Rewilding Language Education: Emergent Assemblages and Entangled Actions

.

material conditions. Humans learn as part of dynamic ecologies constituted by interaction with other people and engagement with a diversity of material and symbolic tools, artifacts, and environments. In the contemporary era, designed learning materials play a central role in human development and are intentionally created to enhance and make more efficient the learning of specific knowledge and/or processes. Institutional settings such as schools and universities, for example, offer instruction in subject matter with predefined and explicitly stated learning objectives. In instructed and noninstructed contexts alike, contemporary language learners have access to structured programs of study, textbooks and grammars, print and digital texts, dictionaries and translation tools, tutorial websites and apps, video-and audio-based language learning content, and a wide array of tasks and activities designed to enhance and accelerate learning processes. Scholarly inquiry into language learning materials has a long history in fields such as education and applied linguistics (e.g., Garton & Graves, 2014;Guerrettaz & Johnston, 2013;Tomlinson, 2011), yet definitions and theorizations of language learning materials-and specifically, careful empirical analysis of materials use-remain underexplored in much of language education and classroom research (Guerrettaz, Engman, & Matsumoto, 2021, this issue; cf. Vygotskian sociocultural treatments, e.g., Lantolf & Poehner, 2008;van Compernolle, 2014).
In the research reported here, we investigate human action and place-based contexts of language use and learning that involve the use of locative (location-aware) mobile technologies (Frith, 2015), with an emphasis on the relationships constituting human embodiment and communicative action with and through assemblages of material and semiotic resources. Our contribution to the theme of language learning materials use emphasizes process ontology, that is, the idea that material processes themselves are the necessary focus of empirical inquiry. Taking a process ontology approach contends that process "is fundamental, and entities are derivative or based in process" (Sawyer, 2002, p. 286) and further, that "process is not only a guiding orientation but is also the fundamental nature of reality" (p. 291). An analogy from the biological sciences is the distinction between the genotype and phenotype of an organism: Genotype is the basic genetic structure of an organism, a static map of its expressive potential; phenotype is the 'observable characteristics' of an organism resulting from the history of interaction of its genotype with the environment. Two Douglas fir trees (pseudotsuga menziesii) with identical genotypes, for example, may mature into strikingly divergent phenotypes; one seeded in rich soil could exceed 90 meters in height while the second, growing in less hospitable conditions, may attain only a fraction of that size. Similarly, the use of identical learning materials (tasks, textbooks, assignments, technologies, syllabi) do not result in identical-or in many cases, even similar-processes or learning outcomes due the contingencies of differing material processes and conditions (Coughlan & Duff, 1994;Seedhouse, 2005;Thorne, 2003). In the empirical cases discussed here, an emphasis on material processes rather than static entities is further extended to suggest that people and the social-material world together produce assemblages of observable "morphologies of action" (Thorne, 2016, p. 189).
This process-oriented research on materials use presents an empirical micro-interactional examination of adult English language learners playing an augmented reality (AR) place-based mobile game. Mobile AR games, tours, and communitybased education projects typically involve participants ambulating through spaces such as cities and neighborhoods, museum exhibits, historical sites, natural areas, and event venues while using a location-aware (GPS-enabled) device, typically a smartphone. As participants enter (or are guided to) specific places on a digital map, they are presented with additional information such as images, video, textual information, or activity prompts that relate to those locations. The locative or place-based aspect of mobile AR technologies has recently received considerable attention for the potential educational value and capacity to hypercontextualize learning experiences (Holden et al., 2015;. In a parallel to commercial recreational games that have been studied as learning environments (e.g., Gee, 2007;Plass, Mayer, & Homer, 2020;Reinhardt, 2019;Thorne et al., 2012), AR games represent a shift away from models of learning based on decontextualized information delivery and toward approaches that emphasize situated problem solving and learning experiences rooted in particular places. Three instances of materials use are presented, each of which involves small groups of university students playing a mobile AR game out of doors on an urban university campus and adjoining areas.

