Sequence Initiation or Self-Talk? Commenting on the Surroundings While Mucking out a Sheep Stable

ABSTRACT This study investigates comments on the physical surroundings while a group of people are shoveling dung in a sheep stable. In this setting, where the auditory space is mostly open for talk, some comments launch a conversational sequence, while others are treated as self-talk. The article discusses how the speaker’s body posture, speech volume, and gaze, as well as the nature of the referent, contribute to attracting a response. Comments treated as self-talk are typically uttered with low volume, while the speaker is bending forwards with his or her gaze toward the ground. Comments that launch a sequence and achieve a focused interaction are instead spoken out loud, with the speaker’s body oriented toward the other participants, and deploy the recipients’ current attention focus. Furthermore, the timing of the comment just before an upcoming activity junction from shoveling to a brief rest increases the chances of a conversational sequence being developed. The data are in Estonian with an English translation.

nonlinguistic doing" may "occasionally speak (their) passing thoughts aloud," thus discussing selftalk in multiactivity. Language use and embodied behavior has been studied in different professional settings where the speakers engage in multiple activities, such as teaching and archaeological excavation, driving and talking, teaching and performing surgery (e.g., Goodwin, 2000;Mondada, 2011Mondada, , 2012Mondada, , 2014, as well as mundane interaction (Haddington, Keisanen, Mondada, & Nevile, 2014). Furthermore, it has been demonstrated how a speaker can display his multiple engagement by posturing his body in a twist (Schegloff, 1998), and a teacher may control a student through touch while engaging in a discussion with other students (Cekaite, 2015). Apparently, people can also deploy (parts of) their body to progress an alternative course of action while continuing their verbal engagement unimpeded (Raymond & Lerner, 2014). In the current sheep stable setting, the essentially embodied work task and the talk do not necessarily impede each other. However, talk is also produced by the body, which can be problematic when engaged in arduous activities. While most interaction analytic work considers talk running in parallel with activities that are not particularly strenuous for the body, such as watching TV, studying, drinking, or driving (e.g., Hoey, 2015;Mushin & Gardner, 2009;Szymanski, 1999), the current article focuses on a parallel activity that not only smells bad but is physically taxing. The article thus also begins to explore how talk is organized in settings where participants are legitimately and intensely preoccupied with a heavy physical task.
Conversational openings are methodologically and jointly achieved (Schegloff, 1968), and turns are generally produced with no gap and no overlap (Sacks, Schegloff, & Jefferson, 1974), so that silence between turns is minimized (Stivers et al., 2009 on question-answer sequences). There are, however, situations of "open state of talk" (Goffman, 1981, p. 134) or "incipient talk" (Schegloff, 2007), when people are co-present and speakers move between engaging in and disengaging from conversation. The organization of these openings has been studied considerably less. For example, Goffman (1981, pp. 74, 93-94) argues that a coparticipant may or may not react to someone speaking "his passing thoughts aloud," even though its subject matter may be informative. Szymanski (1999) has shown how questions, noticings, and outlouds can be deployed as methods of reengagement during group work at school. Hoey (2015) has analyzed the systematic arrival at and organization of lapses. Among other things, he discusses settings where participants are oriented to the optionality of talk, such as when watching TV, where at the occurrence of silence the speakers just go on doing what they were already doing (Hoey, 2015, pp. 437-442). The current study that likewise features a setting of optional talk builds on these findings and scrutinizes the embodied delivery of a particular action, a comment on the physical surroundings, and its potential to initiate a conversational sequence.
The earliest interactional literature on comments triggered by observable features of someone's surroundings builds on two concrete utterances: "Hey you have a hole in your shoe" (Sacks, 1992 II, pp. 87-97) and "You didn't get an ice-cream sandwich" (Schegloff, 1988, pp. 119-131). These two authors call them noticings. In addition to selecting a single aspect of the complex passing world to comment on, noticings are characterized by their timing as (a) interruptive of current talk (Sacks), or (b) sequentially positioned after conversational openings (Schegloff). As it happens, the authors only discuss noticings that are more or less intimately connected to, or even owned by, the recipient and thus make relevant a response by that person. Importantly, however, Sacks and Schegloff talked about noticeable matters as a choice among an infinite number of locally contingent options, including some that must have taken a longer period to develop or be developed, such as a hole in the shoe (Sacks, 1992 II, pp. 87-97) or a new arrangement of living quarters (Schegloff, 2007, p. 86). In contrast to this, recent literature on noticings (Kääntä, 2014;Keisanen, 2012;Kidwell, 2009;Laanesoo & Keevallik, 2017) has chosen to focus on the suddenness of the event being commented on. This literature has dealt with verbal as well as embodied noticings of abrupt events, such as when a small child involved in a harassment episode suddenly sees a caretaker (Kidwell, 2009), when a driver just failed to make a turn (Keisanen, 2012), when a student immediately redirects her gaze after a teacher's mistake (Kääntä, 2014), or when a parent or a dance teacher notices a breach (Laanesoo & Keevallik, 2017). The current study embraces comments on both abrupt and longitudinal matters.
