Beyond learning opportunities: focused encounters in a sociocognitive approach to second language acquisition and teaching

2.1 Interaction

Sometimes considered the greatest sociologist of the latter 20th century (e.g. Collins 1988), Goffman defined social interaction as “the reciprocal influence of individuals upon one another’s actions when in one another’s immediate physical presence” (1959]: 15).^[3] Goffman (e.g. 1961]) further proposed that focused face-to-face interaction creates its own powerful form of immediate social reality; thus, if you and I are engaged in serious conversation, our primary reality is the joint-attentional, intersubjective, interaffective, and multimodal reality we are co-constructing and experiencing at that moment; everything else is secondary. Additionally, our normative social conduct in such interactions – gesture, gaze, facial expressions, tone of voice, body positioning, etc. – not only shapes that reality but can destroy it: We are extremely sensitive to others’ moment-to-moment expressive actions because they are the ultimate stuff human cooperation is made of, based on which “hyperprosocial” (Marean 2015) human animals survive/exist. The general question we investigate in this paper is: If learning/teaching are interactional, as is widely agreed in SLA/T studies, and if Goffman’s distinctive understanding of interaction has merit, then how can it help us to understand the role of interaction in SLA/T?

2.2 Focused encounters

Goffman (e.g. 1961], 1963]) referred to focused face-to-face interactions variously – as “encounters”, “focused gatherings”, “situated activity systems”, etc. – which Kendon (1990) helpfully reduced to “focused encounters” (FEs). Goffman (1961) described FEs as:

one type of social arrangement that occurs when persons are in one another’s immediate physical presence …. For the participants this involves: a single visual and cognitive focus of attention; a mutual and preferential openness to verbal communication; a heightened mutual relevance of acts; an eye-to-eye ecological huddle … maximiz[ing] each participant’s opportunity to perceive the other participants’ monitoring of him [sic]. Given these communication arrangements, their presence tends to be acknowledged or ratified through expressive signs, and a “we rationale” is likely to emerge, … a sense of the single thing that we are doing together at the time …. [FEs] provide the communication base for a circular flow of feeling among … participants, as well as corrective compensation for deviant acts. (17–18)

Let us unpack this dense description, simultaneously suggesting relevances to current issues in SLA/T studies. In our view, FEs have seven major features:

1. Joint Attention (“Single visual and cognitive focus of attention”) – Interactors sharing a common attentional focus are engaged in joint attention, which has been widely studied for its role in human development and communication (e.g. Tomasello 2014). In contrast, individual attention has played a key role in the historically dominant cognitivist school of SLA/T studies (Schmidt 2010), including the cognitive-interactionist approach introduced above. Cognitivists often treat individual attention as a requirement for in-the-head learning, whereas joint attention is a sociocognitive phenomenon enabling intersubjectivity, interaffectivity, and ecosocial action. Heft (2013) further regards “joint attention [as] the basis of all pedagogy” (165; see also O’Madagain and Tomasello 2022).

2. Openness to Communication/Participation (“Mutual and preferential openness to … communication”) – This term signifies a pro-communicative orientation, a prerequisite for all communication. van Lier (1996) proposed receptivity – an “active response” to language featuring “curiosity” and “exploration” – as an underlying condition for SLA/T, leading to “engagement” (see below). Partly following Goffman, Evnitskaya and Berger (2017) introduced the concept of willingness to participate (WTP), as marked by students’ multimodal displays of willingness to respond to teacher prompts, or to show attentiveness if unable to do so. WTP is a social variant of the individual-differences variable willingness to communicate (MacIntyre et al. 1998) and incorporates Goodwin’s notion of “embodied participation framework”, described below.

3. Engagement (“Heightened mutual relevance of acts”) – Goffman (1963) employed “engagement”, “engrossment”, and “involvement” synonymously to characterize states of “cognitive and affective engrossment …, some mobilization of one’s psychobiological resources” (36): “When an individual becomes engaged in an activity …, it is possible … to become caught up …, carried away …, engrossed in it …. By this spontaneous involvement in the joint activity, the individual becomes an integral part of the situation, lodged in … and exposed to it, infusing himself [sic] into the encounter” (1961]: 38).

   In a revolutionary move in SLA/T studies, van Lier (1996) proposed that “language development is primarily a question of active engagement, not the transfer of linguistic items from one person to another” (67). He offered “warmth”, “connectedness”, “investment”, “intensity”, “socio-biological vigilance”, and “participatability” as partial synonyms for engagement, citing conversation as a paradigm context and quoting philosopher Merleau-Ponty: “In … conversation … my thoughts and his [sic] make up a single tissue, … insert[ing] themselves in a common operation of which neither of us is the sole creator. A double-being comes about, … we are … collaborators in … reciprocity, our perspectives glide one into the other” (168).

4. Engagement Processes (“Eye-to-eye ecological huddle … maximiz[ing] each participant’s opportunity to perceive the other participants’ monitoring of him [sic]” as “ratified through expressive signs”) – Goffman described the dynamic processes producing/sustaining engagement: Consciously or unconsciously, we continuously emit “expressive signs” of our ongoing alignment with our ecosocial environment, and humans are monumentally expressive, head to toe. Likewise, we continuously monitor each other’s expressive actions, monitor each other’s monitoring, monitor our monitoring being monitored, etc., leading to continuous mutual behavioral alignment. This occurs most intensively in FEs. An increasing focus on multimodality in SLA/T studies gives embodiment a role (e.g. Block 2014), but Goffman’s concept presents powerful new perspectives.

