together tackle complex challenges (Majchrzak et al., 2015;Sydow and Braun, 2018). In practice, however, these collaborations often fail or perform below expectations (Yström and Agogué, 2020). The collaboration literature stresses the importance of common purposes and shared interpretations of activities and goals to prevent frustration, inefficiencies, and so on: collaborators need to come up with a collaborative strategy that clearly lays out what their group does and where they seek to go (e.g., Gray and Schruijer, 2010). However, when competing perspectives and obscure authority relations confound decision-making, defining this shared path can be very problematic (Huxham and Vangen, 2000;Koschmann and Burk, 2016).
In fact, we have little insight into the actual practice and concrete ways of accomplishing this integration, alignment, or unanimity. That is where we seek to contribute. Members need to build their collaborative strategy as part of their situated efforts, but how precisely they do so, how they manage their collaboration's ambiguous authority structures, and which exact role the different perspectives play (individual, organizational, team, etc.) are questions that yet remain to be more thoroughly problematized. Already in her early work on collaboration, Gray (1994: 290) writes, "whose voices are dominant and whose are silenced is a key question to explore." We interpret this statement as a prompt to investigate in more detail how the many perspectives of interorganizational collaboration become aligned into a collaborative strategy (or not) and pose the question: Whose voices do we hear in interorganizational teamwork and how do they take part in coauthoring strategy?
To realize this ambition, we place communication front and center and consider collaboration, strategy-making, and authority as ongoing communicative processes: what a collaboration does and who has a say on that emerges and continuously evolves in collaborators' situated conversations. We can find traces of such an understanding already in the early collaboration literature (Donnellon et al., 1986;Gray et al., 1985). Novel work on strategy and particularly Strategy-as-Practice also shows an increased interest in the performative effects of language-in-use, most recently evidenced in a Strategic Organization themed issue on communicative perspectives (Vaara and Langley, 2021). Vaara and Langley call for research that illuminates in detail how strategy-making happens through language, and they draw attention to novel theoretical and methodological approaches to communication, such as the Communication Constitutes Organization perspective (Cooren et al., 2011). This perspective argues that organizations and organizational phenomena come to exist in communicative practice and, therefore, take seriously communication's consequentiality for how strategy is constituted (Bencherki et al., 2021) and authority is produced (Benoit- Barné and Cooren, 2009).
Specifically, we use this perspective's concept of ventriloquism. Cooren (2010) introduced this metaphor to illustrate how interactants make present additional voices whenever they communicate-much like a ventriloquial artist expresses the voice of her puppet. For example, an interorganizational team member could speak in the name of her organization and make present their value statement to express disagreement with a strategy (e.g., "for us, this is not acceptable; what you are suggesting has no connection with our value proposition's sustainability standards"). From a ventriloquial viewpoint, not only the team member is doing the talking here; we also hear her organization's voice through what she says. Ventriloquism considers these voices as agentic participants in conversations and proposes that they make important differences for how situations unfold (Cooren, 2015), for how authority is enacted (Vasilyeva et al., 2020), and thus also for what can eventually become a group's strategy. A ventriloquial analysis (Nathues et al., 2021) then enables a novel and fine-grained take on the voices we hear in interorganizational collaboration and their implications for coauthoring collaborative strategy.
We analyzed 11 video-recorded meetings of an interorganizational team that deviated from its original plan and reauthored its objectives and actions. This gave us the unique opportunity to study how a collaborative strategy is built from scratch, in situated communication. This article proposes three coauthoring practices (proposing, appropriating, and expropriating voices) that unravel the dynamics of the collaborative strategy-coauthoring process and that illuminate the performative and authoritative effects different voices can have on strategy. We also sketch a conceptualization of collaborative strategy as a relational assemblage (Kuhn, 2021) of voices that integrate or separate, and argue that strategy materializes from the connections formed between these voices. This proposition invites us to think about strategy-making as a process of piecing together and moving apart, enacted in and through communication.
Our findings and ideas make important contributions to the literature on (interorganizational) collaboration and strategy. They show how the integration of different voices into a collaborative strategy is constituted in communication and accomplished through ongoing interaction. This is important because it accentuates how collaborators can enact strategic agency through subtle but powerful communicative practices. Our ventriloquial analysis unpacks these communicative practices in rich detail, captures their consequentiality for strategy-making, and demonstrates how seemingly small and conventionally overlooked communicative details matter for strategy.

Theoretical background
We start by revisiting three important elements of our study: (a) interorganizational teamwork and strategy-making, (b) multi-voicedness and ventriloquism, and (c) authority and authoring.

Interorganizational teamwork and strategy-making
Whether to combine resources, exchange knowledge, or co-develop solutions, organizations increasingly recognize that partnering with others is essential for innovation and continuation (Majchrzak et al., 2015). At the same time, interorganizational processes are more complex and structures less clear than in intraorganizational settings (Sydow and Braun, 2018). Organizational aims and interests frequently diverge, leading to conflict and frustration (Gray, 1985;Vangen, 2000, 2005). Collaboration furthermore requires that members share power (Gray, 1994) which means that they cannot rely on pre-existing hierarchical relations. Instead, coordination and control-for example, over which organizational aims will be prioritized-need to be exercised through interactional practices, such as managing the agenda or mobilizing organizational resources (Koschmann and Burk, 2016). To move forward their collective, interorganizational teams are therefore faced with the continuous challenge of channeling their plurality into a shared direction that all parties find acceptable (Gray and Schruijer, 2010). In other words, members need to construct a collaborative strategy: What is it that they collectively seek to do?
The Strategy-as-Practice literature unpacks how professionals formulate and implement strategies as part of their everyday activities (Johnson et al., 2007;Vaara and Whittington, 2012). This work demonstrates a growing interest in language's performative effects on strategy-making, most recently evidenced in a Strategic Organization themed issue on communicative perspectives (Vaara and Langley, 2021) but also apparent in earlier work. For instance, studies have identified the discursive practices that characterize strategizing (Vaara et al., 2004), have examined how texts contribute to strategy-making (Spee and Jarzabkowski, 2011), or have elucidated the situated role of ambiguity and multivocality in socially accomplishing strategy (Aggerholm et al., 2012). Samra-Fredericks (2003) provides one of the most detailed accounts in her conversation analysis of how managers build shared meanings of strategic matters. Her analysis reveals the consequential effects of seemingly trivial communicative moves, such as invoking a metaphor or using varying personal pronouns. Asmuß and Oshima (2018), Bencherki and colleagues (2021), and Cooren and colleagues (2015) likewise showcase communication's important role for strategizing. These authors draw from Communication Constitutes Organization thinking, which puts forward a dynamic conceptualization of organization as forming in everyday communication (Cooren et al., 2011;Schoeneborn et al., 2019). This perspective negates the distinction between organization, on the one hand, and communication, on the other, and argues that communication always defines-constitutes-what becomes organizational (Cooren, 2012), and therefore also what becomes strategic. Specifically, Asmuß and Oshima (2018) study how proposed strategy changes are dealt with and show how entitlement questions and different strategic roles emerge and are (re)negotiated in situated interactions. Cooren and colleagues (2015) unpack the complex web of matters and concerns that substantiate strategy and illustrate how this web forms and is (de)valued communicatively. Finally, Bencherki and colleagues (2021) focus on strategy's initial formulation and provide detailed insight into the communicative practices that concerns need to move through to eventually constitute strategy (i.e., presentifying, substantiating, attributing, and crystallizing).
These authors' "communication as constitutive" take on strategy-making appears valuable also for our ambition: when communication is constitutive of strategy, then paying detailed attention to conversations should help us understand how interorganizational team members decide what to work on. Moreover, the two phenomena that we believe complicate collaborative strategy-making in interorganizational teamwork (multi-voicedness and ambiguous authority relations) are communicative by their very nature (Aggerholm et al., 2012;Bourgoin et al., 2020), which renders such a take even more relevant.