More broadly, comments on the physical surroundings constitute instances of local sensitivity in conversation, as described by Bergmann (1990). "Local sensitivity is meant to capture the tendency built into every topic talk [i.e., talk on a specific topic-LK] to focus on elements of the encounter's context which are situated or occur in the participants' field of perception but have not been topicalized so far" (Bergmann, 1990, p. 207). Interestingly, Szymanski (1999, p. 6) shows that during group work in a classroom noticings require their recipients to look somewhere or do something in order to verify the claim being made or to resolve the issue being raised, thus attracting an (embodied) reaction. Locally sensitive utterances thus at least potentially create a shared experience of some event (Bergmann, 1990, p. 215). Among other things, Bergmann argues that topicalizing local matters is a prime method of reopening a stalled conversation.
The current article scrutinizes the initiating capacity of locally sensitive utterances by focusing on a subset, namely those topicalizing the material environment that is mutually perceivable by the participants. In this regard they are similar to noticings and different from online commentary (Heritage & Stivers, 1999), for example, where only one participant may have visual access to the source. In contrast to at least some delimitations on noticings (Keisanen, 2012), comments here feature declarative grammar and concern matters that may or may not have occurred abruptly. They are not tied to specific positions in a sequence or an encounter, the referents are nobody's personal property, and they are thus not subject to any kind of preference (compare to Pillet-Shore, 2017, on registerings at the beginning of encounters). In contrast to announcements, comments on the physical surroundings do not concern personal news or other events that have occurred beyond the here and now (see also Pillet-Shore, 2017, in regard to various similarities between noticings and announcements). Aligning with Bergmann (1990), this study will be talking about comments because a comment is more neutral than a noticing, a registering, or an announcement with respect to its position, recipient-directedness, and suddenness.
As will be shown, comments on the surroundings need not always attract a response. This is in line with the argument by Stivers and Rossano (2010) that noticings together with announcements and assessments form a group of actions that have a low response relevance. They find that "these less canonical actions do not, via their initial sequential position and action alone, normatively require response" (Stivers & Rossano, 2010, p. 27), but response relevance can be increased by interrogative grammar and interrogative prosody. Neither of the latter occurs in the case of comments on the physical surroundings, which should accordingly have very low response relevance according to the model. In addition to (lack of) response, the analysis here focuses on the embodied delivery (body orientation and loudness), the timing of the comments in relation to ongoing activity trajectories and others' current attention, as well as the surprise value of the object commented on. Along with the discussion of Stivers and Rossano's study by Couper-Kuhlen (2010), the argument here is that in an essentially nonfocused interactional setting certain additional aspects of the speaker's embodied behavior and the exact timing of the utterance can actually contribute to attracting a response and achieving a focused interaction. The article thus expands on the recent "embodied turn" in studies of interaction (Nevile, 2015). By analyzing the vocal as well as bodily behavior of the participants, we will be targeting the hitherto unexplored difference between a locally sensitive comment that elicits further verbal actions as opposed to one that is treated as self-talk.
The data, the collection, and the method The study is based on a recording of a single whole-day event where 11 young people have to manually muck out a sheep stable. According to an age-old Estonian farming tradition, they are offered food and weekend accommodation for helping out with the task, but they are not hired. One of the participants, the host, has cleared the stable beforehand and is therefore occasionally treated as an expert. All the others apparently lack prior experience of the job. At their disposal they have two wheelbarrows and a number of forks and spades. Without any prior plan, they switch tools and the related assignments during the day. The video recording comprises 5.5 hours filmed inside the stable. The recording was made with a single camera on a tripod, carried around the space in relation to the workers. Even though a single recording angle occasionally sets analytical limits, care was taken to have the largest possible number of people in view at all times. All the participants have consented to the materials being used for research, and their names have been replaced in the transcripts.
