A Multiscalar Coordination of Languaging

3.1 Research site and participants

The research project took place in Monkey King Middle School (MKMS, pseudonym) in Changchun, China. At the time of the study, MKMS was a regular city school, but, more students from the countryside and rural areas attended the school due to the recent trend of urbanization. Because of the low education rate of Chinese farmers, the teachers at MKMS observed that students from the countryside had less autonomy in their study habits and in everyday care from the parents. This meant that the parents from the rural areas did not spend much time taking care of their children’s after-school activities and helping them with homework assignments.

The baseline data and counterpart QA group data were collected in typical classrooms. Each class consisted of about 50 students. Students were sitting in rows as shown in Figure 1 below. Students usually wore school uniforms and spent 8–10 hours per day at school including morning and late-afternoon self-study times. Students stayed together as a class, taking main subjects, such as math, Chinese, and English, in their own classrooms. Teachers traveled to the classes. Standard textbooks and student workbooks were typical learning materials. A subject-area teacher usually taught two classes and served as a head teacher in one of the classes.

Figure 1

Baseline classroom of MKMS

Implementation of QA and data collection took place in the school’s computer lab, which was larger in size than the classrooms. As shown in Figure 2, the space and arrangement allowed students and teachers to move around and have one-on-one interaction. Student seating was not assigned. Friends often sat next to each other and helped each other move along with quest narratives. The researcher and teacher did not specifically encourage them to work together but told them to help each other if someone got stuck due to a technical problem or fell behind in completing a mission.

Figure 2

Students working on the Diversity Mission in QA in the school’s computer lab

3.2 Research design and materials

This article compares the languaging patterns across three different environments: baseline, online (which took place in the school’s computer lab), and print-based version. The baseline group was represented by typical English classroom instruction that followed a standard textbook. The print and online groups were asked to complete a quest addressing diversity. The print-based group completed the quest in class, while the online group completed the quest in QA in the computer lab.

Diversity, in its most conventional definition, was a new topic to the Chinese students. To help students from a monocultural view experience diversity in an international context, the mission portrayed an adopted Chinese girl non-player character (NPC) and a French-Canadian NPC in a school context. This design afforded the players’ development of a sense of respect for multiple perspectives. Diversity Mission was the first mission the students worked on in QA.

QA, an ActiveWorlds 3D educational metaverse with avatars designed for children ages 9–13, was the virtual learning environment for this study. QA allowed users to travel to virtual lands where they selected educational activities (a.k.a. quests) and interacted with NPCs; spoke with other users and mentors via chat, telegrams, and emails; and built virtual personae as avatars. QA was different from other 3D virtual worlds in that it was designed with game-like activities for children (e.g. a mythological back story), a point-rewarding system (e.g. Lumins and Cols as the QA collectible currency), rich graphic object-oriented worlds, and themes of social responsibility. It also included embedded educational quests and instructional affordances for teachers to give just-in-time instructions.

3.3 Data collection and selection

Classroom sessions were video recorded. Screen recordings of the students playing QA were recorded using both Camtasia (screen only) and Morae (screen and live video of participants). The video recordings of the baseline class, the QA screen recordings of a student doing the diversity quest, and the video recording of the same teacher’s class doing the print-based activity on the Diversity Mission were selected for the focal transcription and analysis. The selection criteria were based on the research design assumptions, observations, quality of video recording, and participants’ participation intensity.

3.4 Multimodal transcription, unit of analysis, and coding

The Transana application was used for handling video data. The video segments were transcribed, communicative projects were identifyed and segmented, and the languaging events were coded. In order to determine what materials (semiosis) learners were attuned to, we transcribed different modes of languaging: talk and gestures from classroom video data and talk, gestures, avatar actions, and mouse clicking behavior from screen video data. We adapted Linell’s (2009) communicative projects as a unit of analysis to parse out talk-in-interactions and actions during QA and classroom sessions. Using communicative projects in action as our units of analysis instead of individual utterances solved an analytical puzzle for empirical studies adapting distributed perspectives. It also echoed Kravchenko’s (2011) view of languaging as an interactional behavior occurring in space-time as parties monitor, initiate, and respond. A communicative project was often centered around a task that required concerted effort by two or more individuals. Each communicative project varied in length (represented as a clip in Transana, see the colored square blocks in Figures 4–12) but always carried a coherent thematic meaning.