Multi-voicedness and ventriloquism
The notion of multi-voicedness has increasingly appealed to researchers of organizational communication. In its most basic reading, it stresses how different organizational actors can have different viewpoints (Jørgensen et al., 2012) but also how one and the same person can express various voices (Trittin and Schoeneborn, 2017).
The metaphor of ventriloquism was introduced to spotlight the different voices that interactants can make present when they communicate (Cooren, 2010;Cooren and Sandler, 2014). Broadly speaking, just as a ventriloquist speaks and acts in the name of her puppet, organizational actors can speak and act in the name of their supervisors, their organizations, rules they must follow, and so on. Invoking these voices makes these other agents participate in the conversation: they actively contribute to what is discussed (Cooren, 2015). Ventriloquism hence broadens the interactional scene beyond the present humans and puts forward a decentered conceptualization of agency where action, including strategy-making, is distributed across a chain of invoked voices (Castor and Cooren, 2006). A ventriloquial analysis can therefore enable strategy researchers to pay close attention to the diverse voices that can be heard in interactions and to consider them as potentially influential for the direction a collaboration is heading toward.
One could argue that especially in interorganizational teams, multiple voices are at play as professionals with dissimilar views, organizational interests, and so on are assembled (Gray and Schruijer, 2010). Matters brought up can be of differing personal or organizational relevance, but collaborators also need to form a unified group and speak with a shared voice (Hardy et al., 2005;Sydow and Braun, 2018). Members could also make present the voices of external stakeholders and larger contexts (Phillips et al., 2000), or even of seemingly abstract sentiments and considerations such as ideals, principles, and feelings (Cooren, 2012). Lewis and colleagues describe members' focus on their organization as a "me-orientation" (2010: 468); Hardy and colleagues use the label of "assertive talk" (2005: 69). Both notions stress how collaborators prioritize organizational needs over team concerns. We would argue that collaborators' individual professions, backgrounds, and so on can also be part of this group, when personal interests are prioritized over team interests. Lewis and colleagues propose a "we-orientation" (2010: 468) to depict when collaborators speak as one team and prioritize their collective benefit; something that Hardy and colleagues call "cooperative talk" (2005: 69).
Overall, this multi-voicedness complicates the process of coauthoring a collaborative strategy: the different voices can jostle one another and push and pull collaborators in between them. Collaborators will want their voices to be heard to contribute to the collaborative strategy. However, some voices might be louder than others. Eventually, the different voices might therefore come to "engage in a contest for audibility and power" (Belova et al., 2008: 493). But how then does this contest play out?

Authority and (co)authoring
Compared with traditional organizations, interorganizational collaboration is characterized by more elusive authority relations (Koschmann and Burk, 2016). Authority is less anchored in centralized positions or structures (Barley, 1996) and needs to emerge from interaction (Bencherki et al., 2020) to provide for the minimum direction that is needed for collective action (Taylor and van Every, 2014).
In keeping with the Communication Constitutes Organization perspective, Porter and colleagues (2018) move from authority to studying the practice of authoring and emphasize authority's emergent, local, and communicative nature (see also Koschmann and Burk, 2016). Moreover, they frame authority as "which voices will be heard" (2018: 873), which is similar to Benoit-Barné and Cooren's (2009) idea of accomplishing authority by invoking others' agency through speech (see also Bourgoin et al., 2020). It also links back to Gray's (1994: 290) question about "whose voices are dominant and whose are silenced" and reminds us of Vaara's (2010: 44, italics added) call to study "who gets to participate in the authoring of strategies." We therefore adopt Porter and colleagues' (2018) conceptualization of authority for our inquiry and, given interorganizational collaboration's multi-voicedness, further specify it as the practice of coauthoring. We propose that coauthoring the strategy of an interorganizational collaboration requires articulating one's voice as speaking for the team amid manifold additional voices.
These additional voices can have important consequences for coauthoring attempts. Voices that say the same can be compelling allies (Benoit-Barné and Cooren, 2009) as they can enable a single person to speak with the power of a crowd (Bourdieu, 1991). For instance, members could strengthen their suggestion with their organization's reputation, but they might also invoke the industry's voice or mobilize their expertise to legitimate their proposal (Kwon et al., 2014). In these examples, the additional voices that members invoke back up their utterances and thereby help them coauthor their preferred course of action.
Voices can be also invoked to resist a strategy, such as when a group's voice is skillfully distanced from a proposal (Bourgoin et al., 2020) or when different organizational voices push for incompatible concerns. For instance, one member's attempt to change a group's strategy might be inhibited by disconnecting this member's voice from the collective's shared voice and thereby excluding this member from strategy-making (Tavella, 2021). Alternatively, an attempted strategy alteration might also be dismissed by invoking voices that help outvote it.
Interorganizational collaboration's multi-voicedness and its ambiguous authority relations thus complicate creating alignment and forming a shared interpretation of what to do. We seek to unpack this complexity, studying situated interactions: Whose voices do we hear in interorganizational teamwork and how do they take part in coauthoring collaborative strategy?