REWILDING AS A GUIDING PEDAGOGICAL METAPHOR
The focal pedagogical intervention described here involved creating mobile AR activities, the development of which was inspired by research on learning 'in the wild' (Hellermann, Thorne, & Haley, 2019;Hutchins, 1995) and diverse cases of digitally mediated language use outside of classroom settings (Holden et al., 2015;Reinhardt, 2019;Sundqvist & Sylvén, 2016;Thorne, 2010;Thorne, Black, & Sykes, 2009;Thorne, Sauro, & Smith, 2015). Reverse engineering from studies of learning in the wild, we introduce a pedagogical approach termed 'rewilding' for its emphasis on designing supportive conditions for goal-directed interaction in spaces outside of classrooms. The development of our mobile AR activities emerged in tandem with the awareness by one of the authors of rewilding approaches to environmental and ecological restoration. First introduced by conservation biologists (Soulé & Noss, 1998), rewilding encourages reintroducing fauna and flora to ecologies that have lost biodiversity through human habitation, cultivation, and urbanization. Applying the rewilding approach to instructed language education, the challenge becomes how to dynamically augment and integrate formal learning settings with the vibrancy of linguistically and experientially rich engagement occurring elsewhere in the social-material world. In essence, rewilding language education is a renewed call for increasing the ecological alignment of domesticated instructional spaces vis-à-vis the heterogeneity, complexity, and unpredictability of interaction in the wild (the latter defined here as extramural contexts). While not explicitly termed 'rewilding,' establishing linkages between instructed language learning settings and external communities and material environments to structure pedagogical tasks and student-initiated exploration has a substantial history, especially in digital contexts such as online gaming (Reinhardt, 2019; Thorne, 2008Thorne, , 2012, in fandom participation and fan fiction authoring (Sauro & Sundmark, 2019;Sauro & Thorne, 2021), in social media and online communities (Lamy & Zourou, 2013;Thorne & Reinhardt, 2008), as well as in routine service encounters (Piirainen-Marsh & Lilja, 2019) and in the integration of world languages into academic and professional contexts (Thorne, 2013a) and community-based service learning (Dubreil & Thorne, 2017). In our research on mobile AR, the application of rewilding to instructed language learning settings is intended to introduce structured unpredictability to the language learning and use experience by having participants, in teams of three, engage in intentionally underspecified tasks that involve spatial navigation and route finding, collaborative problem solving, and discovering and learning about sustainable technologies on and around an urban university campus. Details of the mobile AR game are described in the empirical analysis section of the article, but in preface, we outline theoretical, conceptual, and methodological issues that inform our approach to learning materials and materials use.

THE USE OF MOBILE AUGMENTED REALITY IN LANGUAGE EDUCATION
One of the first applications of AR technology for language learning (L2 Spanish) was Mentira, a place-based mobile game set in a Spanishspeaking neighborhood in Albuquerque, New Mexico, where learners work together to solve a prohibition-era murder mystery (Holden & Sykes, 2011). University students playing the game complete a jigsaw-style set of tasks in which each player receives different roles, clues, and parts of the story, prompting collaboration with one another and with residents living in the area, to complete the task. In a study of the place-based AR game Guardians of the Mo'o, Zheng et al. (2018) described how "place evokes a learner's effort for making meaning and realizing values through embodied action, collaboration and coordination," and noted that "experiencing place is critical for learners to break away from institutional norms and previous thinking patterns in order to develop skilled linguistic action in actual events that lead to prospective actions" (p. 55). This is illustrated via way-finding activity, such as anchoring their next actions in what is physically present in the environment.
ChronoOps (Thorne, 2013b), a quest-driven mobile AR game used in this study, was created and located in Portland, Oregon, and is currently available in seven languages, including English. Participants play the role of an agent from the future who has time-traveled to the present moment under mysterious conditions. The game begins by describing that, in the year 2070, the planet has suffered massive environmental degradation and they (the player-agents) have been sent back in time in order to learn from the green technology projects that are in evidence on and around the university campus. Once participants arrive at a designated green technology site, they file video reports that describe the advantages and disadvantages of (and in some cases additional uses for) the green technologies at those locations (solar energy, alternative forms of transportation, captured rainwater, and free parking and charging stations for electric cars). The video reports are sent to an artificial intelligence in the year 2070 in the hopes that this information will reverse the environmental cataclysm that is this planet's future.
ChronoOps was intentionally designed as a series of open-ended and underspecified tasks with the pedagogical motivation of having players construct their actions in interaction with the game goals and content. In prior research on ChronoOps,  used ethnomethodological conversation analysis to investigate how groups of L2 English students sharing one smartphone orient to the device and the information it displays, develop practices for way-finding, and use talk to bring shared attention to features of their physical surroundings.
That research emphasized the importance of how the game moves the language experience out of the classroom and how the group dynamic around one device influences students' interactional practices. In related research, Hellermann, Thorne, and Fodor (2017) described the complex interactions associated with the literacy event of reading aloud during mobile AR game play, illustrating that collaborative practices for playing the game that involved reading emerged and consolidated over the duration of the activity. Addressing the hypercontextualization and place-based potential of AR, Thorne and Hellermann (2017) analyzed video data of ChronoOps game play and described how problems in understanding, as well as moving next actions forward, are often enmeshed with and supported by the immediate physical environment. Their analysis demonstrates the relevance of embodied and distributed approaches to human activity, illustrating that participants utilize gaze, gesture, vocalizations and talk, pointing, and embodied deixis, in an orderly manner, to coordinate virtual-digital (iPhone) and sensoryvisual information, to navigate to next locations, and to complete the oral narration tasks comprising the game. Looking cross-linguistically, Hellermann, Thorne, et al. (2019) investigated the orientation to the built environment during the AR game process by groups from three different languages. The groups noticed and used a water fountain for the game task even though it was not explicitly part of the game design. This shows both the improvisatory nature, and further, the participant construction and use of material affordances in the environment that were emergent of the place-based aspect of the AR game.