Figure 3

Keyword groups are the black-colored keys, and subcategories of keywords are the indented, red-colored keys. Distributed meaning-making keywords represent relational dynamics in linguistic actions when meaning-making with others. Individual meaning-making keywords represent actions taken place alone.

Figure 4

Actions in baseline class. Linguistic behavior is dominated by distributed meaning-making. Each color block represents a communicative project where the identified action(s) took place.

Transana coding was programmed through keyword groups and subcategories of keywords (for the relationship used in this study, see Figure 3). The first step of coding was to identify and parse the transcribed verbal and non-verbal action data into communicative projects. The second step was to assign keywords to communicative projects. Keyword groups and keywords in this article emerged in the dynamic process of overlaying distributed concepts, such as affordances and distributed meaning-making over the data, and vice versa. The process was both top-down and bottom-up, thus resulting in an abductive coding process (e.g. Newgarden and Zheng 2016; Zheng et al. 2017).

Communicative project parsing and keyword assignments were checked between the first and second authors in two steps: First, internal reliability within the transcript was checked by referring to the definitions of the keywords top-down and bottom-up during coding. When necessary, definitions were revised (e.g., due to a narrowness of the definition), and thus the previously coded clips were double-checked and revised. Secondly, cross-rater agreement was also checked between the two raters and disagreements were checked, discussed, and resolved once agreed upon. Some keywords such as affordances (Gibson 1979) and parallel and collectivized interaction (Reed 1996) came directly from literature. Others were co-constituted from the dynamic coding process. Those keywords are explained below in the findings.

The multimodal transcription and the Transana keyword visualization tool allowed for a full description and visualization of languaging in terms of salient keywords symbolizing linguistic and non-verbal actions along with other co-varying ties, such as the affordances described below. We then used visualizations of keyword maps generated in Transana to demonstrate patterns. The patterns resulted from post hoc hypotheses and analyses. For example, we first looked at a keyword map containing all the keyword groups. When we observed a density of keywords occurring overlapping or intermittently, we would form a hypothesis and then take a further look at the raw video data to confirm our hypothesis qualitatively and discursively. This method has been applied extensively by Zheng and her colleagues in previously published qualitative investigations (e.g., Newgarden and Zheng 2016; Zheng 2012; Zheng et al. 2012), quantitative studies (e.g., Newgarden et al. 2015), and mixed-methods studies (Zheng et al. 2017).

4.1 Emergent salient keywords of understanding observer-actors languaging

From observations, the most salient actions in the computer lab were students interacting with the computers and advancing the Diversity Mission narrative; in the classroom, the motion was driven by the teacher. Both environments were observed with characteristics of individual meaning-making and distributed meaning-making. These categories were abductively selected and emerged from viewing the data through a distributed lens. Individual meaning-making refers to the individual linguistic actions, whereas distributed meaning-making refers to linguistic behavior done with others. These categories are represented in keywords and the sub-keywords as shown in Figure 3.

The second salient category, the degree in which the interaction was synchronized toward a common goal, primarily drew on distributed literature to further differentiate distributed meaning-making. This category depended on whether a) the students’ repetition with their teacher, peers, or a non-playing character was to remember a fact; b) both parties recognized the nature of the task and took actions together; or c) both parties co-acted to take the meaning of verbal patterns to do things with it. The first was labeled as parallel interaction, following Reed (1996). Parallel refers to mere interaction, a prerequisite for coordination of effort, but only as a prerequisite. Parallel interaction is similar to task-based interaction where learners mainly exchange information to complete a task (Zheng and Newgarden 2012). The second was labeled collectivized interaction and motivation. Reed (1996) believes when both parties recognize a task and its process, this counts as a collectivization of effort. The third was labeled co-action. Co-action is identified as action of feeding forward, anticipation, and risk-taking in order to reach the flow state of co-authoring, co-construction, and co-producing (Cowley 2007). A third party, such as a set of sociocultural norms, is always involved in co-action, either virtual or real. The effort is then distributed throughout the environment, depending on meaning and values (Linell 2009). In this study, co-action consisted of two types: agent and agent, or agent and the agent’s avatar (Zheng and Newgarden 2012). In her study of co-questing interactions in the Second Life virtual world, Zheng (2012) found that collectivized interaction was closely coupled with co-action, implying that when individuals’ utterances and actions served as the context for the party, seamless action moved forward. As shown in the next section, collectivized interaction and motivation is closely related to co-action in this data as well (see Figures 8 and 12). However, co-action does have the additional feed-forward characteristics found in collectivized interaction and motivation.