Case and data
As part of a larger project, we secured access to an interorganizational collaboration initiative aimed at strengthening regional business development. The initiative was supported by a regional investment fund and had a duration of 3.5 years. It brought together 23 organizations, consisting of high-tech companies and service providers that differed on characteristics as maturity, structures, and markets. Organizations formed teams to work on joint strategic challenges (e.g. servitization implementation, business model transformation). As perspective exchange was an essential goal, teams were always composed of members from different organizational and professional backgrounds. Teams had to report their progress to the steering group overseeing the initiative but, apart from that, were themselves responsible for determining what exactly they worked on or how they distributed responsibilities and executed activities. This confirmed that authority structures were elusive.
The first author joined the initiative as an academic researcher. She was not actively involved in any subject matter discussions but followed teams through observations, interviews, and database access. Her continuous access to multiple teams enabled her to build a profound understanding of the initiative, its motivations and structures, and its members (as both organizations and individuals). The second author joined the initiative as an academic consultant: she was part of the steering group and helped monitoring the project but never attended an actual team meeting. This positioned her as a distant insider. The third author joined the research after data collection was finalized. He never had any contact with the initiative and hence added a valuable outsider perspective.
The present study is built on the material we collected in one of the initiative's teams (Table 1). This team was supposed to work on the implementation of a digital product sheet that would follow products through their production cycles, logging and communicating information through automated procedures. However, members considered it more relevant to work on automation mechanisms beyond production processes as problems typically occurred not in manufacturing but in the procurement of product components. They hence reauthored their objectives and activities once they started collaborating, exploring the possible implementation of a decentralized hub that brought together suppliers and organizations to automate and accelerate procurement processes 1 . That gave us the unique opportunity to study how a new collaborative strategy was built from scratch. We followed the team from the first to their final meeting. The team had four members from three organizations (pseudonymized): Max worked as an engineering manager at Proto, a large organization operating in sensitive markets. He joined the team as his organization was searching for more efficient ways of organizing their prototyping. Tom worked as an IT architect for Flex. He described himself as a technology expert but as less experienced in production processes. He joined the team because of his technological interest and because his organization sought to test more flexible production and procurement handlings. Ben and Anna worked for Hali; Ben as a senior project manager and Anna as a temporary consultant. They joined the project because of their (organization's) technological interest. Moreover, their company aimed to build a better network with companies in the region. The members did not closely know each other before their joint work started. Max and Tom had met twice (in the context of the overarching initiative). As Anna had just joined the company, she and Ben had met only once.
The team came together for eleven meetings. As a non-participant observer, the first author attended nine meetings (Meetings 1-8 and 11). She was replaced by a trained assistant for the two meetings she could not attend (9 and 10). All meetings were video-recorded with a 360 degreecamera, to capture the fine-grained, interactional data that we needed to understand how members communicatively established their collaborative strategy. Additional materials were used to develop a comprehensive contextual understanding and corroborate emerging insights. These included fieldnotes, documents (e.g., meeting minutes, reports) and two interviews with each member, once in the collaboration's initial phase and once halfway through (Table 1).

Data analysis: a ventriloquial take on coauthoring strategy
We chose ventriloquism as our analytical lens. With its detailed attention to communication's constitutive effects (Cooren, 2012), voice invocation (Cooren and Sandler, 2014), and authority (Vasilyeva et al., 2020), ventriloquism appeared useful to help us understand how members built a collaborative strategy amid diverse perspectives and elusive authority structures. It also offered a novel and potentially enriching take on strategy-making that we sought to further explore. Following the ventriloquial analytical framework recently proposed by Nathues et al. (2021), we moved through four phases.
Phase 1: Identifying. The framework lists three analytical question to identify ventriloquial effects. The first question (what is a person invoking with what she is saying?) aids in identifying explicit voice invocations, such as when a professional openly speaks in the name of her organization (e.g., "for us, for our company, it is very important that supply chain processes run smoothly"). The second question (what voice(s) can be recognized in what a person is saying?) helps to unpack implicit voice invocations. Here, a voice is not straightforwardly named but present in a more enfolded manner, such as when a professional implicitly refers to a rule members must follow at her company (e.g., "we have to do it that way at my company, there's no way around that"). Identifying these implicit invocations requires intense familiarization with the material and a deep understanding of the setting; in our case enabled by the first author's prolonged field presence. The third question (what appears to lead a person to say what she is saying?) aids in identifying possible animation effects, that is, when voices can be identified as the driver or motivator of someone's utterances or actions. Typically, animation effects are characterized by attachments, such as when the attachment to a principle or value makes a person interrupt and disagree (e.g., "I'm sorry, this doesn't work for me, maximum data protection is really important to me").
Equipped with these analytical questions, the first author analyzed all video-recordings; combing the data for individual, organizational, collective, and other ventriloquized voices (also considering pronoun use, Taylor and van Every, 2014). Paying attention to implicit invocations, for example, helped identify the voice of caution that one member was implicitly invoking when suggesting an initial focus on investigation. The animation question was particularly helpful to grasp that the same member's behavior was driven by the strict security regulations that his organization had in place. All passages where members discussed strategic options were marked using computer-aided qualitative analysis software (Atlas.ti, for easy traceability) and transcribed (92 pages, single-lined). To avoid dependence on a single person's interpretations, all authors jointly analyzed portions of these transcripts. The first author also studied meeting minutes, the team's final report, her fieldnotes, and interview transcripts for better orientation and corroboration of emerging insights. This additional material was especially helpful for cross-checking our interpretations of implicit voices and animations. We also presented our final analysis to the team we followed, to review our interpretations and indicate possible inconsistencies. None were found.
Phases 2 and 3: Ordering and relating. Our interest in interorganizational collaboration's multi-voicedness already provided us with some structure (me-orientation and we-orientation; individual, organizational, team, etc.). However, ordering and relating the many invoked voices that we identified proved to be even more complex. For example, we found that members spoke with a singular first-person pronoun ("I") while referring to their organizations. Similarly, when members spoke with a plural first-person pronoun ("we"), this could refer to both their organizations and their team. We reviewed the literature for guidance on how to clarify this nested complexity and found Ashforth and colleagues' (2011) distinction between intrasubjective I, intersubjective we, and generic subjective it helpful. While their take on pronouns addresses the nestedness of identities in organizations, it helped us articulate the complexities of invoked voices in strategy-making, too. Figure 1 provides an overview of the voices we identified. Specifically, we found that members spoke as an intrasubjective I to refer to themselves as individuals or as organizational representatives and vice versa addressed each other as individuals or as organizational spokespersons with what we call an intersubjective you. We could also group two ways of speaking as an intersubjective we, where we referred to either one's organization or the team. Finally, we found instances where members referred to the team or organizations as it or where they made present additional voices, such as when they spoke from positions of caution or expertise or when they invoked the initiative's voice. We grouped these additional voices under the label generic subjective it.
To better understand how the collaborative strategy emerged and evolved, we created thick descriptions for each meeting, chronicling the main discussions threads. We particularly focused on strategic tensions, that is, moments where members had to decide between different options. We then grouped discussion threads, identified voices per meeting, and abstracted timelines of how threads, tensions, and voices fluctuated (Langley, 1999). For further zooming in, we selected two central threads of the team's strategy: (a) their general decision to work on a certain technology (which we call "hub") instead of the digital document they were supposed to develop and (b) their discussions on whether this hub should be an open or closed system. Within these two threads, we differentiated voice invocation between members (who invoked which voices) which showed how individuals were invoking different voices as they were engaging in different strategy-coauthoring practices. We identified three such practices (Table 2): (a) proposing individual, organizational, and other voices to coauthor the collaborative strategy, (b) appropriating the team voice to coauthor the collaborative strategy, and (c) expropriating the team voice to detach themselves from the collaborative strategy.
Phase 4: Showing. The presentation of our findings broadly follows a "conceptualized composition" format (Berends and Deken, 2021). Phases 2 and 3 have already introduced our insights on nested multi-voicedness and coauthoring practices. In Phase 4, we selected five meeting episodes to present as vignettes, to open up richer insight into the empirical dynamics. We use the discussion section to develop a process model of the coauthoring practices and their implications for strategy-making.