The empirical research reported in the present article takes materials use as its central theme. The concept of rewilding, introduced previously, describes the pedagogical approach used to create the materials we analyze-in this case, a progressive series of mobile AR tasks (though as mentioned, rewilding refers more broadly to ecologically aligning instructionally situated activities with extramural contexts and practices). All data are unique to this paper and are analyzed using sequential analysis coupled with insights from 4E cognition and sociomaterialism, each of which are defined and further discussed in the next section.

Humans as Open Systems and 4E Cognition
A diverse array of scholarship, from dynamic systems theory to ecological and sociomaterial approaches, sees humans as relationally porous and open systems, with the implication that development arises as a function of interaction within historically formed-and dynamically changingsocial, symbolic, and material ecologies (de Bot et al., 2013;Pennycook, 2018a;van Lier, 2004). This view aligns with a number of approaches that theoretically and empirically redefine cognition as embodied, embedded, enacted, extended and distributed (i.e., 4E cognition; e.g., Atkinson, 2010;Bucholtz & Hall, 2016;Clark, 2008;Hellermann, 2018;Hutchins, 1995). Extended and enacted cognition refer to approaches to understanding human action, such as thinking and communicating, as processes that are fundamentally supra-individual and that include, but are distributed beyond, neuronal activity of or symbolic representation in the brain. When viewed this way, human activity and development constitute an 'ensemble' process along a brainbody-world continuum (e.g., Spivey, 2007).
The term 'distribution' emphasizes the idea that thinking and doing involve the body and coordination between human as well as nonhuman artifacts and environments. In essence, neither the brain nor the individual are the exclusive loci of cognition; rather, the focus is on understanding the organization of assemblages, or 'cognition in the wild' (Hutchins, 1995), which presumes an ecological view of human action as organized by the interplay between persons and resources that are distributed across social and material environments. An important constraint is that the principle of distribution, applied to both cognition and communicative activity, is not meant to imply a necessary symmetry among individual humans, groups of people, artifacts, or environments (Kaptelinin & Nardi, 2006). Instead, the suggestion is that the density of activity can shift from brains to bodies and to a range of physical and representational media in the network and flow of activity (e.g., Cowley, 2009;Thibault, 2011;Thorne, 2016). The notion of distribution suggests an additional entailment, namely that of units of analysis such as 'organism-environment systems' (e.g., Järvilehto, 2009), which describe how change within an organism is accompanied by change to the environment and a reorganization of organism-environment relations. In these ways, distributed, situated, and extended approaches to cognition suggest that human action and development are fundamentally emergent of, and contingent upon, temporal, social, and material conditions, a viewpoint that is more strongly voiced in sociomaterial theorizations, discussed in the next section.

Sociomaterial Perspectives and Sequential Ethnomethodological Analysis
Sociomaterial contributions to this research include the illuminating lens afforded by taking 'matter' seriously, and in many disciplinesincluding language education and applied linguistics (e.g., Canagarajah, 2018;Fenwick, 2015;Toohey et al., 2015)-eliding a central or sole focus on humans has opened up compelling opportunities for deep engagement with the question of what, in actual fact, it means to be human (Pennycook, 2018a). At a general level, sociomaterialism 1 makes visible often unaddressed biases in the doxa (Bourdieu, 1977), or seeming common sense validity, of human exceptionalism, that is, the belief that humans are of a special kind in the universe, and by extension, that human-centric interpretations of the world are self-evident (for example, the right to exploit the natural environment without regard to consequence). More specifically, a number of sociomaterial concepts have illuminated new possibilities for interpretation and analysis. In application to this research, these sociomaterial concepts include entanglement, assemblages, and emergence.