5. Intersubjectivity (“A ‘we rationale’ is likely to emerge, … a sense of the single thing that we are doing together”) – Mutual access to others’ expressiveness promotes intersubjectivity, widely considered crucial for social learning. For Goffman, intersubjectivity comprises individuals’ overlapping “definitions of the situation” – sociocognitive frameworks answering the crucial question: “What is … going on here?” (1974/1986: 8), thereby enabling a “working consensus” for cooperative action. In SLA/T studies, conversation analysts have studied intersubjectivity as the ongoing achievement of social organization via talk-in-interaction, often in tandem with embodied action.

6. Interaffectivity (“[FEs] provide the communication base for a circular flow of feeling among … participants”) – Although usually considered a cognitive phenomenon, intersubjectivity is also affective (Trevarthen 2006). For Goffman, the circulation of shared emotion – or ‘interaffectivity’ – is a primary feature of FEs. SLA/T scholars have likewise noted a possible “affective turn” in the field (Prior 2019).

7. Normativity (“[FEs] provide … corrective compensation for deviant acts”^[4]) – Goffman regarded social action as inherently normative, based on “ground rules of a restrictive and enabling kind” (1971]: x). As participants become engaged in FEs they present themselves as committed, agreeable, emotionally stable, mentally competent, etc. vis-à-vis the conventional norms of the activity engaged in. Such normative “presentations of self” (1959]) are continuously monitored and regulated by other participants through the engagement processes described above, resulting in convergent behavior: “The individual tends to modify this activity, … with its public character in mind”, yielding “an idiom of appearances and gestures that calls forth in the actor what it calls forth in … others” (1963]: 33). These presentations of self leave participants changed: Self/identity is constructed moment-by-moment in interaction for Goffman – we express who we are/are becoming through language use and embodied action – and is “accumulative” (Goodwin 2018). “Deviant acts” are substantially self-correcting via this process, but wayward participants can be nudged into line with a glance, intonation, etc. without explicitly disrupting interaction. The normative power of FEs thus operates both bottom-up and top-down as group members monitor, self-regulate, and align, thereby constructing social selves. Such normativity is central to learning/teaching, as specified below.

2.3 Extensions of FE concept

3.1 Alignment

Humans dynamically and sensitively adjust/adapt to their ever-changing ecosocial “surrounds” (Goffman 1981). Language and other communicative modes afford such alignment; language learning thus means learning how to coordinate adaptive action together – an ongoing organic, affective, relational process – rather than the private acquisition of linguistic systems. For van Lier (2004), “knowledge of language for a human is like knowledge of the jungle for an animal. The animal does not ‘have’ the jungle; it knows how to use and live in it. Perhaps we can say … that we do not ‘have’ or ‘possess’ language, but that we learn to use it and ‘live in it’” (253).

Language learning from this perspective is more about learning to (co)operate effectively in a human world/ecology – a “highly structured form of mutual fatefulness” (Goffman 1961: 35) – than internalizing grammatical code. The brain in this view is primarily a “relational organ” (Fuchs 2018), not an internal information-processing device.

3.2 Embodied multimodality

Our primordial form of human sociality – face-to-face interaction – is highly embodied and highly multimodal. For Goffman and his followers, expressive multimodal behavior connects cognition, embodiment, and social action because it indexes individuals’ cognitive perspectives, attitudes, values, and attentional and affective orientations vis-à-vis social others, who we mutually and continuously monitor and align with. Multimodal behavior also connects us to our material world (e.g. through pointing, gazing, and manipulating objects). In fact, from a sociocognitive viewpoint, “internal” thought, perception, affect, knowledge/competences, etc. are not separable from “external” behavior/action or the material environment; instead, they exist and co-constitute one another in an integrated “mindbodyword” ecology (Atkinson 2014). Multimodal action likewise impacts learning “because it occurs within an embodied participation framework that creates a visible, public locus for attention and action that includes both relevant structure in the environment and the actions and bodies of other participants. This is especially important in situations of apprenticeship and education” (Goodwin 2007: 59).

3.3 Natural pedagogy (NP)

Teaching-based learning is a powerful form of ecosocial adaptivity that is innate in humans (Csibra and Gergely 2009; O’Madagain and Tomasello 2022; Tomasello 2016). That is, as a species occupying almost every terrestrial habitat on earth, our main adaptive strategies have been cultural rather than genetic; e.g. Arctic peoples developed complex dwelling and clothing technologies to survive the cold rather than growing fur (Boyd et al. 2011). Yet such cultural survival strategies (1) can’t be transmitted genetically; (2) are usually technologically complex and therefore non-intuitive/“opaque”; and (3) therefore require novel forms of inculcation/learning. NP evolved to fill this gap. Atkinson and Shvidko (2019) reviewed NP research across disciplines vis-à-vis SLA/T, while Goodwin (2018) tied NP directly to his theory of co-operative action within EPFs, showing how the solution to the problem “of creating actors with specific relevant understanding of objects that are intrinsically opaque … emerge[s] systematically from the organization of co-operative action” (308).