The general impression of these data is that many utterances are not responded to. Besides announcements, assessments, noticings, and other comments, responseless actions include swearing, turns addressed to birds and animals, announcements about speaker's own actions, and general instructions about the organization of work, but even wisecracks may be flagrantly ignored. Comments constitute a type of action that may or may not be related to the work task. The dung in particular constitutes an object of wonder and disgust, being a novelty to most of the workers but also an obstacle to overcome before the pleasant evening activities can begin. Its texture and weight, as well as the depth of the dung layer, are all relevant factors that can be recalculated into time units for the remaining work.
The collection for this study includes 48 comments on the physical surroundings. The criteria of selection were sequential, topical, and turn-constructional. The utterances could not be produced in a responsive sequential position, and thus they cut into silence or disrupt the ongoing conversation. They had to be formulated as if based on a perceptual inspection and express a proposition about something that exists or happens in the context. Furthermore, they could not be immediately preceded or followed by a question or a directive by the same speaker because this would have created a different level of response relevance. A typical boundary case between comments and directives is prefaced by the verb "look," as in vaadake kus pojad on seal ("look, there are chicks there"). This particular utterance attracted almost everybody's gaze and launched an extended discussion on the well-being of the chicks in their nest. These kinds of directive-noticing turns were excluded because the main task here is to dissect how a comment as such can initiate a sequence.
Accordingly, the comments feature declarative syntax, such as in juba muld paistab ("the soil is showing already"), siin oli mingi nukik ("[there was] some kind of a lump here"), or a mere noun phrase, as in vai ("a pole"). They are mostly neutral but may also involve an evaluative element, as in siin ei ole väga palju ("[there] is not much here"), or sitke, ei taha ära tulla ("tough, doesn't let go") (commenting on a piece of dung stuck to the speaker's fork), thus simultaneously functioning as assessments 1 (15 instances). The utterances overwhelmingly feature positive valence, with only a few cases in the negative, as in keegi pole veel selfiet teinud ("nobody has taken a selfie yet").
Topically the comments target abrupt events, such as when an object emerges from the dung: a pole, a trapdoor, or "something really hard" (10 instances). A similar sense of local discovery is involved in commenting on animals and birds, such as an entering cow (three instances). Other ostensibly inspection-based comments attend to events that happen across extended periods of time and typically concern the work's progress, such as juba muld paistab ("the soil is showing already") (commenting that the layer of dung has almost been removed), or ring tõmbab koomale ("the circle [of workers around the area still to be dug] is becoming smaller") (11 instances). At the same time, the comments may provide guidelines for what is yet to be done-siit on mõned labidatäied ja siis on sellega korras ("a few more spadefuls and then it'll be done") (four cases, some of them simultaneously attending to work progress)-or make a complaint-see vist läbi ei lähe ("this doesn't seem to go through") (while poking with the fork) (four cases). Many of these comments are occasioned by not merely visual inspection but raw physical engagement with the surface. Finally, a group of five comments concern peoples' activities or appearance, such as see on juba nagu see hiina toidu söömine pulkadega ("this is like eating Chinese food with sticks") (about lifting straw with dung forks) or Jaan on nagu laudatööline, valge särk seljas ja ("Jaan is like a true stable worker, a white shirt on"). In two cases, as in the latter example, individuals were targeted in the comment, which of course increased their response relevance-and these are not discussed in the following. However, it could also happen that the targeted participant was unlikely to provide a fitting response, as in see pull on ilmselt kade et me saame siin sees jahedas mõusalt olla ("this bull is obviously envious of the fact that we can be comfortably cool indoors"). There were six comments in the collection, 1 Similar "assessing" noticings are included in Szymanski (1999, p. 5-6), and Stivers and Rossano (2010, p. 14-15).
such as the latter, that could be seen as an attempt to joke, and four that could be characterized as complaints. The rest were more or less neutral. None of the aforementioned categories featured exclusive initiation or self-talk, and overall, 22 of the comments are treated as self-talk.