The notion of entanglement (Barad, 2007) eschews simple binaries and emphasizes the relational constitution of the material universe. Assemblage (Deleuze & Guattari, 1987) constrains the principle of entanglement by narrowing the scope of inquiry to a heterogeneous constellation of relations that, perhaps only for a moment or with greater temporal durability, produce certain actions. Emergence helps to avoid presuming a priori states and relationships and rather refocuses attention to how assemblages produce kinds of social-material ontologies. In a critique of social science research, for example, Latour (2005) stated that "it seems that the most important decision to make before becoming a social scientist is to decide first which ingredients are already there in society" (p. 28). Rather than reifying social aggregates, Latour's suggestion is to trace processes through which the social is generated, which is to say, emphasize emergence (a chapter in Latour, 2005, is titled "No group, only group formation" p. 27). Finally, the sociomaterial perspective has enhanced our conceptualization of the term 'mediation,' with its frequent association with Vygotskian sociocultural theory (e.g., Cole, 1996;Vygotsky, 1978). Appadurai (2015) convincingly argued that "mediation and materiality are coproduced effects, which never exist apart from one another (. . .) [and] are best treated as mutual conditions of possibility and as effects of each other" (p. 233). Framed this way, mediation is a 'mode of materialization,' and from the phenomenological vantage point of human experience, the senses (seeing, hearing, feeling, smelling) are the modes of materialization through which "matter becomes active, vital, energetic, agentive, and effective in the world around us" (Appadurai, 2015, p. 233).
Regarding commensurability between sociomaterial approaches and sequential analysis informed by ethnomethodological conversation analysis, we note that Latour himself has remarked that sociomaterial approaches (actornetwork theory in his case) are "simply another way of being faithful to the insights of ethnomethodology: actors know what they do and we have to learn from them not only what they do, but how and why they do it" (Latour, 1999, p. 19). We find ethnomethodology to complement sociomaterial approaches because of its radically emic approach to the analysis of interaction. An ethnomethodological approach seeks to uncover the practices that members use to produce actions, not through the application of a priori theory but through careful observation and the microinteractional description of accountable actions that members use to produce and maintain intersubjectivity. Members are accountable to one another for working toward intersubjectivity, and accountable action can be instigated by phenomena outside the minds, bodies, and voices of the human participants (Garfinkel, Lynch, & Livingston, 1981). As described in a recent article applying a sociomaterial lens to medical education, Medical education is a messy tangle of social and material elements. These material entities include tools, like curriculum guides, stethoscopes, cell phones, accreditation standards, (. . .) natural elements, like weather systems, disease vectors, and human bodies; and checklists, internet connections, classrooms (. . .) and an endless array of others. (MacLeod et al., 2019, p. 177) In our research, assemblages of people, humancreated objects, and material entities are entangled catalysts for communicative action (the latter our primary focus). An example from recent research  involved mobile AR participants reading aloud a prompt for the task from the iPhone that asked the group to brainstorm uses for collected rainwater. The group was walking toward a fountain and as it came into view, one of the participants named it as one of the uses for collected rainwater. In this way, physicality of the water fountain-its visibility, size, and the sound of the flowing water in the fountain-prompted the noticing of the fountain by a participant, and subsequently the fountain became a resource to list and discuss as part of completing the task. Informed by a sociomaterial perspective, the excerpts and analysis that follow trace temporally sequential actions among heterogeneous assemblages of human and nonhuman actants that contribute to understanding not what things are, but what they do, and do together, in the emergent unfolding of entangled actions that comprise language learning materials use.

EMPIRICAL CASES
Our analysis of AR materials use draws on video recordings (18 hours) of small groups of students playing the quest-type AR game described previously: ChronoOps (Thorne, 2013b). The game was played by learners of a number of languages including English, French, German, Hungarian, Japanese, and Spanish. Video data of three-player groups were transcribed using conventions from multimodal ethnomethodological conversation analysis (Mondada, 2018; see the Appendix). The video data and transcriptions were then analyzed in group data analysis sessions. The analysis draws from multiple approaches (primarily multimodal ethnomethodological conversation analysis with concepts and terminology drawing from sociomaterial approaches) and illustrates the achievement of ongoing co-action through observable embodied displays, the performance of new actions through coordinated (re)use of public semiotic and material resources (Goodwin, 2013(Goodwin, , 2017, and objects and the physical surround as actants in the sequential production of action (Latour, 2005).
The analytical focus of the case studies in the next section centers on the actions of the human participants-for after all, our primary concern as social scientists interested in real-time communicative practice is to investigate human languaging, and more specifically here, how people use and learn additional languages in settings that include materials designed for this purpose as well as affordances such as other people, artifacts, and proximal environments. What we discover in sequential analysis of mobile AR game play underscores the relevance of a number of principles advanced in recent sociomaterial research and philosophy-namely, that human participants, created artifacts such as mobile devices and learning materials, and social-material context are relevant and necessary for holistically describing the observable and audible activities represented in the video recordings (Pennycook, 2018b;Toohey, 2019).