3.4 Emotion/affect/value

Cognitivism adopts an engineering perspective: The brain is a data-processing machine. Yet humans are biological organisms, their brains are in their bodies, and their first priority as such is to keep their bodies viable in dynamic environments filled with affordances and threats. Valuing and affect are therefore at the heart of human behavior, perception, cognition, and learning/teaching (van Lier 1996). Furthermore, as organisms dependent on cooperative action for survival we are highly sensitive to interpersonal environments and relations; in Goffman’s words, “there seems to be no agent more effective than another person in bringing a world for oneself alive or, by a glance, a gesture, or a remark, shriveling up the reality in which one is lodged” (1961]: 41). Classrooms are affectively fraught in this sense because their very existence requires displays of incompetence and associated judgments, which are inherently face- and identity-threatening. A basic requirement for learning, as well as for who one is learning from/with, therefore seems to be that they have value for the learner.

We conclude this section by noting that our sociocognitive approach does not seek to replace other SLA/T theories, but to dialectically complement them and thereby broaden understanding of the enormously complex phenomenon of SLA/T. Cognitivism, in particular, is not simply wrong from a sociocognitive perspective; rather, like all theories of human action/behavior, it is incomplete.

4.1 Data collection

Data were collected via convenience sampling (Patton 2002): We searched the internet for videos of classroom interaction wherein participants’ multimodal expressive actions were maximally visible and audible. Of 21 videos found, we chose the video best meeting this criterion. One might critique this approach as a form of ‘cherry-picking’ – we found data we liked and used them to support our approach. This is true, but only trivially: A precondition for multimodal interaction analysis (see below) is that interactors’ use of embodied expressive tools is visible, audible, and therefore available for analysis; this was all we knew when we selected the data, as explained below. A second possible critique is that the video selected represents a single case, with nothing relevant to say about other cases. In contrast, we believe that anyone with any experience with classroom interaction will be able to ascertain familiar shapes and patterns of classroom interaction by merely inspecting the data. In fact, a nearly unique strong point of our data is their free and full accessibility to readers, unlike virtually any other classroom data set we know of. The video chosen is a 13-min YouTube video of an ESL lesson filmed in Australia, described further below and publicly available at: https://www.youtube.com/watch?v=DhCUVPkpLdk .

4.2 Data analysis

Our data were analyzed using multimodal interaction analysis (MMIA), as employed in our past research and detailed in Atkinson et al. (to appear).^[5] Goffman, arguably the conceptual originator of MMIA, mandated that “the study of every unit of social organization must … lead to an analysis of the interaction of its elements” (1961]: 7), focusing, in the case of FEs, on “the glances, gestures, positionings, and verbal statements that people continuously feed into the situation, whether intended or not. These are the external signs of orientation and involvement – states of mind and body ….beginning with the littlest, for example the fleeting facial move an individual can make in the game of expressing his [sic] alignment to what is happening” (1967]: 1). More specifically, we (1) focused on interactional episodes that appeared rich vis-à-vis Goffman’s description; (2) transcribed them using modified CA and multimodal transcription conventions;^[6] and (3) analyzed them in detail, integrating Goffman’s approach with our own past use of MMIA (Atkinson et al. to appear).

4.3 Contextualization: site, participants, and teaching approach

In order to contextualize our multimodal interaction analysis, we interviewed the teacher by Zoom.^[7] The teacher, John Bartik, told us that the class took place in March 2014 in a church in Sydney, Australia. The church offered ESL classes, and John was tasked with training teachers. He had previously completed a one-month CELTA (Certificate in Teaching English to Speakers of Other Languages) course, and taught at local language schools.

The students were 10 Mandarin-speaking residents of Australia, placed by their “beginner” level and L1 background. The class met weekly for 2 h, but John taught it just once, in order to: (1) model an active “communicative” teaching approach; and (2) test his personal theory of language learning/teaching:

Language acquisition is all about setting up scaffolded sentences and changing one thing at a time. And so … you’ve got “Do you like my [blank]”, and then just having that as my template and then substituting – it was all about substitute one word to build that deep construct. That’s what was going on in my mind ….You’ve got a stock sentence and you substitute, “Do you”? “Do I”? “Do you like my jacket?” …, making sure that [students are] good at substituting things in and out. (Interview, 8/10/20)

John’s actual teaching approximated this account: He used a single question form, “Do you like my ____?”, filling the blank with lexical items related to his person. The four resulting prompts in the video’s first 6:20 – when John was sole prompter/question-asker – were “Do you like my” (1) “jacket”; (2) “blazer”; (3) “haircut”; and (4) “breath”.^[8] After introducing each prompt, John (1) elicited repetition and provided feedback in full-class mode, and (2) prompted individual students, providing feedback on grammar and pronunciation. He sometimes did so in a dramatized way, which elicited much laughter.

5.1 Initial elements in classroom FE

From 00:50 to 6:20 on the video counter, the classroom is configured as in Figure 1: Students (Ss) stand in a rough semi-circle around teacher (T).

Figure 1:

Classroom configuration from 0:50 to 6:20 of video; “S1” = Student 1, etc.; “T” = Teacher.

5.1.1 Joint attention

Goffman (1961] and above) identified “visual and cognitive” attention-sharing as key to establishing/maintaining FEs. Here we first use Goffman’s “visual and cognitive” formulation in our analysis before complicating it.

Once the initial participation framework emerges around 00:50 (see Section 5.1.2), Ss evince joint visual attention, gazing actively at T when he speaks, or at fellow Ss individually prompted by T. Figure 2 shows Ss’ shared gaze-following as the FE progresses: In 2A, T asks S10 “Do you like my blazer?”; in 2B, T prompts S1 to repeat this question; and in 2C, T playfully responds to S8’s response to a different question. In all cases, Ss attend either T or fellow Ss prompted by him closely.