In the following, we will be asking what it takes for a comment on the physical surroundings to initiate a conversational sequence-that is, not to merely attract a brief gaze but to occasion another conversational action that builds on it. To answer this question, multimodal interaction analytic method (Goodwin, 2000;Mondada, 2012) will be used. We will thus be scrutinizing the participants' body posture, speech volume, and gaze. The overall activity trajectory and the target of the comment will also be considered, to argue that its initiatory potential is intimately tied to its embodied delivery and timing.

Comment as a sequence initiation
Let us start by looking at an instance (Excerpt 1) where a comment launches a conversational sequence. The workers have been toiling in silence for 1.08 min, and the wheelbarrow is filling up. Vello puts a last forkful of dung on top of the load ( Figure #1A), while Rain is already bending down to grasp the handles and take the wheelbarrow out of the stable. After 1.3 more s of silence Vello utters line 2: see sitajunn on seal õlgedel nagu kirss koogil ("That piece of shit on the straw is like a cherry on a cake"). This comment targets the piece of dung that Vello himself just lifted onto the wheelbarrow. It is based on perceptional information from the surroundings, featuring affirmative declarative syntax and a nonresponsive position after a long silence, all typical features of the comments in the current collection. The abrupt appearance of the target, along with the perceptional basis of the claim, could also count as criteria for this utterance to be considered a noticing (Keisanen, 2012;Kidwell, 2009;Kääntä, 2014;Laanesoo & Keevallik, 2017;Sacks, 1992II, pp. 87-97, Schegloff, 1988. (1) Sõnnik 6; 02:10 #1B That piece of shit on the straw is like a cherry on a cake.
#1A Vello puts the last piece of dung #1B Jaan (behind Rain) raises in the wheelbarrow standing in his gaze towards Vello. the middle of the circle of workers, Rain is bending over to start pulling it out of the stable.

(3.5#2A)
At least four people have stopped working before line 2, but no one gazes toward Vello at the beginning of his utterance. The utterance includes two elaborate references, the noun phrases see sitajunn ("that piece of shit") and seal õlgedel ("on the/that straw, lit. there:ADESSIVE straw:ADESSIVE"), and it can therefore be appreciated even without visual access to the referent. During the utterance, the speaker does not select a recipient by gaze, as he is looking at the leaving wheelbarrow. This is a situation that has been shown to reduce response relevance by any single participant (Meyer, 2010, pp. 153-154;Mushin & Gardner, 2009;Stivers & Rossano, 2010). Nevertheless, Jaan lifts his gaze toward Vello by the end of the utterance (as shown in Figure #1B) and then turns toward the leaving wheelbarrow. Several participants are already bodily oriented toward the target of Vello's utterance, as they are standing in a circular formation, waiting for the wheelbarrow to be replaced.
A pause of 0.9 s ensues. During that time three gazes turn to the wheelbarrow. The only vocal response is by Mari (äh in line 4), and it is not appreciative or aligning in any sense, at most a neutral receipt. By the end of an even longer ensuing pause (3.5 s) most people are gazing toward the wheelbarrow, and it is possible to trace a smile in Rain's face ( Figure #2A)-an appropriate reaction to the jocular comment. Finally, after almost 5 s Renee counters with another joking utterance about the dessert (line 6). This statement clearly builds on Vello's turn and would be incomprehensible on its own. Renee's turn receives a hearable laughter by Jaan, who lifts his head to gaze at Renee ( Figure #2B). Renee himself is bending down and digging the dung. More silence ensues, and as soon as the next wheelbarrow appears in the doorway, people start working again.
A note on the temporal organization of conversation at the stable is in order at this point. On the basis of talk-focused settings it has been established that conversational turns are produced with no gap and no overlap, so that silence between turns is minimized (Sacks et al., 1974). In contrast, in the current activity setting, extensive silences are tolerated. Conversational sequences can be developed across considerably longer time frames when people are continuously co-present and preoccupied with physically strenuous work #2A General orientation towards the #2B Renee still looking down and wheelbarrow. Vello and Rain smiling.
digging, Jaan staring at him and laughing briefly.