Although the instructions for how to play the game were minimal by design, there were fixed aspects of the AR activity that all participants had to engage with: Reading the game narrative and instructions for the game itself, using the GPSenabled map to find the five technologies named in the activity, and responding to the prompts for making a report on sustainable technology at each location. Around these fixed aspects of the game, a number of contingencies arise due to the multiple routes game players may take to the different locations and the different environmental features that may become relevant material and semiotic resources participants use to complete the tasks comprising the place-based AR game. The empirical analyses that follow sequentially describe three assemblages, the nonhuman elements of which are a water fountain, a rectangular metal utility plate embedded in a sidewalk, and the sun.

Case 1: A Water Fountain Beckons
The following excerpts from Case 1 illustrate how joint attention to a contingent environmental artifact is achieved by a group. The group has completed three of the AR game tasks at specified destinations and has just arrived at what they deem to be the place to stop for the fourth destination represented by the blue circle in Figure 1. The fourth destination names a building, the Academic Student Recreation Center (pictured in Figure 1) and highlights the fact that the building collects rainwater on its roof. A large fountain (which the group refers to later) is marked by the dashed circle. The exact text of the task that participants read from their smartphone is as follows: "The toilets in the Academic and Student Recreation Center (ASRC) flush with rainwater collected from the roof. What are some other ways that rainwater can be used? Record your answer in the game's notebook." It should be noted that the 22 groups we have recorded doing this AR activity stopped in a number of different places on or near this plaza adjacent to the named building to complete Task 4. The place to choose to stop to complete the task is a first contingency. Some of the ways that groups thought about the use of rainwater included washing cars, watering plants, community gardens, washing windows, and producing drinking water. The uses that each group named were related to where that group decided to stop to complete this task and what structures were perceptible to that group at that time. This shows the role that place and the contingent nature of stopping at particular places played in the task-as-process. The group in this example was part of a pedagogical intervention that had expert speakers of English play the game with English language learners. Their pseudonyms are Sula, Wes, and Rex (pictured from left to right in Figure 2). For their report at Destination 4, they stop about one meter to the north of a large water fountain. The rushing and falling water in the fountain is audible in the recording.
In Excerpt 1, the group has just stopped walking when in line 3, Wes, who is holding the phone, reads aloud part of the text from the phone. During the 12-second gap (4), Sula also appears to be reading from the text. Her gaze is on the phone and her lips move. In line 5, Wes makes relevant to the group one aspect of the text just read ("water") which receives confirmatory receipt by Sula (6). However, immediately afterward, Sula and then Rex inquire about the question from the text (8-9). Wes then locates the question in the text and reads aloud: "What are some other ways that rainwater can be used?" After his reading, Sula highlights the water (as Wes had in line 5) but specifies the type of water ("rainwater," 13).
In Excerpt 2, after a 5-second gap in talk in line 14, Wes shifts his gaze away from the phone and toward the water fountain (Figures 3-4). Rex provides a minimal receipt token to either Wes's turn (11-12) or Sula's turn (13) and just after Wes proffers a nonverbal second pair part to the question prompt that he read in lines 11-12, he produces the indexical and iconic gesture in Figures 5 and 6 (15). Figure 6 shows that Rex's gaze shifts toward one end of the fountain where Wes had originated his gesture.
EXCERPT 1 EXCERPT 2 However, at just this point, Sula shifts her gaze upward ( Figure 5) and her turn at talk (17) shows she is oriented to the collection area for rainwater (the roof of the building). Her turn is an overly explicit reformulation of the prompt for the task (17, 19), which receives minimal receipt from Wes. This reformulation orients explicitly to a different source for the response to the task than that to which Wes and Rex are oriented.
In Excerpt 3, Sula continues with a reformulation of the task question itself (24) and a tag exhortation (27) after which Rex joins her in visually scanning the area.
To summarize: at the start of the group's performance of this task, all members of the group orient to the immediate environment as a resource for task completion. There are, however, different alignments with Wes and Rex oriented to the water fountain nearby; Sula, to the roof of the building. The different alignments are resolved in the following excerpts.
After the 4-second lapse in talk when the three participants make cursory looks around the area, in Excerpt 4, Rex makes a query about the water in the fountain (29) just in front of the group. Midway through his turn, he indexes the fountain EXCERPT 3 EXCERPT 4 with his right hand (Figure 7) which Sula orients to by shifting her posture and gaze to the fountain ( Figure 8) and then pointing to it (Figure 9). Wes also shifts his gaze to the fountain in Figure 9. When there is no response to his query, Rex adds incremental expansions to his turn: the proposal (31) and the 'or choice' tag (33).
In Excerpt 5, Sula responds to the query with some speculation and accounts (34-35; 38, 41). Rex's turn in line 42 is a suggestion that is initiated with a consequential marker so connecting the lack of understanding of the origin of the water in the fountain with the need to investigate. Sula gives minimal verbal receipt to the suggestion but then initiates the movement in the direction of the fountain.