The third salient category, which also drew on distributed literature to describe emerging affordances of the three environments, was learning opportunities in the distributed perspective. The guiding question in this section was as follows: How did each communicative project provide varying activities for languaging styles? Three subcategories emerged. The first of these was directing and regulating one’s own behavior. The guiding principle for this subcategory was that “language is not a means of transmitting ideas or representations; it is a means of making information available to others” (Reed 1996: 155). What was said and heard could be used to regulate one’s activities and to assist collectivized prospective control. For example, talking between peers was often not just to express a thought or frustration. What was uttered was both “ambient and public” (Hodges 2007b: 589), pointing to a further action in the environment as an individual or collective effort. In other words, languaging regulated the behaviors of both the participants themselves and others. The second subcategory was establishing or maintaining a natural rewarding relationship. To direct and be directed, or to initiate a topic in joint work, “is an act of coordinating multiple values and multiple relationships in creative ways” (Hodges 2007b: 598). A rewarding relationship with the environment and with agents in the environment is usually achieved when affordances and semiotic resources are available to act on. The third subcategory was modifying and constructing one’s own environment. Ecological psychology emphasizes that an agent encounters its environment to construct its niche. A learner’s ability to monitor, track, and adjust to significant variations in circumstances is a learning opportunity for its own sake (Reed 1996).

4.2 Languaging style of answering questions, repeating, and note-taking in the baseline classroom

As with other English classes offered at MKMS, the baseline class followed a teacher-centered tradition: the teacher led the major meaning-making process by giving instruction and presenting corresponding information on the screen. Then the students did substitutive drill practice by using the words or phrases shown on the screen to answer the teacher’s questions. Figure 4 demonstrates students’ actions carried out during the process. They include answering questions, correcting, inquiring, repeating, and writing. The first four fall into the category of distributed meaning-making, featured by their distribution in consensual domains in a typical teacher-centered classroom (Cowley 2019). For instance, students answered the teacher’s questions, following the teacher’s verbal and physical cues (e.g., the teacher paused and looked at the class after asking the class a question). The action of writing fell into individual meaning-making, featured by students’ individual efforts with material attachment (e.g., writing the vocabulary in their notebooks). It is particularly interesting that the writing occurred either before or after repeating a new word, as students followed the teacher’s direct instruction to copy the vocabulary they perceived on the screen, such as “Write it down in your notebook.”

Parallel interaction was salient in the baseline class in the form of information confirmation. More specifically, students responded to the teacher’s instructions by answering questions and repeating the teacher’s demonstrations. As shown in Figure 1, these actions were seldom accompanied by other actions that were ambient and public for probing and exploration. In other words, such actions were mere information exchanges between the students and the teacher for one-time task completion (e.g., saying and repeating the expected answer). They did not necessarily constitute any contexts that would feed toward further (inter)actions, nor were there opportunities for students to take the time to synchronize their task, aligning a common ground (see Newgarden et al. 2015). In the two clips where answering questions was coupled with inquiring, students’ inquiries used the given sentence pattern and vocabulary to do substitutive drills with their partners, and the inquiry–answer exchange in each unit of conversation occurred no more than once.

To investigate how the classroom environment facilitated this languaging style, we also coded the affordances that provided specific contextual information and meaning potential for students’ actions, as well as the learning opportunities such actions entailed. As shown in Figure 5, the students’ distributed meaning-making was coupled with the teacher’s direct instruction, such as “read after me”; scaffolding, such as modeling; body language, such as the teacher looking at the class after asking a question; and correction, other students’ responses to the teacher’s questions (which also served as modeling), and the information shown on the screen. In other words, the teacher took charge in providing and making salient meaning potentials in the learning environment for students to further regulate their actions. Students, on the other hand, seldom got to explore a third party (e.g., by manipulating an artifact or coordinating an activity with their teacher or peers).