Empirical episodes
In this section, we zoom into five episodes to illustrate how interorganizational team members coauthored their collaborative strategy in their situated conversations, through ventriloquial acts. Table 3 provides an overview of the vignettes.  Reflections of the strategy are authored

Vignette 1: Authoring a new strategy
When the team starts collaborating, meetings are characterized by great voice diversity. Two members in particular ventriloquize many voices: Max and Tom are eager to change the team's strategy and mobilize different voices to do so. Instead of working on an advanced digital product sheet, they seek to work on a hub to solve their organizations' strategic challenges around procurement 2 . They have just suggested this new strategy. Vignette 1 is taken from the team's first meeting. Only Max, Tom, and Ben are present. Picking up on the new strategy that Max and Tom have suggested shortly before, Ben expresses his view of what the team ("we") will work on (line 1). However, he is interrupted by Max who counters Ben's suggestion with a more detailed and simultaneously less ambitious version: rather than "develop[ing] a [hub]" (line 1), the team should investigate "whether it is eh (.) feasible to develop" this hub (line 2). By correcting Ben, Max positions himself as being able to speak for the team and decide on its strategy: he knows better than Ben what the team should do. It also appears that Max is mobilizing the voice of caution here. His humbler suggestion of focusing on investigation first (not development right away) could be read as expressing his ability to lead by the use of reason, further substantiating that Max should be the one speaking and deciding for the team. Tom agrees with Max and elaborates on their rationale for dropping the original idea and instead working on a hub (lines 3-5). Tom appears to be able to speak with the team voice too, as if he and Max would make up the collaboration and Ben would not (yet) be part of it. In other words, Max and Vignette 1. Authoring a new strategy. Tom appropriate the team voice to outline a new strategy and, simultaneously, keep Ben from doing the same.
Max then provides another argument for changing the team's strategy and working on a hub by explicitly invoking his organization's voice and implicitly invoking his colleagues' voices when he refers to "we" and "[Proto]" and stresses how their product sheet "doesn't need any eh improvement" (lines 6-7). Max's exact phrasing here is a literal example of how voices are multiplied ("that's one and what we also see," line 6) to strengthen one's authoritative position. In what seems to be an attempt of adding further voices, Max then seeks a similar organizational affirmation from Tom; now addressing him as an organizational representative ("you have eh within your company," line 7). Tom once more backs up what Max says (line 9), and Max continues on the point he is trying to make: that product documentation during manufacturing and the product sheet should no longer be focused on. He again invokes his organization's voice ("we see actually that this is convenient enough," line 11) and also starts speaking as Tom's organization (lines 11-12). Tom then himself elaborates on his organization's strengths and structures, substantiating the fact that his organization does not need a new product sheet by invoking their fast turnarounds (lines 13-15). We also see him invoking the voice of Max's organization, referring to "over there" (line 16) while pointing at Max. Tom explicitly points out that the sheet is "not really a problem" (line 16) to which Max agrees (line 18). Multiple voices are hence proposed by Max and Tom as pressing for a new team strategy: not just their own (as the ones able to speak in the team's name), but also the voices of their organizations, their organizational colleagues, and their product sheets.
Max then translates the various concerns brought up into a succinct collaborative strategy. He switches to talking with the team voice and authors what they must do: "What we have to do now is investigate whether that is even possible" (lines 18-19). Note how Max uses the verb "investigate" to describe the team's core activity, as he did at the very beginning (line 2). Ben, who has been silent for a while, questions Max for more detail (lines 20, 24). While he initially referred to the team as "we" (line 1), Ben now addresses his collaborators as "you" (lines 20, 24). It appears as if he is granting the others (and their organizations) full authority to decide on what the team is going to work on: Max and Tom are able to speak as the team (appropriating its voice) and to coauthor a new strategy through invoking (proposing) multiple voices, while Ben expropriates the team voice and does not join the strategy-making. The new strategy hence emerges from an asymmetrical mobilization of voices. Because Max and Tom switch between the practices of proposing and appropriating, their organizations' voices and the team voice are being pieced together. In contrast, Ben moves from appropriating to expropriating the team voice, which, at this point, separates his voice from the team and its emerging strategy.