Excerpts 1-5 illustrate the relevance of contingencies a group encounters during the AR activity and the work done by the group to shape and make sense of those contingencies for purposes of accomplishing the activity: making a report about uses of collected rainwater. The different orientations to the competing environmental contingencies (made public by group participants via gaze, gesture, and talk) are resolved when a verbal query is made about a fountain that had been noticed and oriented to by two group members.

FIGURE 10
Trek, Max, and Prius (From Left to Right), Participants in Excerpts 6-10 [Color figure can be viewed at wileyonlinelibrary.com]

Case 2: A Metal Utility Panel Serves Language-Related Work
The second case illustrates how gesture and a contingent environmental artifact help the participants define a term from the game text. In these excerpts, a different group (from a classroom of upper intermediate English language learners, Trek, Max, and Prius (see Figure 10) has arrived at the second destination, at which they need to discuss the advantages and disadvantages of solar energy.
The following is the full task prompt as it appears in the game: "The solar panels on the roof EXCERPT 5 EXCERPT 6 of Lincoln Hall supply almost 3% of this building's energy. What are some of the advantages or disadvantages of using solar energy? Record your answer in the game's notebook." The group stops outside the front of the Lincoln Hall building. As the text indicates, the solar panels are on the roof of the four-story building and are not visible from the street level. In Excerpt 6, Max is reading the task prompt aloud (3-7) and then lists some advantages (9-16). There is minimal verbal and nonverbal receipt: Trek's repetition (10) and weak agreement (12), Prius's nod (15). As Max continues with the preparation for the report by projecting a list of disadvantages of solar power (18), Prius reaches toward the phone and possession of the phone is transferred during Max's turn at talk (Tuncer & Haddington, 2020) in line 19.
Prius's acquisition of the phone forecasts an upcoming trouble source for the activity, as we see in Excerpt 7. Max continues with the list of disadvan-tages of solar power (21, 23) which is supported by Trek's repetitions and receipt tokens, and as that sequence comes to a close, Max shifts his posture to Prius (25) and begins to reach for the phone to begin the recording of the report as he utters the transitional (Beach, 1993) item okay (25). At that point, Prius initiates repair (27) by requesting a definition. At the start of his repair initiation, Max stops his reaching for the phone and shifts his posture away from the phone (27), which allows Prius to then show the unfamiliar word to his peers on the phone screen.
In Excerpt 8, when there is no uptake to his repair initiation, Prius continues with the question in line 28 and the repetitions of the trouble source word in lines, 32, 34, 36 producing the trouble source with different pronunciations (Kääntä, 2017). Trek displays her understanding of Prius's request when she repeats the word (38) and begins a depiction of a generic panel. The depiction starts with Trek moving into position EXCERPT 7 EXCERPT 8 ( Figure 12) and then moving her hands and arms to indicate a rectangle in the air, one side of which is at the level of her face and which is made in a plane that is directed specifically at Prius and which Prius orients to (Figures 13-14).
After Trek's depiction, in Excerpt 9, Max repeats the trouble source word (40), shifts his gaze to Prius, and begins a turn that includes an agreement token, a generic categorization, and then a deictic marker (43). Trek offers a more specific description in lines 44-45, 47 ("it's the thing that holds the sun"), which is in overlap with Max. In line 46, Max adds to his deictic marker with a comparison this size as he identifies a rectangular metal plate on the sidewalk at his feet (Figures 15-16) and then reformulates the information given by Trek (48). Figures 15-16 show that Max has placed his left foot on one side of the panel and moved his right foot from one corner of the panel toward the top side to emphasize the rectangular shape.
As Max produces that turn, Excerpt 10 illustrates how Prius shifts his gaze downward toward the panel on the sidewalk and utters the change of state token (49). Trek also orients to Max's depiction and steps on the panel as she utters a strong agreement and explicates further, giving the usual color of a solar panel-not the gray of the metal EXCERPT 9 EXCERPT 10 panel on the sidewalk, but black. Prius also points to the panel to display his orientation to the referent being used to illustrate the trouble source word and makes another change of state token with the sequence closing marker (53).
After Max checks on Prius's understanding of the meaning of the word panel (54), Prius closes the sequence with an indication of a change in his epistemic state (55-56) and the group moves forward with making their report on solar panels.
The sequence of Excerpts 6-10 illustrates how when one group member displays a lack of access to a lexical item from the game text, other participants in the group can find an immediately visually perceptible feature of the built environment (in this case, a metal plate in the sidewalk on which they stand) to aid in their depiction of the trouble source lexical item (see also Eskildsen, 2019).