For dinner we'll now be having a cherry cak-cake with a cherry, #2B 9 ((Several people bending down to anticipate the arrival of an empty wheelbarrow.)) (Keevallik, 2018). The moral norms of participating in talk are somewhat relaxed in this setting, since it is equally moral-or even more so-to participate in the work task and get it done as quickly as possible. 2 Here a comment can be retrospectively treated as an initiation across an extensive silence, as happened in the current Excerpt (1). Only two people are preoccupied with digging and one with pulling the wheelbarrow when the response could be expected, but the continuation is still delayed, at least according to the norms of talk-focused activities. This initial excerpt thus also illustrates how difficult it may be to initiate a conversational sequence in this co-present state of incipient talk. Despite the pauses in lines 3 and 5, this exchange can be seen as containing a conversational sequence proper: two topically related turns and two vocal reactions (äh, laughter) together with at least one smile. The factors that contribute to the treatment of Vello's comment in line 2 as an initiation are its timing in relation to the work task, Vello's bodily orientation, and his loudness of voice. To start with, the leaving wheelbarrow constitutes a natural moment for a brief rest (and some of the workers take a short break). Vello simultaneously deploys the fact that the workers are orienting toward the center, the wheelbarrow, whose contents he refers to in his utterance. In fact, the bodily configuration of the participants is reminiscent of a large dinner table, with everybody being positioned facing the center. This bodily arrangement provides the basis for a common attention focus. Furthermore, Vello himself is standing up straight, with his head high, and orienting toward the group of workers. His line is uttered in a loud and clear voice, which cuts into the long silence, where even the clicking of forks and spades on the rim of the wheelbarrow has ceased. This kind of embodied presentation holistically amounts to the line being produced for the audience. Last but not least, the comment contains a surprising comparison between dung and food, which can only be appreciated by an immediate look at the wheelbarrow before it disappears out of sight. All of these features are likely to have contributed to the fact that the turn attracts a receipt and is eventually built on topically. In the following, we will compare this instance to a comment that is produced and treated as self-talk.

Comment as self-talk
In comparison with initiations, comments treated as self-talk feature a markedly different embodied delivery in terms of gaze, posture, orientation, and loudness. They do not make use of upcoming activity transitions or others' attention foci. Neither do they target a surprising or transient object to be appreciated right at this very moment.
Excerpt 2 starts just when a previous topic has been terminated in lines 1-2. There is a 3-s silence, and then Toomas, who is poking the ground, utters kurat see on nagu põhjatu sin ("damned it's like bottomless here"). The stable is built on a dirt floor, and the dung has to be dug out until the soil, the "bottom," shows. Toomas's comment concerns the depth of the dung layer, which is highly relevant for the accomplishment of the joint work task. However, since he formulates an extreme version-that there is no bottom at all-the utterance sounds like a complaint. As a comment on the surroundings, it features a nonresponsive position after silence, a claimed perceptual basis of the comment, and affirmative declarative syntax. It is followed by a silence of more than 36 s. The comment is treated as self-talk in this case, it does not attract a verbal reaction of any kind. (It is possible though that Veiko and Piia briefly gaze toward Toomas, Figure #3B). Among other things, this excerpt shows how even a comment that is potentially relevant for the progress of the work task is left with no response. (2) Sõnnik 5; clip 02 Several temporal and bodily factors contribute to Toomas's comment being treated as selftalk. To start with, it is produced in the middle of filling a wheelbarrow, not at an activity junction to rest. Second, everybody apart from Renee, who is manipulating the wheelbarrow, is orienting toward their own work spaces and tools. They are thus not configured as an audience. Renee is looking toward the center of the stable, away from Toomas, who is working slightly behind him ( Figure #3A). Third, the utterance is produced while digging: Toomas is bending forwards, poking the ground with his fork on see on ("it's") and pushing it deeper into the dung on the word nagu ("like"). He is thus presenting himself as preoccupied with shoveling. Furthermore, his body is half-turned away from the center of the group, toward the wall, and his gaze is steadily on the ground. The words are uttered softly and relatively quickly. All of this is in contrast with Excerpt 1: the activity phase, the posture and bodily orientation, the gaze, the clarity of speaking, and the availability of recipients.