Vignette 2: Spelling out strategy details
Members have decided to drop the original plan. Consequently, they need to further specify their new strategy: what exactly will they focus on in working on this new subject? In Vignette 2 (from the second meeting), two aspects are discussed: prototypes and security requirements. Anna is present for the first time but mostly takes on a listening role. As she did not join the first meeting and hence did not coauthor the strategy that was discussed then, she might possibly not yet feel authorized to contribute.
The vignette begins with Ben-speaking with a collective "we" (line 1)-suggesting how to approach one aspect of the new strategy. He appears to be slightly unsure about speaking as the team (and adding to their strategy), indicated by his hesitation ("we eh but eh," line 1). This changes after Max and Tom agree with his proposition (lines 2-3); their affirmation seems to put Ben in a more comfortable position when he authors the collaboration's next steps in his next turn (lines 4-5). Note how Ben is picking up on the feasibility objective ("if it is really feasible," lines 4-5) that Max has emphasized in Vignette 1: Ben is essentially appropriating the team voice and contributing to its new strategy by appropriating what Max said as representing the team before.
We see Max attempting to add another aspect that is of relevance to his organization to the next steps Ben has outlined (lines 6-7). Interestingly, Max starts speaking as an individual ("I") but then switches to speaking as his organization's spokesperson ("we"), presumably to lend more weight to the addition he is trying to coauthor into the strategy: it is not just him but also his organization that needs this addition to be included. Max explicitly invokes his organization's "industry" and "firewall" (line 6) and then switches to speaking in the team's name again ("things we need to investigate," line 7) as if to ensure that his organization's concerns are adopted by the team. Subsequently, Max addresses Tom and his organization (specifically also their firewalls, line 7), presumably to add additional voices that push for including his security concern in the team's strategy. However, Tom weakens rather than supports Max's suggestion (line 8). Specifically, by stressing his close proximal distance "to the boss of the firewall" (line 8) and by his joking tone, Tom presents Max and Proto's security concerns as not being of much relevance to himself or his organization. We can observe an interesting dynamic of strategy-coauthoring here: voices proposed as important for the collaborative strategy by one member can be accepted as such by others, but they can likewise be questioned, contested, or made to say something else. In other words, once expressed, proposed voices escape the full control of their articulator-especially so when a voice that is proposed belongs to someone else that is also in the room.
Max resorts to adding additional organizational voices to emphasize their need for security. Specifically, he invokes their "grades of [. . .] Internet" (line 9) and their "physically separated" parts (line 10). These virtual and physical structures as well as the organizational voices Max made Vignette 2. Spelling out strategy details. That's also things we need to investigate. You also have your firewall =((looking at Tom)) 8 Tom: =Yeah, but I work next to the boss of the firewall so ((laughing)) 9 10 11 12 Max: It's a bit easier! But we have eh three different grades of eh-of eh Internet. And-and-theand the so-so some parts are actually physically separated so-(.) so we need to understand a bit also what are the IT requirements [. . .] defining a prototype is not that difficult. We are making many and also [Flex] you are making ma-eh only prototypes. =Yes but-but-but we, we have some additional eh requirements in these kinds of areas, so we also want eh we also want to know what kind of export licenses are behind those components. 22 Ben: So that comes in the requirements? 23 Max: Yes, that comes in the requirements.
present before (the industry and firewall) thereby join him in authoring details of what the team should work on (lines 10-11). Altogether, they essentially dictate or require that security concerns are coauthored into the team's strategy. We can observe how Max continues appropriating the team voice but also how he proposes many additional voices to coauthor the new strategy's details. These voices help him accomplish a strong authoritative position: because we also hear Max's organization speaking through him, the significance of his additions is amplified.
Only afterwards Max comes back to the prototypes (line 11). Now talking for both his and Tom's organizations (line 12), he emphasizes how defining prototypes is not difficult. This time Tom agrees (line 13). Anna then has a question on the prototypes (line 14), but Tom swiftly explains how a few numbers are sufficient for what the team intends to do (lines 15-18). Tom seems to be speaking from a position of experience and expertise, materializing through his effortless enumeration as well as his exact phrasing: he categorizes Anna's question as a risk the team cannot "dare" (line 15), thereby at least partially expressing that he knows better. This points at his continued authority over the collaborative strategy, even if Tom is not explicitly speaking in the team's name here.
Tom is interrupted by Max who voices additional requirements. Speaking once more for his organization ("we," line 19), Max names "export licenses" (line 20) as crucial information that needs to be included "in the requirements" (line 23). It could be said that Max's organization again joins him in authoring details of the new strategy, but we could also say that Max implicitly positions himself as animated by his organization and its interests, acting merely as a medium through which organizational concerns are brought forward. The interviews we conducted with Max's team colleagues corroborate this impression: team members repeatedly described Max's behavior as strongly driven by his organization and its strict procedures. Ben instantly accepts Max's addition (line 22). This once more shows how Max, just like Tom, still seems to possess much authoritative power over the team's voice and strategy, greatly influenced by the proposition of organizational voices.
Overall, the team's new strategy gains further detail in this second meeting. Especially Max's switches between proposing organizational voices and appropriating the team voice increasingly enmesh these different voices into the bundle constituting the team's strategy.

Vignette 3: Authoring opposing strategic options
With the joint work progressing, the team's strategy is further specified. However, Max and Tom cannot agree upon one aspect: whereas Max appears driven by his organization's strict security policies and presses for a closed system, Tom envisions a more open structure to meet his organization's needs for adaptiveness and flexibility. In Vignette 3 (from the fourth meeting), members discuss this tension.
Anna raises a question about whether the hub should be "anonymous" (line 1). She speaks with a generic vocabulary, talking about "suppliers" (line 2) and "company" (lines 2, 3) instead of naming any particular one or speaking with her organization's voice. This contrasts with many of the behaviors we have observed in previous meetings, especially those of Max and Tom who often spoke for their organizations and named their suppliers. Interestingly, Anna seems to be addressing her question about anonymity at Max and Tom (line 3) not as individual professionals but as organizational representatives ("as a company," line 2). This might be a consequence of the increasing entanglement of the team's strategy with Max and Tom's organizational voices.
In his response, Tom, however, starts speaking not as an organizational representative but as himself ("I," line 4) and then switches to using the team's "we" (line 4) as if to emphasize that keeping the hub public is important not just for his organization but for the entire team. He also starts speaking "as a supplier" (line 4), invoking this supplier's voice to make a better case for his argument (lines 4-5). It is not just him saying that the hub should be public but also one of this hub's main participants, which helps Tom emphasize the significance of his point. In other words, while appropriating the team voice Tom also proposes the voice of an imagined supplier to strengthen his "public" (line 4) version of the strategy. The fact that he mobilizes the team's and this supplier's voices, although being addressed as an organizational representative, points at a possible rearrangement of authority sources: at least for Tom, representing the team now seems to weigh more than representing his organization.
Max disagrees with Tom, stating that the hub should be "restricted public" (line 8). He explains his general attitude toward data-sharing, describing how he prefers to be cautious (lines 11-12). One might wonder how far Max's attitude is influenced by his organization, which has strict security regulations in place (as we learned in Vignette 2, but also across meetings and in interviews with Max and the others). Max brought up security issues many times, often in relation to his organization, which is why we could assume that his organization here is leading him to disagree. In many ways, Max appears to be the voice of security concerns in the team. It is through him that security aspects are expressed and eventually become coauthored into the collaborative strategy.
Whether to design an open or closed hub is not resolved in this meeting. Tom and Max are authoring different strategies, which also changes their way of addressing one another: while they spoke with a shared voice before, there now exists a me-you dialectic between them, where each seems to be appropriating a different version of their team voice. Tom and Max continue surrounding themselves with additional voices; however, these voices now lay out opposing strategic details. What is best for the team, presented as such by Tom, is no longer also what is best for Max's organization.