Case 3: Sun = Rainwater: Accounting for Weather During a Report
A third example illustrates a perceptual mismatch between the instructions for the activity and the contingent environmental condition of local weather. The mismatch is accounted for during the group's report at Destination 4 show-ing the highly improvisational nature of materials use and materials creation during these kinds of 'rewilding' tasks. In this excerpt, a third group (also from a classroom of English language learners, Sam, Rick, and Red; see Figure 18) has arrived at the same destination as the group in Case 1, in which they need to discuss other uses of rainwater.
The following is the full task prompt for Destination 4 as it appears in the game: "The toilets in the Academic and Student Recreation Center (ASRC) flush with rainwater collected from the roof. What are some other ways that rainwater can be used? Record your answer in the game's notebook." The group reads the prompt for Destination 4 and begins discussing what to include in their report as they walk to the destination. They mention several uses, including public drinking water, cleaning the streets, and watering plants on farms. When they arrive at Destination 4, they position themselves in the configuration for making a recording ( Figure 18) and Rick indicates (Excerpt 11, line 9) that he will make the report. When Red asks how he will do that (12), Rick asks for the topic of the report one more time (13) and when this is clarified, he turns to the camera operator, Sam, and makes a hand gesture with his right hand to Sam ( Figure 19) and then claps his hands, which is taken as a signal that he will start the report. At the start of the report, there is a misuse of a term; when that is resolved, Rick starts the report a second time (see Excerpt 12). He greets the audience and introduces the report's subject matter (20)(21). At the end of his turn, Red starts a turn that is prefaced with two disjunctive discourse markers, the reduced well and actually, which is hearable, despite or because of its design, as a dispreferred action turn-shape (Pekarek Doehler & Pochon-Berger, 2011) that breaks progressivity and is interpreted as interruptive by Rick. Evidence for this is in line 22 when Rick shifts his posture toward Red and makes a two palms up gesture (Bavelas et al., 1992) "to express uncertainty" (Chu et al., 2014, p. 700) and the open-class repair initiation (24; Enfield et al., 2013).
As Red continues with his assessment of Portland weather, the alignment to this shift in topic gradually becomes shared. In line 27 of Excerpt 13, Rick begins to laugh at the narrative that Red is presenting: a report of long-suffering rainhabituated residents enjoying the sunny weather at the time of the recording. And as Red continues his assessment, 'it rains throughout the year,' Rick collaborates to co-construct the assessment describing the current situation ("it is sunny today;" 29). Red continues building his turn with a contrastive clause (30) and raises his arms and gaze to the sky as he says "suddenly." Again, Rick makes a collaborative completion (31) to Red's turn, uttering the key point of the assessment, the punchline, the term sunny.
Red repeats Rick's predicate from line 31 in line 32 and they quickly close the assessment sequence in lines 33-34 and the report on other uses of rainwater started by Rick in Excerpt 12 (20-21) continues in lines 35-36.
The series of Excerpts 11-16 shows that a report on uses of collected rainwater was prompted by the game instructions and was planned. One turn EXCERPT 12 EXCERPT 13 into the planned report, however, Red makes relevant the contingent aspect of the weather, and particularly that it was a sunny day in a city that is well known for its gray skies and consistently rainy conditions. The disjunctive aspect of reporting on uses of collected rainwater while standing in bright sunshine hits him-actually as well as metaphorically-in the eye, and Red treats it as an accountable matter. Though his peer had already started the report, Red's initiation of a side sequence plays on the irony that although it is sunny on the day of this recording, it is relevant to speak about using collected rainwater because the climate of Portland allows for that. Red's extemporaneous contribution also indexes the insider status of the participants-that they know the city and region well enough to take ownership of the task and remark on uses of rainwater under uncharacteristically blue skies-and by doing so, they perform a culturally and meteorologically framed new action built upon the publicly visible local resource of weather (Goodwin, 2013).

DISCUSSION AND CONCLUSION
The empirical cases of materials use described here illustrate that communicative action is multimodal, embodied, and embedded in material environments that catalyze action among a heterogeneous array of humans and nonhumans. The use of a mobile phone for the place-based AR game provided a framework for cooperative action among the participants and with their environment. Sequential-temporal analysis showed how human actions such as gaze, pointing, reading aloud, bodily deixis, and audible communication are used in an orderly manner to achieve and maintain intersubjectivity; and importantly, that such human actions, enmeshed with nonhuman contributions (a fountain, a metal panel, the sun) together produce the social (Latour, 2005). Even while maintaining a focus on human agents, shifting to the sociomaterial unit of analysis-the assemblage (Deleuze, 1988;Fox & Alldred, 2015)-affords helpful perspectives on our three empirical cases. Tracing from the human participants' actions, the assemblages include (a) a group standing where their sightlines present a water fountain leading them to investigate how the fountain works and its source of water, (b) a member of the second group encounters a trouble source word while standing near a metal utility panel embedded in the sidewalk and its rectangular shape and size is used to describe and define the trouble source word, and (c) while making a report about uses for captured rainwater as they face direct midday sun, a member of the third group produces an extemporaneous side sequence-namely, that Portland is a rainy place-to account for the irony of squinting into bright sunshine while discussing rainwater. From the vantage point of relational sociomaterial ontology, each case shows how emergent assemblages divergently accomplish what are, ostensibly, the same (Cases 1 and 3) or a similar task (Case 2).