Toomas's utterance begins by swearing, which constitutes one of the main types of soliloquy (Goffman, 1981, pp. 97-99). Swearing only potentially attracts the attention of others, who might check out the problem that occasioned it. Here, however, the swearing prefaces a statement about the environment, a discovery that Toomas has just arrived at, which at least weakly makes relevant an alignment or possibly even an offer of assistance to dig down to the dirt floor of the stall. Contentwise Toomas's comment contains bad news to the group of workers, since the time that they have yet to invest is in direct correlation with the depth of the layer of dung. Toomas's gazing at the ground constructs the reference with sin ("here") as a very local one, the space in front of him. At the same time, it could refer to the whole area they are clearing, as the ground is flat and the depth of dung approximately equal across the stable. In other words, the utterance does not require an orientation to the precise spot in front of Toomas. Importantly, the referent in this utterance-the stall floor-is constantly present. There is no urgency in orienting toward it, whereas the piece of dung on a straw bed on top of a wheelbarrow in Excerpt 1 was a temporary sight to be attended to immediately or else it would be missed. In short, Toomas's embodied production of this comment and its timing in relation to the sequence of work phases result in the turn being basically ignored by the others and treated as self-talk.
The first two excerpts have showed that the comment produced in a loud voice, orienting toward the other participants, with the speaker's gaze toward the center and deploying an upcoming activity boundary, was treated as an initiation, while the one produced in a soft voice, with the speaker's head and gaze directed downwards while poking the dung, was not. The cases also suggest that a comment with a surprise value is more likely to be followed up than a complaint. We will therefore continue with a pair of cases where the comments concern objects that are in some sense surprising in the context, thereby targeting the question to what extent the subject matter is of relevance in its subsequent treatment.

Subject matter is less important than the bodily delivery and participant configurations
During the workday the participants find several objects in the dung that do not necessarily belong there, among other things, a pole, string, a trapdoor, stones, beer bottles, and a jacket. Comments on some of them launch a conversational sequence, while others do not. The crucial difference lies once again in the details of embodied delivery as well as the current bodily orientation and attention foci of the participants.
At the beginning of Excerpt 3, Vello and Renee are jokingly discussing the origin of a formulaic saying in Estonian (lines 1, 3), orienting toward each other ( Figure #4A). During line 3 Meelis walks by and steps on some string that has got stuck between the confinement structure for lambs and the wheelbarrow Renee is holding. His stepping on the string slightly tilts the wheelbarrow. While Meelis walks further without noticing it, Renee starts to bend down during his disagreeing turn in line 5 and turns his gaze toward the string, as shown in ( Figure #4B). He then utters the particle ota ("hey/wait a minute"), 3 which is a regular device for initiating a digression (Keevallik, 2003, pp. 126-140). This word is produced softly and with a bent posture, but Renee continues loud and clearly mingi nöör läheb mulle siia g-kärusse ("there is some string that's coming to my wheelbarrow"). He is gazing at the string but standing in the middle of the stable, as if at center stage, and as a speaker he is thus able to naturally make use of the physical configuration of the working bodies around him. Three people turn to gaze at the string, as shown in Figure #5A. Crucially, Vello chooses to develop the topic using word play. As the Estonian pronunciation of the car brand Citroën comes close to sitanöör, lit. "shit string," he seizes the opportunity to use this folk etymology as a suggestion to what Renee has found-a true "shit string" (line 9). This attracts laughter from two people, and the conversational sequence has been launched.
(3) Sõnnik 3; 23:36 #4A Maybe because of that there is a saying close your mouth, heart Close your mouth or your heart will cool down or.
#4B Renee bending over to check Meelis has just stepped on some string on the string. that has slightly tilted the wheelbarrow Renee is holding. #4B°°# 5A #5A Renee grabs the string. Vello, #5B Renee gathering the string. Meelis and Siim gazing at it.

That I don't quite believe.
Hey, there is some string that's coming to my wheelbarrow.
This is a genuine shitstring. ((colloquial for Citroën)) e: While Renee is gathering the string in his hands ( Figure #5B), Vello produces another joke about a tapeworm (line 13), which leads to a further joke by Renee (lines 17, 19). Thus, the comment on a surprising object in the middle of the stable has been treated as an initiation of a conversational sequence, where the participants successively build on each other's turns. While the later developments may be contingent on the success of the first joke in line 9, the production features of Renee's initial comment in line 7 are similar to those in Excerpt 1. Renee makes use of the existing bodily configuration around him and utters his line loud and clearly. Even though his gaze is on the target of the comment and he is bending down to pick it up, his position higher up than the others on the still-to-be cleared layers of dung affords him an orientation toward the group. Last, the object has likewise just been discovered, and the comment formulates a perceptual (visual and haptic) experience, thus also constituting a noticing action along the lines of some recent literature (Kääntä, 2014;Kidwell, 2009). In contrast to Excerpt 1, however, here the speaker is holding the object, which makes relevant others' gazes toward him rather than in some other direction in the stable. More importantly, the comment is not placed at an activity boundary, which shows that the embodied delivery of the comment and the bodily configurations by the potential recipients may be more relevant than the activity boundary. The comment on the surroundings, if not interruptive (as described for noticings by Sacks (1992 II, p. 91)), does not advance the prior conversational sequence and proffers a new locally sensitive topic.