Vignette 4: Defending the collaborative strategy
Members find a compromise in an open but anonymized system. Potentially everyone can access this open system, but identities of buyers and suppliers will be sealed and visible only to an intermediate party. A narrowed version of their strategy hence emerges. In Vignette 4 (from the seventh meeting), a guest is present. Phil, who is also a member of the initiative the team is part of, met Ben at a gathering and wanted to learn more about the team's project. Phil does not have any hierarchical relationship to the others, and except for having met Ben once does not yet know the team members. Max has summarized what the team has been working on, including the system's anonymization, when Phil starts sharing his opinion. Phil starts speaking as himself ("I," line 1) but switches to a more collective "we" (lines 1-2) shortly after, when he explains how he already discussed some ideas with Ben. That positions him closer to the team, presumably to appear less of an outsider, and is reinforced when he starts speaking about the hub as an anonymous and confidential system (lines 2-3). When Phil starts elaborating on precisely this anonymity, he invokes the voice of Max's organization, outlining their needs (lines 3-5). While Max allowed his team members to speak as his organization (e.g. Tom in Vignette 1), he does not yet seem to be granting this power to Phil when he already starts correcting Phil while he is still talking (line 6). Interestingly, Max uses a generic vocabulary in his explanations (lines 6-8). Instead of naming his organization or its suppliers (as he generally did before), he speaks about "seller" and "buyer" (lines 6-7). When he cannot think of one word, he asks his team colleagues for help, speaking with their shared voice: "e:::h how do we call that" (lines 7-8), as if they have established their own collective terminology. Max clearly appropriates the team's voice, contrasting the team's collectiveness to Phil's singularity.
Shortly after, Max invokes his organization's voice (lines 12-13); this time to justify the team's strategy rather than to impose his organization's view onto the collaboration (as in Vignette 2, for example). It appears as if Max's organization is no longer proposed as only Max's ally but now also as the team's. While before, making present his organization's voice helped Max to strengthen his authority over the team's strategy, invoking this very voice now helps him to defend the strategy that he coauthored. One could say that, over time, the different voices became pieced together to such an extent that the concerns that characterize Max's organization now also characterize the team and what it does.
When Anna explains some processes in more detail, she too adopts a generic vocabulary, speaking about "the company" (lines 14, 16) or "the suppliers" (lines 14, 15) rather than naming any particular one. She also appropriates the collective "we" (line 17), something she has not often done before but which here stresses how she positions herself as part of the team, in contrast to Phil: Yes and eh so when I was thinking about it and we also discussed a little bit ((looking at Ben)) we were thinking about how you can have a kind of anonymous e:::h eh say procurement type of e:::h especially when it's confidential. Suppose yeah eh [Proto] they will have a new product and they don't want eh they want to get quotes for that but eh-[but you don't want everybody to know 6 7 8 Max: [But we don't share our bill of material [. . .] even-even in this part, the seller of components they eh it is not necessary that they know who the buyer is, there is e:::h how do we call that there is a-9 Ben: Transport company-10 Max: Yes, so there is a transport company in between like bol.com or something like that. Phil. Tom adopts the team's collective voice as well when explaining how they "wanted to try" (line 18) working on a hub. Max adds to that by making present their goal again (line 20). He stresses their focus on checking feasibility and compares their work with other projects (lines 20-21) that are less exploratory, thereby also emphasizing their uniqueness. That is important as it highlights the team's distinguishability as one collective group that is different from the initiative's other teams and therefore also different from the team(s) that Phil, their guest, is part of.
Overall, it seems as if Phil's presence (i.e., the presence of an outsider) makes the team move closer together. Members speak with a much-shared team voice and collectively stand behind their strategy, clearly distinguished from their guest.

Vignette 5: Authoring strategy evaluations
The team manages to build a first prototype. However, as we see in Vignette 5 (from the tenth meeting), Anna concludes that their way of designing the hub has ultimately not been a good strategic choice. Initially, her skepticism is met with counterarguments by Max and Tom. After a short while, however, it appears that Max and Tom start reconsidering their conclusions.
Anna voices first evaluations of the hub. She starts speaking with the team voice ("what we have done," line 1) but then switches to addressing the others as "you" (lines 3, 4, and 9) and to speaking of herself as "I" (line 9) as if wanting to detach herself from the team's strategic choice to include an intermediate party for anonymization. This decision had been made as a compromise between Tom and Max's opposing preferences for either an open or closed system. Remember that the initial decision to work on a hub had also mainly been coauthored by Max and Tom (and their organizations; see Vignette 1). Anna had always been more hesitant toward this strategy as we observed in initial meetings but also learned when interviewing her. However, her hesitations were overruled by Max and Tom who managed to put themselves in more powerful authoritative positions. While Max and Tom's organizational concerns have become closely knotted to the team's activities, Anna and Ben's organization was less present in their collaborative strategy. As just one of many possible consequences, Anna seems to be in a position to easily expropriate the team voice and distance herself from its partly disappointing outcome-an outcome that the others, more than her, are responsible for.
Tom provides an alternative rationale for why the hub was nonetheless a good idea. He brings forward a typical problem that organizations have when ordering (lines 10-21, 13-15), which helps him highlight the hub's benefits. As in the collaboration's beginning phases, he invokes organizational voices to back up his argument. Max takes side with Tom (line 12) and adds additional details for why their strategy has been a smart choice (lines 16-21). Following Tom's lead, he also invokes his organization's voice, positioning it as representing large companies in general (lines 17-19). Both their organizations and large companies in general are hence proposed by Tom and Max to help them defend their strategic choices as "very beneficial" (line 23). This contrast Anna's more negative evaluation and, to some extent, could be interpreted as Tom and Max's attempt to draw Anna back in.
While Anna appears to agree with some of the benefits (line 22), she sticks to her critical reflection and names another aspect to prove her point. Once more speaking with her individual voice ("I"), she talks about the need "to pay attention to the safety of the [hub]" (line 24). That Anna is bringing up this particular aspect is noteworthy: so far, it has always been Max who made present security concerns (see, for example, Vignette 2); now, it is Anna who invokes this aspect as substantiating her negative evaluation of the team's strategic choices. In contrast, Tom appears to be trying to downplay the security issue. Speaking with their collective team voice ("for our specific goal," line 27), he explains how this concern is "not that important" (line 27). While Max's first reaction to Anna is very short ("okay," line 26), a few moments later, he appears to start aligning with Anna, when he acknowledges not only the hub's "benefits" (31) but also its many "drawbacks" (line 32). Interestingly, then also Tom seems to reconsider his evaluation (lines 34-35).
Overall, voices are becoming more plural again. Because Anna switches from appropriating to expropriating and authors an individual reflection, her voice is initially distanced from the team and its strategy. However, through her powerful proposition of security concerns, first Max and then also Tom appear to start aligning with her more critical evaluation. This is an interesting shift in how members' different voices are assembled and configured. Anna's expropriation of the team voice and her invocation of security concerns appear to start a process of re-aligning the group's shared voice, so that eventually all members stand behind the partly negative assessment of their coauthored strategy. This is also reflected in the team's final report that, for example, says: "The If every benefit is actually (.) performed better in a central database or distributed database with an open-source community, then yeah, that's also a nice conclusion of course.
[hub] has many interesting features, however, these are not specifically needed [. . .] there are simpler and better options."