Figure 2:

Joint attention.

Visual attention is also a proxy for cognitive attention (McMains and Kaster 2009), the latter further evidenced by the “conditional relevance” (ten Have 2007) of Ss’ responses. Thus, T’s initial query, “Do you like my jacket?”, receives three immediate “yes” responses, one apparent question confirmation, and “I like”, “Yeah”, and “(It’s) beautiful” simultaneously or in tight consecutive order, the latter three responses accompanied by pointing gestures. Ss thus show that they are cognitively attending T’s question by responding appropriately, and jointly construct the order and form of their responses.

In fact, Ss don’t just attend visually and cognitively but with their whole bodies (Figure 2), evincing a shared “ethos, … spirit, … [or] emotional structure” (Goffman 1963: 19) of embodied attention. Thus, they stand with hands clasped in front of them, at their sides, behind their backs, or arms folded; their gaze and facial expressions indicate attentiveness; and their bodies and language generally align with T and fellow Ss. This style of attention might reasonably be called “serious, alert attention,” and is highly affective: “A heightened sense of moral responsibility seems to develop” (89). Affectivity is also evidenced in the group’s sense of humor/play (e.g. Figure 2C).

5.2 Embodied participation framework

By 00:50, the EPF shown in Figure 1 has formed, “creat[ing] a visible, public locus for attention and action that … is especially important [for] apprenticeship and education” (Goodwin 2007: 59 and above) and providing at least five learning/teaching affordances. First, the semi-circular arrangement gives Ss maximal access to each other’s embodied action, enabling joint attention, mutual monitoring, and engagement. Second, standing provides greater interactional mobility and flexibility than sitting. Third, height differences between T and Ss – an ethological component of power/dominance (Stulp et al. 2015) – are reduced, thereby promoting affiliation. Fourth, the semi-circular linear arrangement affords a queue-like organization (Hutchins 2005) providing notable learning/teaching affordances: (a) predictable order – on introducing new prompts, T first addresses the full class, then moves up- or down-line prompting individual Ss in roughly consecutive order, thereby indicating what and where social action is likely next and therefore “how to go on” (Wittgenstein 1958); (b) Flexible sizing and shaping of “participation units” – T can address/respond to Ss whole-group, individually, in clusters, or Ss can organize themselves (see Excerpts 1 and 3); and (c) The queue is used to mark episode boundaries – T usually ends a prompting cycle around queue’s end, thus further structuring the learning environment. Fifth and finally, the EPF features linguistic routines, thereby further constraining the sociocognitive action-space by enabling predictable language behavior. T thus begins each prompting cycle with a “Do you like my ____?” question, which Ss are initially guided to answer positively. One overall effect of this EPF is to afford a stable substrate (Goodwin 2013) of social organization on which change – learning – can occur.

5.3 FE-enabled learning: an introductory example

While describing their features discretely, Goffman (e.g. 1964]) treated FEs as integrated interactional microecologies. Here we provide an introductory example from our data, showing how different elements in Goffman’s definition function together to afford learning through focused interaction.

As noted above, T’s initial prompt was “Do you like my jacket?” T subsequently substitutes “blazer” for “jacket”, models “blazer” for choral repetition, then starts a new prompting round with “Do you like my blazer?” After first eliciting choral response, T models both question and answer, and Ss repeat. T then moves to near-end of line and engages S10, first prompting repetition of the full question, then focusing on pronouncing “blazer”. S10’s responses are mostly inaudible, but S6 is hearable responding “I like your bla/dʒ/er”, then repeating “bla/dʒ/er” twice as T models “blazer” for S10. At this point S7, who has been observing T intently, intervenes. Excerpt 1 and Figure 3 capture the action.

Figure 3:

Pronouncing “blazer”.

First, consider the role of joint attention and engagement/mutual monitoring as Excerpt 1 commences. Preceding the excerpt, T had prompted S10 with “Do you like my blazer?”, then focused joint attention on pronouncing “blazer”. Ss aligned with T by experimentally pronouncing “blazer”, repeating it at least six times in the 8 s preceding Excerpt 1, five coming from either S5 or S6 in close consecutive order. In doing so, Ss 5 and 6 offer what may be competing candidates for the most teacher-like pronunciation, S5’s “blazer” being inaudible and S6 saying “bla/dʒ/er”. Thus, in addition to maintaining joint attention with T, S5 and S6 appear to mutually monitor each other’s pronunciation, and may be producing competing, norm-seeking pronunciations of “blazer” on that basis.

Second, mutual monitoring and norm-seeking appear explicitly when S7 turns to S6 and presents her own pronunciation of “blazer”, apparently repeating it twice (lines 2–4) with gestures. The normative (i.e. teacher-ratified) focus of joint attention – the accurate pronunciation of “blazer” – is thereby amplified/supercharged by mutual monitoring and action, revealing a substantially more complex “mutual entanglement” (Clark 2022) than found in much classroom interaction research, e.g. instead of a top-down imposition of Initiation–Response–Feedback (IRF), we see a complex, patterned microecology of interaction.