Figure 5

Distributed meaning-making in baseline class

As mentioned above, individual meaning-making was only attained by writing in the form of note-taking. As shown in Figure 6, such writing or note-taking action co-emerged with the information on the projected screen and direct instruction. Again, students directed and regulated their own behaviors largely based on what others (i.e. the teacher) foregrounded or made salient in the learning contexts.

Figure 6

Individual meaning-making in baseline class

4.3 Languaging style of individuating multiscalar coordination between reading and writing in the QA Diversity Mission

4.3.2 Other-oriented multiscalar reading

Distributed meaning-making was dominated by collectivized interaction, confirming, and direction, all of which featured co-action between the students and their avatars. Actions and affordances led seamlessly one to another within the same communicative project. As instantiated in Figure 5, the focal student player and his avatar collectively perceived and acted on the NPCs and text. He regularly switched between the first-person view (see Figure 6), with which he projected what I as an avatar am doing (i.e., awareness of my own action at the moment), and third-person view (see Figure 7), with which he was witnessing the actions of other avatars and how others perceived him. This is a significant virtual world affordance for other-oriented action. The ambient action was undertaken at a rapid speed and with multiscalar complexity, such as confirming information with the NPC and changing his avatar’s location according to the given text. By doing so, the student directed and regulated his own behavior according to the instruction he elicited from the NPC, established and maintained a natural rewarding relationship with the avatar in terms of perception and action, and modified and constructed his own semiotic environments by moving around and clicking on different NPCs to obtain more comprehensive and finer information. This demonstrates how the observers-actors and the environment co-constituted each other, creating and affording sense-saturated activity (Cowley 2017).

Figure 7

Actions in QA diversity

Figure 8

Collectivized meaning-making in QA diversity by the focal student

Figure 9

First-person perspective in QA

Figure 10

Third-person perspective in QA

4.3.3 Rapid intervals of meaning-making

Reading was the most prominent action in individual meaning-making within QA. One form of reading is “conversing” with non-playing characters in preprogrammed text-based dialogues – similar to interacting with a bot providing a player with multiple text responses. Reading is also done by directly reading the Diversity Mission narrative. In the former, reading was done in a perception– action loop coupled with clicking links and writing to respond to a quest (as shown in Figure 11). Sometimes during reading, co-action took place between the students and their avatars (as shown in Figure 7). In other words, when students perceived a potentially useful text or utterance of an NPC, they could click on links or NPCs to read the texts. Afterward, and based upon what they perceived during their reading, they a) clicked other links to move on or b) wrote down their opinions as responses to questions, using information they extracted from the texts.

Figure 11

Individual meaning-making in QA diversity

Figure 12

Actions in print-based diversity class

Nevertheless, these sequences of actions were not always linear. There were occasions where the students moved back and forth between the same texts, NPCs, and links to seek more detailed information and figure out how to move on (for similar languaging style, see the mobile place-based data gathered by Zheng et al. 2018). More specifically, in some cases, students scanned the texts when they first encountered them and later on found that they had to come back for more detailed information that was necessary for their progress. For instance, in the co-action that happened right after 10:00 in Figure 8, the student went back to the previous NPC to re-confirm the instruction it provided. During this process, in addition to only directing and regulating their own behaviors as responses to their perceptions, students kept modifying and constructing their own environment and thus created more learning opportunities for themselves.

The learning opportunity provided by modifying and regulating behavior was also identifiable in the languaging of translating. Happening either between reading or writing (see Figure 11), the student used an online dictionary to look up new words to make meaning of the situation he was in. This interdependent looping of reading and writing gave rise to individual meaning-making by using material tools, such as a dictionary, for effective and dynamic coordination. However, and also inseparably, this action had an impact on his languaging behavior, meaning-making, and values realizing all within a short interval of time.