Discussions and contributions
The empirical episodes have provided insights into the rich interactional details of coauthoring collaborative strategy. We have shown how the convergence of different voices into a (more or less) shared strategy is constituted in communication and accomplished through ongoing situated practice. Next, we seek to more profoundly discuss and theorize our findings. We first further zoom into the communicative practices through which collaborators can enact authority and strategic agency, to shape but also to reorient what their collective does, says, and thinks. Afterwards, we sketch a conceptualization of strategy as a relational assemblage (Kuhn, 2021;Nail, 2017) to abstract the workings and implications of multi-voicedness for strategy-coauthoring.

Strategy-making as a communicative authoritative practice of invoking voices
Kuhn argues that authority is about "authoring the configuration participating in the assemblage" (2021: 116).
Having authority over what a (multi-voiced) group works on then means being able to define which voices are part of its strategy, and which ones not. Our ventriloquial analysis has provided rich insight into these strategy authoring dynamics (Vaara, 2010) and confirms communication's great consequentiality for strategizing (Bencherki et al., 2021;Cooren et al., 2015). To some extent, one could describe strategy-making as the practice of speaking for a collective and enlarging one's say by invoking additional voices, as these voices are more than trivial linguistic elements. They can have powerful authoritative effects for how strategizing unfolds in situated interaction, akin to how a few well-chosen words can twist a strategic conversation (Kwon et al., 2014;Samra-Fredericks, 2003). To another and more surprising extent, however, strategy-making also seems to be about detaching from a collective and speaking as an individual. This latter practice seems to offer room for critique, resistance, and possible re-orientation. First, our findings show how professionals can substantiate their strategy-making and strengthen their authority by making present-proposing-additional voices that back up their suggestions. Max and Tom expressed their experience, invoked the voice of a supplier, or invoked their organizations to multiply the agents behind their proposals; comparable to how "Larry [. . .] leveraged his position as Engineering Director and expertise in aerospace engineering" in Kwon and colleagues' study (2014: 286). In both our and their work, it was not only the professionals pushing for their strategy, but also the voices that they surrounded themselves with (Benoit-Barné and Cooren, 2009;Bourgoin et al., 2020). Intriguingly then, our work shows that not all voices are equally loud: some voices appear to be more impactful than others, which also changes over time. Even a voice that is presented as silenced (e.g., Anna's invocation of the voice of security in Vignette 5) can be highly performative.
We therefore second Vaara and Whittington's (2012) call for more decentered and complete accounts of who or what can act as a strategy agent, but also want to extend it: can we get a fuller grasp on who or what participates in strategy-making that also elucidates the nuanced performative effects of each of these agents? Why are some agents, or voices, louder than others? How is it that some voices fall more and more silent in the strategy-coauthoring process, while others amplify?
Second, our findings show how professionals can enact authority over a group's strategy by exercising firm control over the group's collective "we." In the team we studied, the same two members that successfully invoked voices as authoritative allies initially also constituted this "we" and therefore appropriated the right to speak in the team's name. This meant that they could push for including own or organizational interests (voices) in the team's strategy from a privileged position. Their appropriation also meant that they could (de)legitimize others who tried to speak for the team, either giving them the voice to define what to do or refusing their participation altogether (Mantere and Vaara, 2008;Tavella, 2021). Much more than an opportunity provided by the collaboration (Koschmann and Isbell, 2009), our findings show that a collective's voice and the possibility to contribute to its strategy appear to be provided by those interactants that constitute and thereby appropriate and usurp it (Bourdieu, 1991). Just as voices can be asymmetrically proposed, so can a group's voice be asymmetrically appropriated by only some members. Hardy and colleagues' (2005) distinction between cooperative talk (talking in terms of we and us) and assertive talk (prioritizing own interests) or Lewis and colleagues' (2010) separation of we-orientation and me-orientation can hence easily blur in practice: what is presented as cooperative and collective can in fact be greatly assertive and individualistic.
Finally, our findings point at a third practice of enacting authority and strategic agency: professionals can produce powerful positions by expropriating a collective's voice and breaking their association with their group and its momentary strategy. Their expropriation can function as a protection mechanism from undesirable plans or outcomes, but it can also underline professionals' disagreement with and resistance toward a suggested route. As we will explain in more detail in the subsequent part of our discussion, expropriation is therefore not solely a way of distancing from a group's strategy; it can also be a powerful move to re-orient the latter. We think that a more thorough investigation of expropriation practices and their consequentiality for strategy-making is an exciting and currently underresearched path to pursue. While we know about the general power of quitting (Huxham and Vangen, 2004), we yet have little insight into the nuanced and situated unfolding of such distancing work. How exactly is expropriation accomplished, when is it particularly impactful, and why? How precisely is strategic resistance and change enabled through expropriation? And how is it that, sometimes, a silenced voice rings loudest of all? Conceptualizing strategy as a relational assemblage of voices can be helpful to address such and similar questions.