Third, consider the social action afforded by the EPF in Excerpt 1. Following S7’s intervention, S6 momentarily shifts attention to S7, pauses, and repeats “bla/dʒ/er” (6). S5 then joins in, affectively upgrading S7’s action by stepping forward, turning toward S6, flashing eyebrows, and demonstratively responding “BLA/s/ER. Bla/ʒ/er”, marked verbally by increased volume and repetition (7–9). Neither S7’s nor S5’s actions are predictable/conventionalized parts of the pedagogical action/EPF; they are nonetheless afforded by it.

Fourth, consider the two final points of Goffman’s (1961] and above) definition: “[FEs] provide the communication base for a circular flow of feeling among … participants, as well as corrective compensation for deviant acts” (18). First, feeling/affect apparently circulates between S7 and S5, who are physically closest to S6 due to the EPF’s configuration, and therefore directly privy to her behavior. They appear to feel that S6 needs to be corrected, or at least made aware of the targetlike form: S7 begins by pronouncing “bla/z/er” quietly but insistently and then repeating it, accompanied by beat gestures while turning toward S6. Following S6’s reiteration of “bla/dʒ/er”, S5 intensifies the shared action with body movement, facial expression, and raised volume. In this sense, Ss5-7 constitute a temporary participation unit (Goffman 1981) within the larger participation structure, although S6 neither acknowledges nor responds to S5.

Second, it seems clear that Ss7 and 5 are attempting to norm S6’s pronunciation of “blazer”. In Goffman’s (1971) terms, they collectively “sanction” S6’s pronunciation, first mildly then more demonstratively. From another perspective, however, this “sanctioning” may constitute cooperative co-teaching: Ss7 and 5 try to help S6 produce a valued norm. Goffman’s (1967) notion of face may apply here – participants actively protect one another’s positive image/identity – or face – in order to preserve group functionality.

5.4 Extended example of FEs in action

5.4.1 Context

T introduces a new prompt at 2:22: He smiles, takes three steps along the student line while sweeping it with his gaze (Figure 4A), stops and leans in (4B), then points to his right temple while flashing eyebrows, smiling, and intoning rather quietly, “Do you like my haircut?” (4C). After a one-second pause, Ss respond with laughter. T then asks S9 or 10 (off-camera) the same question with raised pitch (4D), and, following another one-second pause, asks S8 the same question with further pitch rise (4E). Following yet a third one-second pause, S8 answers “Yeah” while nodding (4F), to which T responds “YES” with a smile and quick palm-up gesture (4G), turns left, takes a step, puts RH to forehead in a theatrical “whew” gesture while exclaiming “Whe:w” (4H), turns away from Ss and then back, grinning broadly. Ss laugh enthusiastically (4I).

Figure 4:

Initial set-up and prompting of “Do you like my haircut?”

5.4.2 Learning action

Excerpts 2–3 and Figures 5 and 6 present learning action occurring simultaneously with the foregoing. Here, as T begins prompting Ss with “Do you like my haircut?”, S6 undertakes a quest for the meaning of “haircut”. What ensues illustrates the inter-individual, multimodal, natural pedagogical character of learning/teaching within FEs.

Figure 5:

“Cut” + embodied action.

Excerpt 2

(2:33–2:45): “Cut. Cut. Cut.” (Marginal numbers mark S6’s repetitions of “cut”)
	1	T:	Do you like my haircut?	Figure 5A
	2	S8:	(1.0) Yeah.
	3		[
1–>	4	S6	oCuto ((gazing at T))	Figure 5B
	5	T:	((Gazing at S8)) YES. (.9)
	6		Whe:w
	7		[
2–>	8	S6:	oCuto ((after tilting head L while	Figure 5C
	9		gazing intently at T))
	10	Ss:	((Laugh enthusiastically, except S6))
3–>	11	S6:	((Visibly utters “cut”, still gazing at T.	Figure 5D
	12		Turns L to Ss7&8 with RH extended,
4–>	13		establishes eye contact, then visibly repeats	Figure 5E
	14		“cut” with downward beat gesture))
	15		(.8)
	16	S7:	((Puts RH on head)) Haircut	Figure 5F
	17	S6&8:	((Put hands to head simultaneously &	Figure 5G
	18		move fingers. S6 lowers hand & refocuses
	19		gaze on T))
	20	S7:	(Jiǎn de). Jiǎn tóu fa.[Cut. Haircut.]

Excerpt 2 begins with T asking S8, “Do you like my haircut?” (line 1; Figure 5A). As S8 replies, S6 utters “cut” in a quiet, somewhat breathy voice while gazing at T, apparently indicating non-understanding of “haircut” (4; 5B). S6 then utters “cut” twice more in T’s direction, the second visible but inaudible due to group laughter (8 & 11; 5 C & D). Instead of “letting it pass” (Firth 1996) and co-laughing, however, S6 initiates a “side-involvement” (Goffman 1963) word search, turning to Ss7 and 8 while extending RH, securing their attention, and visibly repeating “cut”, punctuated by a downward beat gesture (12–14; 5E). In response, S7 quickly raises RH to head (16; 5F) and utters “haircut”, followed immediately by Ss6 and 8 raising hands to heads and moving fingers there for about 1 s (17–18; 5G). S6 then shifts her gaze/attention back to T, and S7 utters (Jiǎn de). Jiǎn tóu fa. [Cut. Haircut.] (20).

When T comes back down-line 2.7 s later, S6, who was smiling before commencing her word search (Figure 4D), is smiling again. T questions S7, then moves to Ss9 and 10 (off-screen), whom he prompts extensively, supplemented by translation help from Ss7 and 8. Excerpt 3 and Figure 6 show what follows.