In contrast to string, one might think that finding two bottles of beer in the dung would occasion a discussion. In Excerpt 4, after a silence of 2.4 s, Vello lifts up a bottle of beer and utters a proposal Oo, teeme peo ("Wow, let's have a party") (line 1, Figure #6). Only Renee seems to be gazing toward Vello, but he has been in this immobile position for a while and thus cannot be seen to clearly react ((turned towards someone else in the back 4 ; Siim bends down to help to disentangle the string)) #5B ((Mari talking to someone else in the back)) It was once a tapeworm. ((Renee puts the string into the wheelbarrow)) Yeah.
It's the dental thread that a dog has eaten up probably.

4
In this case, there is also a parallel conversational exchange in another part of the stable between two people, but they speak softly and orient only to each other. With up to 11 people present, schisming (Egbert, 1997) is of course always an option.
to Vello's turn. Renee produces a vague vocal reaction äh, which is hardly an enthusiastic agreement with the proposal. After a lengthy pause (in line 3), he instead utters a directive to another coworker, Margus, to push the wheelbarrow toward him-that is, to the back of the stable where most of the people are digging. This directive breaks sharply into the potential slot for developing a conversational sequence on partying with the beer. Renee's directive is loud; he is standing up straight with his body oriented toward the center of the stable and the circle of workers. In contrast to the nonwork-related nonserious proposal by Vello, the work-related instruction by Renee leads four more people besides himself to look toward Margus, who is currently in charge of the wheelbarrow ( Figure  #7A). It seems that organizational talk about work is given priority here in relation to the odd finding. While Margus starts pushing the wheelbarrow toward the back of the stable, Vello finds another bottle of beer and utters a comment about the brand-they are both homemade (line 6). This action also qualifies as a sudden noticing, as Vello is visibly and observably just discovering the bottles and characterizes them in his utterance. He is bending forward over his findings, half-turned away from the center (see Figure #7B), and his voice is markedly quieter than Renee's was in line 4. His voice is also somewhat muffled, which may contribute to his comment being treated as self-talk. In addition, the workers are visibly orienting toward the coming wheelbarrow that implies more digging for them very soon. Vello's comment in line 6 ends up as placed unfavorably in regard to the activity structure, as there is no break coming up. When he lifts up the second bottle, several people are already digging ( Figure #8A).
huh Listen Margus, roll it here.

1
Finally, after a lapse of more than 7 s, Vello walks toward Margus ( Figure #8B), who is waiting for the wheelbarrow to be filled, and asks a question that is uniquely addressed to him: sa ei ole kaotand õlud ära ("you have not lost the beers"). Vello is now gazing at Margus and uses a second person singular pronoun to address him. This is the first time that his utterance attracts a proper response, albeit a curt one, in the current excerpt.
The comment in line 6 is treated as self-talk. Crucially, Vello has not secured the bodily orientation of others. In contrast to the comment about a piece of string at "center stage" in Excerpt 4, this one is done with the speaker's back almost turned to the circle of workers. Vello's gaze is on the bottles, away from the other participants, and he is muttering while bending forward. Indeed, the lifting of the bottles amounts to a large visible gesture by Vello ( Figure #8A), but that does not seem to compensate for the lack of favorable bodily configuration among the possible recipients. Thus, regardless of the potential high surprise value of the target objects in this excerpt, who is now looking at him.

#8B
Oh, both of the beers are homemade.
You have not lost the beers.