Consequently, the languaging of writing was different from that done in the baseline classroom context. Writing in forms of note-taking was a valuable skill to practice. However, writing in QA was more public and other-oriented. It fed forward into ambient action. It involved self-expression, reasoning, and reflection, contributing to a style of languaging interdependent on dialogical partners and material affordance bundles. As illustrated in Figures 7 and 11, writing in the QA context often happened after the co-action and reading. This rapid style of languaging loop between reading and writing resonates with Xu (Wang 2016): to continue and to use in more creative ways that which has been proven effective. However, this data was more social, actional, and other-oriented.

4.4 Languaging style of self-expressing and reflecting in the print-based Diversity Mission class

In the print-based version class, a handout of the Diversity Mission was used instead of a textbook. The students were asked to think about the questions on the handout before the class. In class, the teacher checked the questions with the whole class by inviting different students to share their answers and then asked the students to do group discussions on those topics again. Due to quality issues with the video of the group discussion, only the class discussion part was able to be transcribed. However, from the class discussion, some interesting phenomena regarding students’ learning experiences were identifiable. As Figure 12 shows, there were no individual meaning-making actions identified in the class discussion. Instead, the dominant action pattern was the parallel interaction between the teacher and the students in the form of students answering the teacher’s question. The dominant language pattern was self-expression. Unlike the baseline class setting, the action of answering questions was always accompanied by self-expressing and sometimes by reflecting as well. This indicated that the learners had opportunities to language about their life experiences and values – action that was more ambient, public, and other-oriented than actions from the baseline classroom but less than actions in the QA environment. As in the baseline classroom, learning opportunities were only created as students directed and regulated their own behavior. Even though the group discussion video was not available, the observation notes provided more evidence that students’ discussion and writing were more toward experiences and outlook of respect for diversity, rather than toward taking notes to remember words.

In one interesting incident during a parallel question–answer exchange between the teacher and students, agent to agent co-action occurred for less than 2 minutes. As shown in Line 1 of CP1, the student raised his hand to actively answer the teacher’s question. In Line 11, the teacher also became emotionally engaged in the topic. From that point on, actions of the teacher and the student began to impact each other’s subsequent efforts, and thus formed a momentum for co-action. Nevertheless, with the teacher becoming stronger and stronger in her opinion and acting as an authority in the classroom, agency faded. This was revealed in the change of the student’s actions moving from looking at the teacher to looking down at the book. Consequently, the co-action ended as well.

CP1: Jiang's clip: Li Ying paper-based QA --> DMM --> no school

1. (0:13:44.0)Jiang raises his hand.

2. (0:13:49.7)Jiang stands up, looks down at his book, answers the question

3. and then looks up at the teacher.

4. (0:13:51.2)Jiang: Students don't go to school.

5. (0:13:52.5)T: Speak loudly, OK?

6. (0:13:52.8)Jiang: The students don't go to the school

7. (0:13:57.2)T: The students what?

8. (0:13:58.9)S: go to school.

9. (0:14:00.6)T: Oh, don't go to school.

10. (0:14:00.7)S: ((laugh))

11. (0:14:01.5)T: Then where can we study? At home?

12. (0:14:04.7)Jiang: Yes

13. (0:14:05.3)T: On the computer?

14. (0:14:05.9)Jiang: Yes.

15. (0:14:07.5)T: But some students maybe play the computer games, not, not study.

16. Do you think so? Now Jiang Zhengyu, for example, I give you lots

17. of freedom. I ask you to study at home. And I give you computer, I

18. put all the textbooks information in the computer, will you study well?

19. (0:14:23.6)T goes towards Jiang, students look at them

20. (0:14:29.0) Jiang looks down at his book while he answered the teacher's

21. question.

22. (0:14:30.2) Jiang: Yes.

23. (0:14:31.0)T: Are you sure? Yes or no?

24. (0:14:32.8)Jiang (keeps looking down): Yes.

25. (0:14:34.7)T: Yes↑ OK, maybe let's have a try. I give your texts and information

26. in the computer, and I ask you to study by yourself. And you don't

27. go to English class, don't listen to Jacelyn's lesson, maybe your

28. grades are better and better? I think so? Do you think so? No. I think

29. if you study in, on, computer. and you don't watch me. Maybe you don't.

30. because I can give you some rules. OK. Now I give you two minutes to

31. write the answer on this paper, OK? OK.