Figure 6:

“I don’t like your haircut”.

Excerpt 3

(3:13–3:17): “I don’t like your haircut” (Arrows mark S6’s responses)
	1	T:	((Moving up-line, gazing toward Ss1 & 2))	Figure 6A
	2	S6	((Leans forward, tilts head, with serious
–>	3		face)) I don’t, I don’t like (.3) oyour
	4		haircut.o ((Raises L index finger on “your”))
	5	T:	((Stops at “I don’t”, shifts gaze to S6 while
	6		putting hands together, leans forward, flashes
	7		eyebrows, & points)) Good. ((Steps back as
	8		Ss laugh)) (1.5) Say I don’t	Figure 6B
	9		[
–>	10	S6:	(Very long) ((points at T
	11		with R index finger on “very”, then puts RH
	12		to head on “long” & laughs))
	13	T:	I don’t like your haircut. ((Extends RH &	Figure 6C
	14		gestures on syllable beats))
–>	15	S6:	((Smiles broadly & nods)) I don’t like (.5)	Figure 6D
	16		your haircut.
	17	T:	((Continues extending RH)) >Good good<.

As T reverses direction and starts up-line, S6 leans forward, tilts head back, and declares, “I don’t, I don’t like (0.3) your haircut” (3; Figure 6A). As S6 speaks, T pauses mid-stride, shifts gaze to her, and puts hands together (5–6). He then moves closer, leans forward, points briefly with R index finger, responds “Good”, then pauses a second-and-a-half, steps back, leans forward again, points, and says “Say I don’t” (8; 6B). S6 overlaps T’s speech with “(very long)”, apparently justifying her negative assessment, and laughs (10–12; 6B). T then restarts and finishes his utterance, “I don’t like your haircut” (13; 6C), which S6 repeats while smiling broadly and nodding (15–16; 6D). T responds, “Good good” (17) and moves on.

To summarize and interpret this extended example of FE-based learning/teaching from a sociocognitive perspective: T introduces a major new semiotic object, “haircut”, into the focused learning/teaching encounter, which S6 doesn’t understand. She therefore takes adaptive action by initiating an environmental search. Her first stop is T himself–S6 utters “cut” three times in T’s direction, receiving no response. She then searches elsewhere in the environment, enlisting Ss7 and 8 in her search. This approach works: S7 answers verbally (“Haircut”) and gesturally, Ss6 and 8 repeat and elaborate the gesture, and S7 adds Chinese translation. What has transpired so far, then, is rich, complex, and ecologically distributed learning/teaching action within an FE: Following a “we rationale” (Goffman 1961), Ss and T operate as an integrated organism.

The action continues: Rather than simply storing the new word in her head, S6 redistributes it in the environment. As she does so she transforms it, creatively embedding the new object within meaning objects already accessible to her – “I don’t like your haircut” – thereby undertaking co-operative action. This utterance is further noteworthy for what it adds to the learning/teaching event: S6’s contribution marks the first appearance of negation in the class – a powerful tool for the activity they are engaged in – answering “Do you like my ____?” questions/opinion-sharing. S6 therefore increases the class’s learning potential significantly, which T and other Ss actualize by incorporating negation into the lesson: The next three teacher-student exchanges (not included here) feature negative answers. In sum, the total teaching/learning action is highly “inter-” – ecological, distributed, intercorporeal, and interaffective; it is thus the product of dense sociocognitive alignment, or, in Goffman’s (1961) terms, a focused encounter. It is furthermore a natural pedagogical encounter, because it is grounded in intrinsic skills of human interaction: gaze, gesture, naming/labeling, repetition, body movement, etc.

6.1 Joint attention and action

Schmidt (2010) influentially defined attention for SLA/T studies as a set of internal cognitive mechanisms “controlling information processing and behavior when existing skills and routines are inadequate” (no page numbers), hypothesizing that SLA requires attention to occur. Following Schmidt, a major goal in instructed SLA studies is to direct conscious attention to linguistic form, which may temporarily preempt meaning-based communication (Long 2015).

In our view, however, attention is an ecological/sociocognitive relation: Organisms naturally attend to what their environment affords their adaptivity and survival (Gibson 1979). For humans, this regularly means attending with and to others, because our main survival strategy is acting adaptively in groups: Joint attention is essential for cooperative action.

In FEs, joint attention further includes mutual monitoring – participants monitor each other’s actions in order to co-align. Such attention may be intensely focused, but occurs naturally without requiring highly conscious behavior: “The individual … can become unthinkingly and impulsively immersed … Conjoint spontaneous involvement is a unio mystico, a socialized trance” (Goffman 1957: 47). Accommodation, interpersonal synchrony, co-construction, intersubjectivity, and intercorporeality are other concepts indicating how we naturally co-orient/align. In this sense, attention represents ongoing ecological sensitivity and connectedness, not an internal cognitive mechanism requiring special stimulation to occur.

FE-based joint attention may affect SLA/T as follows. First, it may require no special cognitive effort or artificial techniques, only “a sense of the … thing … we are doing together – a ‘we rationale’” (Goffman 1961: 18, quoted above). Second, per Goffman’s just-quoted words, joint attention guides joint action, including learning. Third, it is highly embodied, as shown in Figure 2, where students track their teacher with their whole bodies. Fourth, it is highly affective: The students in our data adopt a shared ethos of serious attentiveness along with a sense of play/humor. Fifth, it is guided by the ecological “surround” (see Section 6.2 below).