No the embodied delivery of the comment as well as spatial orientations and participation configurations define it as self-talk. More generally, this shows that the subject matter of the comment is less important than the embodied contextual details surrounding its production. Among other things, complaining versus entertaining comments in the current collection did not exclusively result in no response versus continuation of talk respectively, as might have been expected. Three out of four complaints were indeed treated as self-talk, and the fourth instance featured an open-class repair initiation ("what?") as a next action, but also one out of six markedly entertaining comments did not attract a response. Accordingly, the previous analysis did not focus on the subject matter as much as posture, gaze, and attention foci in relation to activity trajectories, which systematically appeared to be relevant for how the locally sensitive comments were treated.

Conclusion
This article started out with a collection of comments on the physical surroundings uttered during a workday in a sheep stable, focusing on a subset of declarative utterances that feature lowest response relevance. The main question was why some of them were treated as self-talk and others as initiations of a conversational sequence that transformed a nonfocused interaction into a focused one. The study demonstrated that the embodied delivery of the comment and the current bodily constellation of fellow workers played a crucial role in whether a comment attracted a response or not. It thereby expanded on the earlier study by Stivers and Rossano (2010) that identified four factors that could enhance the response relevance of noticing, announcing, or assessing turns, among them gaze toward the recipient. The current study demonstrated that further factors in attracting a response included the speaker's body orientation in relation to the other participants, whether the gaze was on the ground/object or toward the recipient(s), whether the body posture was upright or bent forward, and how loud the comment was. A mumbled comment while bending forward and gazing down hardly ever attracted as much as a gaze by others and was treated as self-talk. In contrast, a comment that was uttered loudly and built on the current attention focus was more likely to be treated as an initiation of a conversational sequence. Furthermore, it was argued that for the development of the sequence it could be an advantage if the comment was placed at an upcoming activity junction between work and rest-that is, when the wheelbarrow was leaving the stable and the workers could take a short break. This particular timing apparently eased participation in talk as opposed to work. Interestingly, the subject matter of the comment, the targeted referents, played a minor role. Comments on stable entities, such as the dirt floor under the dung layer, were indeed often treated as self-talk, while those targeting odd or transitory objects tended to attract a reaction by others. Still, as demonstrated in Excerpt 4, local bodily configurations and attention foci were more important in terms of whether the potential recipients reacted to a comment or not. In unfavorable participatory moments even comments with high surprise value could be ignored. All of this shows that not only conversational initiations but also self-talk are locally negotiated in current embodied activity trajectories. Responding constitutes the interaction as focused, while not responding constructs it as nonfocused. In other words, even self-talk is coconstructed during states of incipient talk within its immediate social, temporal, and spatial context. During the day at the stable it became obvious that some participants' comments launched a conversational sequence more often than others'. Among the people represented in the current article, Vello seemed to receive more reactions than Toomas. Instead of assigning this to any abstract social hierarchies or personal characteristics, we can now see that the difference depended on the embodied comment design. The explanation lies largely in the two participants' bodily behavior, sensitivity to activity junctions, and local contextual configurations. In addition, action type seems to have played a role. While complaintlike actions could be ignored (three out of four), jocular comparisons, e.g., of dung with food, were more likely to attract ensuing turns of talk (five out of six). In a discussion on locally sensitive utterances, Bergmann (1990, p. 215) argues that the cointeractants abandon, even if for a moment, their roles deriving from extrasituational bonds and take on a shared identity of being, e.g., a "witness" or a "victim" of a local event, achieving a sense of mutuality and "phatic communion." Perhaps taking the role of a complaining "victim" is not such a successful strategy for communion in a joint work situation, while attempts at entertainment were appreciated. The current article has showed how mutuality is not by itself guaranteed by a comment on the physical surroundings. At least in a multiactivity setting of talking and working, it is a joint achievement.
The dual involvement in talk and work at the stable implies that a conversational sequence can involve extensive but unproblematic silences. One could assume that this is due to physical impossibility and exhaustion. As became evident, people can indeed talk while bending forward with a spadeful of dung, poking the surface, etc. (see also Keevallik, 2018). Instead, it was argued that there are embodied ways of producing an utterance as an initiation rather than self-talk and that there may instead be organizational reasons that favor the uptake of a topical thread rather than letting it pass. micropause : lengthening of a sound°low volume .
pitch fall at the end of an intonation unit ?
pitch rise at the end of an intonation unit , level pitch at the end of an intonation unit (not in Estonian)the part is not expressed in the Estonian version (X) unhearable syllable # the moment of the frame grab