Strategy as a relational assemblage of voices
Our ventriloquial approach has elucidated strategy-making's communicational polyphony (Vaara and Langley, 2021;Vaara and Whittington, 2012), in an interorganizational setting: what the team eventually worked on was constituted by a variety of voices, beyond the present team members. Members' organizations but also their expertise or an imagined supplier showed to be important agents of the strategizing process. This once more attests communication's consequentiality for strategizing and draws attention to the importance and performativity of seemingly subtle communicative details for what becomes a group's strategy. Figure 2 presents the process model that we built to depict how different voices can become relationally integrated into a strategy (indicated by the narrowing funnel shape), through situated communicative practice. Specifically, we propose to call this integrating or blending strategy a strategic assemblage. We theorize that as professionals switch between the coauthoring practices proposing and appropriating (a-arrows), the different voices they invoke are assembled into a temporary arrangement that constructs specific relations between them (Nail, 2017). Our conceptual proposal is that a collective's strategy materializes from precisely these communicative connections or relations (Deleuze and Parnet, 1987;Kuhn, 2021; see also Bencherki et al., 2021) and hence can always be traced back to a "chain of agencies" (Castor and Cooren, 2006: 572, italics added). Thinking in terms of strategic assemblage then means thinking about strategy-making as a pluralistic and polyphonic process of piecing together different voices, enacted in communication through combining the coauthoring practices proposing and appropriating.
The relations between the voices that the strategic assemblage constructs are a matter of degree (Cooren, 2020). They can be more or less robust, at different moments in time. Our findings point at a progressive pattern, where the relational integration that occurred in one stage is not lost in the next: what characterized members' organizations eventually also became what characterized the team and its strategy. Our theoretical inference from this is that voices' relations and their integration into a group's strategic assemblage can become stronger the more interactants switch between proposing and appropriating, and hence between the voices they invoke.
While others have described an unstable back-and-forth of integration and differentiation processes in interorganizational collectives (e.g., Hardy et al., 2005), our work provides a more nuanced account where professionals' iteration between proposing organizational, individual, or other voices and appropriating the collective voice can eventually drive a catalytic, centripetal process (Koschmann et al., 2012). Professionals' oscillation between multiple voices then is precisely what channels their group's plurality toward a pieced-together strategic assemblage, even if at times this means that the assemblage is skewed toward only part of the group. This illustrates how a group's collaborative strategy emerges and solidifies in members' situated interactions and converges with Bencherki and colleagues' (2021) ideas on strategy's progressive materialization. Crucially, however, our insights highlight how a presentified strategic concern can be substantiated by and attributed to many different voices. Paying close attention to these different voices can help elucidate strategy-making's polyphonic inner-workings (Vaara and Whittington, 2012) and problematize by whom or what strategy is enacted. This type of work is important as it can reveal the complexities of language-in-use and their performative implications on strategy; aspects that so far often go unnoticed (Vaara and Langley, 2021).
The right side of our process model ( Figure 2) depicts how voices become relationally separated from a group's strategy (indicated by the widening funnel shape). We theorize that as professionals move from appropriating to expropriating (arrow b), they detach themselves (and possibly the voices they had proposed before) from the temporary arrangement that lays out what the group does. Consequently, these voices are cut out of the chain of agencies (Castor and Cooren, 2006) constituting a group's strategic assemblage: at least momentarily, they do not get expressed in the group's aims and activities. Thinking in terms of strategic assemblage hence also means thinking about strategy-making as a process of moving apart, enacted through switches from appropriating to expropriating.
Expropriating the collective voice can be accompanied by proposing individual, organizational, or other voices (arrow c). Comparable to the progressive integration pattern that we named before, professionals' iteration between expropriating and proposing (arrow d1) can drive a catalytic centrifugal process (Koschmann et al., 2012) and eventually reinforce the relational separation effect. However, professionals can also propose additional voices to sketch an alternative strategy for their group when they (or others) connect their proposed voices back to the collective voice (arrow d2). This starts a new cycle of piecing together an arrangement that lays out what to work on. This is where expropriation becomes a powerful move to re-orient a group's strategy. Anna's expropriation of the team voice and her proposition of Max's security concern is a highly illustrative example that shows how voices that momentarily get written out of the collaborative strategy can have a great effect on what this strategy subsequently becomes or, as in our case, how it is evaluated. One could say that expropriated voices are at once absent and present: they are excluded from a group's collective voice, but their absence has a performative presence that can possibly lead to collective re-orientation.
We believe that it is important to not only consider which voices are written into a strategy but to also consider those that are written out. Convergence and divergence-or centripetal and centrifugal forces-are always both at play in strategy-making (Tavella, 2021;Vaara and Langley, 2021). So far, however, research tends to prioritize alignment over exclusion and therefore leaves unattended or invisible the silenced and marginalized voices of strategy-making. Consequently, we can often explain which stakes are part of strategy and why, whereas we can say much less about those matters that have been excluded (Gray, 1994). Thinking about strategy-making as a relational assemblage and thus as a process of piecing together and moving apart can be helpful to capture more complete and transparent understandings of who gets to coauthor strategy and who is excluded from doing so, at least momentarily. Conceptualizing strategy as a relational assemblage means rejecting "unity in favor of multiplicity" and rejecting "essence in favor of events" (Nail, 2017: 22), which in turn means that analytical efforts need to unfold the processual development of strategy as a performative effect of how different perspectives and agencies are assembled-or not-in communication and interaction.
As the different routes indicate, our model does not necessarily imply a linear process. Professionals can move from proposing to appropriating to expropriating linearly over time, but they can just as well switch between practices or move from one to the other at multiple moments. Moreover, the practices do not by definition imply successful strategy-coauthoring: a proposed voice, for example, can be contested by another proposed voice or can be silenced by someone powerfully appropriating the team voice. We observed a certain order and a broad range of voices in the interorganizational team that we have studied, but this order and the specific voices that are invoked could look a lot different in other settings. For example, which voices are proposed as authoritative allies and what are their nuanced effects on strategy when organizational membership itself is fluid, temporary, or even contested (Bechky, 2006;Dobusch and Schoeneborn, 2015)? Diverse voices might also compete for strategy authorship in less pluralistic settings, such as intraorganizational groups. We believe that our process model can help to elucidate the polyphonic nature of strategy-coauthoring also in such settings. From afar, these contexts might appear more single-voiced, but a ventriloquial lens and fine-grained analysis of how our three coauthoring practices are at play might reveal a much more intricate and complex picture. There is one important boundary condition for the applicability of our model and conceptual proposals, however: the Notes 1. Because members were working on this technological subject, some of the conversations that we present include technological/operational details. Sometimes, the connection to strategy-making might therefore not immediately be apparent. However, it was also through such conversations and technological/operational details that the new collaborative strategy emerged, was refined, and so on, which is why these are important. 2. Note that the team had sufficient freedom to change their activities and objectives. While a plan already existed, reauthoring the strategy was not constrained by important structural factors.