And we should not be surprised by this. All learning, and indeed all activity, in the phenomenal universe, is entangled in the process ontology of unfolding activity across temporal scales. As Tomasello (2000) has described it, "all linguistic knowledge . . . derives in the first instance from the comprehension and production of specific utterances on specific occasions of use" (pp. 237-238), where each 'occasion of use' is situated in, and potentially catalyzed by, a particular social-material context (see also Eskildsen, 2015). In essence, everything happens in a place and at a time, while each place and moment is sequentially nestled against immediate temporal and material neighbors and also entangled with more distal flows of space, time, experience, and memory. As such, when language learning materials are considered genotypically, as stand-alone entities-blueprints for action-or treated as isolated objects decontextualized from their mediational role in human activities, the result is an impoverished and potentially inaccurate representation of the phenotypic dynamic comprising situated brain-body-world activity.
As described in the introductory article to this special issue (Guerrettaz et al., 2021), definitions and theorizations of language learning materials and materials use remain underexplored in the majority of language education and classroom research. The analyses presented here attempt to provide a fine-grained process-focused account of emergent heterogeneous assemblages that illustrate the surprising and creative morphologies of action that emerged in mobile AR game play. Theoretically and pedagogically, this research illustrates the significance of embodied, extended, and ecological views of cognition and learning as 'material processes,' and in particular, it outlines the affordances of mobile AR as one example (among many) of rewilding approaches to creating situated and affordance-rich conditions for action and language learning.
The biologist Lewontin (1974) described the goal of population genetics as mapping between two spaces, from a genotype space, representing the expressive potential of genetic material, to the phenotype space, the observable characteristics of that genetic materiality in situ, when entangled in assemblies of myriad other entities and emergent dynamics in the process ontology of change over time. The development and use of language learning materials run parallel. We design for the desired phenotype space but can never fully predict, or sometimes even imagine, what will happen there (and perhaps this is the art and alchemy of pedagogical innovation). It is for this reason that detailed, microinteractional, moment-bymoment investigations into materials use are so critically important: They offer insight into the ways that learning materials, as purpose-designed constituents of broader human and nonhuman assemblages, potentially enhance (or have some other effect on) the developmental trajectories of the people who use them (for research specifically describing language use and learning in mobile AR, see Sydorenko et al., 2019, for a discussion of language-related episodes; and Sydorenko et al., 2021, for analysis of directive language use).
In closing, we share a few less formal comments on our design and research experience with mobile place-based learning materials. Humans remain the analytic focus of this work due to our primary concern with language development and the building of social and material infrastructures for learning (Wagner, 2015). The rewilding approach informing our development of mobile AR materials was guided by the notion of 'structured unpredictability,' a game design principle used to enhance engagement by providing adequatebut-limited explicit guidance (structure) so that players evolve their own practices for setting and accomplishing objectives in environments involving some degree of probabilistic, unexpected, or stochastic variability (unpredictability). In essence, when creating rewilding opportunities for language learners, the suggestion is to provide pedagogically informed resources and guidance, but not too much.
The mobile device, AR game, human participants, and features of the material surroundings form material assemblages of humans + artifacts + environment(s), together yielding "the raw materials out of which signs grow, and from which language emerges" (van Lier, 2004, p. 63). In the analyses, these assemblages were shown to produce entangled morphologies of action that were spontaneously emergent and that incorporated elements that extended beyond the learning materials in each instance of materials use. During the iterative process of creating this modest mobile AR game for language learners, analyzing player actions to get a sense of the phenotype of the game-in-use, and then modifying the game for clarity and flow, two recurrent catchphrases were frequently repeated among the researchers. The first was, "simple game mechanics yield complex human behaviors." This phrase arose because we initially fretted about the simplicity of the game, which-while based on a narrative role-playing scenario-was essentially comprised of a serial set of location-specific oral narration tasks based on linguistically uncomplicated prompts (the latter so that language learners at many levels could participate). We were surprised and intrigued by the complexity of observable player behaviors and the emergent assemblages that formed among people, objects, and the built and natural environment. As Sacks (1984) remarked, "from close looking at the world we can find things that we cannot, by imagination, assert were there" (p. 25). For readers interested in language learning materials, iterative design, and pedagogical innovation more broadly, we leave you with the second catch-phrase, which was, "what would something better look like?"