Ingold (2001) and Goodwin (2007) treat learning as “the education of attention” for adaptive ecosocial action. Learners don’t internalize representations – cognitive knowledge structures replicating something in the world – but learn through guided experience to act in/on their environment in adaptive ways. What else could brains-in-bodies-in-the-world be for than responsively acting in/on worldly environments?

6.2 Rich, structured, multimodal environment

For cognitive philosopher Clark, being human involves “radical entanglement” (2022]) of mind, body, and “a complexly structured environment apt for engagement and action” (2016]: xv). For Goffman (1964), FEs are “microecological”: faces, hands, utterances, bodies, material settings, and social practices intertwine in “ecological huddles”. Goodwin’s embodied participation framework significantly develops Goffman’s vision, including how it contributes to learning.

The learning affordances of embodied participation frameworks are well-displayed in our data: Students’ semi-circular standing arrangement allows for both maximal interactional connectivity and specific pedagogical action/response. Through the teacher’s prompting of individual students in queue-like order, students know roughly when their turn is, and can use classmates’ prior responses to calibrate their own. This participation framework also affords emergent participation units of smaller clusters of students, whereby they co-teach and co-learn, e.g. their distributed engagement around the word “haircut”.

SLA studies typically views learning environments thinly – as indirectly influencing or “socially gating” internal acquisition processes (Atkinson 2013). Even conversation analysis, a major social perspective on SLA/T, views interaction as produced by a semi-autonomous “machinery” based on sequential organization, leading Goodwin to critique its “punctual, categorical view of action and actors”, versus “an interactive field … pulling upon distributed resources” (Goodwin and Salomon 2019: no page numbers), while Deppermann and Streeck (2018) recommended refocusing conversation-analytic studies on “simultaneity, multi-party settings, the materiality of multimodal resources, multi-activity, objects, space, and movement” (7).

Viewed ecologically, the world is overflowing with “information”, which we use to continuously orient and act in our world, including through language. Contra many SLA approaches, ecological information is highly affective and value-laden, based on “a defining feature of life, … responsivity” (Sheets-Johnstone 2009: 376). Far from being input processors or interaction machines, real learners are highly responsive humans-in-(e)motion.

6.3 We-process

As quoted previously, Hanks (1991) viewed learning as occurring “in a participation framework, not … an individual mind … Learning is … distributed among coparticipants” (15). This view hardly denies that individual minds, or – better – “mindful bodies” (Sheets-Johnstone 2011), exist, but treats them as fundamentally interconnected and ecologically embedded. Language is a tool for coordinating ecosocial action, which essentially shapes its learning; language learning is therefore a public/relational “we-process” (Atkinson et al. to appear), here as afforded by embodied participation frameworks in FEs. Classrooms are key sites for such learning, although typically viewed as promoting individual knowledge acquisition as measured by individual assessment.

The teacher curates the learning process in our data, but it is solidly grounded in basic processes of social interaction. That is, although novices in English, the students are experts in social interaction. Thus, in Excerpt 1 students produce differing phonological manifestations of “blazer”, but they do so using familiar interactional tools: gaze, volume, repetition, gestures, intonation, movement, facial expressions, etc. Therefore, although discourse/conversation analysts have typically treated classroom interaction substantially sui generis, it is grounded in basic building blocks of social interaction.

6.4 Alignment

Most generally, alignment is organisms’ ongoing responsiveness to their environment. More specifically, alignment is ecosocial – human ecologies are socially constructed, organized, and maintained. More specifically still, alignment is micro-interactional: Individuals conduct focused social action in/through embodied participation frameworks – dynamic micro-niches enabling responsive adaptation via multimodal we-processes.

Alignment has been variously defined in human studies. Conversation analysts, for instance, define it as “the structural level of cooperation”, versus affiliation – “the affective level of cooperation” (Hall and Looney 2019 for SLA/T; Stivers et al. 2011: 20–21). For sociocognitivists alignment and affiliation are inseparable: Signaling cooperation/interaction automatically entails taking an affiliative/affective stance. We align with what is valuable, and therefore meaningful, in order to survive: Feeling and doing are mutually constituted in ecosocial engagement.

A key concept in cognitivist SLA/T studies is feedback – “correction of production” (Behney and Gass 2021) of learners’ formally inaccurate speech/writing, typically by more proficient language users. Feedback is considered necessary to promote conscious attention to form, leading potentially to cognitive restructuring, leading ultimately to accurate output.

Yet feedback can be construed much more broadly, as basically every environmental happening organisms are responsive to. Viewed thusly, feedback represents the “differences that make a difference” in an ecosocial feedback loop (Bateson 1972). Goffman (1981) envisioned effective language use as adaptive responsiveness in presenting self to others, moment to moment, in order to maintain cooperative relations, versus externalizing internalized grammar: “What the speaker is engaged in … is to meet whatever occurs [in the interaction] by sustaining or changing footing …, select[ing] the footing which provides … the least self-threatening position in the circumstance, … the most defensible alignment” (323). How such interactional responsivity may motivate language learning is discussed in Section 6.6.

6.5 Variation

Modern L1A and SLA/T studies were largely founded on the insight that variation is essential for learning, i.e. errors are hypothesis-testing learning tools (Corder 1967). Contemporary SLA/T approaches increasingly give variation a significant role (Larsen-Freeman 2020; MacIntyre 2020), viewing each occurrence of linguistic form/function as a learning increment: “In every recurrence there is a compromise of iteration and innovation” (Brown 2005: 27). L2 teachers have likewise long known that error is essential in language learning; as veteran teacher educator/researcher Reppen (2021) stated: “We want our students to make errors …. If they aren’t … they’re in the wrong level, … not having to push themselves, … not having encounters with communicative, purposeful, meaningful language”.

The importance of variation, then, is well-established in SLA/T studies. Yet variation-as-we-process – variation occurring in/for group learning – has not been investigated to our knowledge. In our data, when S6 produces “bla/dʒ/er”, S7 responds “bla/z/er”, and S5 responds “BLA/s/ER. bla/ʒ/er”, along with gaze, gestures, body movement, and facial expression in an embodied participation framework, the students are learning together, not (just) in their heads.

6.6 Normativity, identity, self

FE participants monitor each other’s embodied expressivity, interpret it as expressing alignment to the group, and respond to it actively in co-regulating group action. For Goffman, this process is norm-driven and identity-driven: By presenting themselves positively vis-à-vis norms of social interaction, participants qualify as eligible interactors – crucial for a species dependent on group survival. General norms for FEs include presenting attentive, competent, prosocial selves/identities, while specific norms in language classrooms may include attentive responsiveness to pedagogical action, pursuit of accurate and fluent language behavior, openness to feedback/correction, and consideration/support for fellow students.

For Goffman, selves are relational: Who we are is based on how we present ourselves socially, and how we present ourselves socially is based on ideal, socially normative versions of self. However, Goffman didn’t view people as rote rule-followers, but rather as responding sensitively and continuously to normative values and practices, including resistantly and partially, in dynamic moment-to-moment interaction.

Goffman’s view resonates with those of major theorists of language and society, including Gee (2015): “Being a who (identity) engaged in a what (activity) is a performance …, visible in the world and situated within ongoing, but changing social conventions. If there are no conventions (social agreements and alignments) there is no understanding to be had, no matter what is in your head” (302), and Bakhtin, “To be means to be for the other …. Man [sic] has no internal sovereign territory, he is all and always on the boundary; looking within himself, he looks … through the eyes of the other …. I must find myself in the other, finding the other in me” (In Wertsch 1998: 116).

For Goffman, Gee and others, identities are continuously formed, negotiated, and transformed in interaction: If interaction is “the reciprocal influence of individuals upon one another’s actions when in one another’s … presence” (Goffman 1959: 15 and above), then we are always being shaped in non-trivial ways. This is patently true in child socialization, wherein children’s identities are continuously being formed, but, as Hopper (1998, quoted previously) highlighted, socialization is lifelong. This shaping is typically norm-convergent – we try to meet the expectations embodied in whatever group/s we identify with.

One major form of norm–convergent shaping is learning. During interaction, as interactors continuously shape each other’s actions according to norms/targets, new social action emerges. Thus, in Excerpts 2 and 3, S6 responsively enacts a self which is both norm-adherent – e.g. in her accurate learning of the form/meaning of “haircut” and use of negation in “I don’t like your haircut” – and norm-innovating – she introduces a consequential new tool into the class, which it then actively takes up — negation functioning in response to the teacher’s basic question: “Do you like my ____?” Significantly, S6 also does so with humor, thereby honoring the sense of play that is a local classroom norm. For Goffman (1971), norms are of both “a restrictive and enabling kind” (x), and in this case, when the class undertakes group learning of a powerful new means of meaningful response, it is enabling. Evidence of enabling norms may also be found in our data when S7 and S5 show regard for S6’s positive face in Excerpt 1. In these ways and others, normativity drives learning.

6.7 Learning as ecosocial homeostasis

In biology, homeostasis is the feedback process whereby bodies maintain viability in a dynamic world: Heartbeat, respiration, hunger, size of iris, etc. automatically adapt to environmental change due to continuous, highly sensitive body–environment feedback loops. Without homeostasis, we couldn’t survive.

When new, potentially useful phenomena appear in our environment, we adapt if possible because we need to be functioning parts of it. As widely acknowledged in the learning sciences, much learning is implicit and incidental, occurring naturally through exposure and experience. Implicit learning can be viewed as adaptive in nature; thus, in the case of language learning, “children do not seem to learn sentences, but … to adapt their behavior to increasingly complex surroundings. It is an enterprise that … continues throughout a lifetime” (Hopper 1998: 161–162 and above.).

Yet simple adaptivity may go only so far – many of the complex technological traditions innovated by humans over generations require more. Humans have therefore evolved an additional form of adaptivity critical to our survival – the ability to adapt culturally, as described in Section 3.3. One powerful form of cultural adaptivity is natural pedagogy – a supercharged ability to convey, receive, and in so doing develop and transform complex cultural solutions to human environmental adaptation problems. A few other species teach (by certain definitions of teaching), but no other species depends so radically on inculcating cumulative traditions of intelligent action for continuous development, refinement, and transformation.

Learning is therefore homeostatic in that when our environments change, we naturally change/adapt with them. Part of this homeostasis, however, is based on a human superpower, which is an additional reason why human learning – and teaching – “is not a one-person act” (Hanks 1991: 15): We learn sociocultural practices together in multiperson settings from more skilled others of all types, including but not limited to classroom teachers.