Towards an embodied Grammar : Gesture in tying practices Constructing obvious cohesion

Laursen, Lone
University of Copenhagen, Denmark, lone@hum.ku.dk

Abstract

It has been thoroughly demonstrated, that gesture and bodily conduct are parts of the interactional organization, especially when it comes to the organization of sequences and overall activities. When it comes to the linguistic structure of utterances, though only few researchers have focused on demonstrating how gestures are part of the linguistic structure of utterances in interaction. From a conversation analytic perspective this article will demonstrate how gestures are parts of the turn constructional unit, and how they are deployed as tying techniques to connect utterances over time, not only ensuring coherence but also constructing obvious cohesion. The examples are taken from a large data collection based on video recordings of everyday conversations and workplace interactions.

Keywords : Interaction, Embodiment, Tying techniques, Cohesion, Conversation Analysis

Though coming from a different field, I was originally inspired by Armstrong, Stokoe & Wilcox (1995). The analysis in this article is closely connected to their belief that an analysis of the physical structure of visible gesture provides insight into the origin of syntax. With the important difference, though, that I'm not interested in the origin of syntax. I am interested in how gesture is an observable part of syntax of a single ongoing turn constructional unit (TCU), and in how we may come to speak of an embodied grammar, here and now. It will be demonstrated that gesture is a systematic part of a single TCU, and thus part of an interactional grammar (e.g. Ford, Fox & Thompson, 2002 ; Ochs, Schegloff & Thompson, 1996 ; Selting & Couper-Kuhlen, 2001 ; Steensig, 2001). In this paper I have chosen to focus specifically on gesture as tying technique, and I shall demonstrate how gestures not only ensure overall coherence (Kuehn 2002), but also construct obvious cohesion between two turns.

1. Data and method

The data examples in this article are taken from a large data collection based on video recordings from naturally occurring spontaneous everyday conversations and workplace interactions. The gestures dealt with in this article can all be described as iconic, meaning that they can be said to depict or describe a concrete or abstract word, and that they are linked semantically to the content of speech. The examples are either translated from Danish into English, or in lingua franca, English. The translations are not in correct English. Rather they are an attempt to use more or less the same words that are being used in the Danish original, so that the relation between gesture and speech is maintained (for the Danish transcripts see appendix). A video clip is linked to each excerpt.

Conversation Analysis (CA) is the methodological basis for the analysis. Because of CA's insistence on including all observable details of an ongoing conversation in the analysis, it offers, through its sequential analysis, a method for describing the construction of the single unfolding turn in interaction, and thus provides for a better understanding of how grammar is constituted in the course of turn construction.

2. Repeated gestures as connectors

Throughout the course of an interaction it can be observed that gestures are repeated by same as well as by other participants. McNeill et al (e.g. 2002) have developed the theory of catchments in order to describe a possible underlying cognitive unit in the language production, which is revealed through the repeated gesture. Other researchers have looked at these kinds of gestures from an interactional perspective in naturally occurring interactions. Close sequential analysis has revealed that these repeated gestures are part of not just one, but several different practices and interactional activities, e.g. word search activities (Goodwin & Goodwin, 1986 ; Gullberg, 1999 ; Laursen, 2002a), alignment practices (Laursen, 2002a ; Lerner, 2004), clarifying activities (Goodwin, 2000 ; Goodwin, Goodwin & Olsher, 2002 ; Laursen, 2002a-c) and tying practices (Kuehn, 2002). The repeated gesture, though, always has an element of connecting.

Beginning from an overall perspective, by e.g. observing a lively debate on TV or a story telling sequence in naturally occurring interaction, one notices immediately that different participants produce the same gesture, and that these gestures are part of ensuring the overall coherence in the interaction (e.g. Kuehn, 2002). In this article it will be demonstrated that these gestures are not just part of ensuring coherence, but part of constructing obvious cohesion between the single sequences and between the stories of different participants.

In Excerpts 1 and 2 participant T tells and completes a story. This story contains a description of a water-skier hurting his shoulder badly. During the story T produces a gesture embodying how the water-skier's shoulder went out of place (fig.1.1). One minute and 18 seconds later participant I initiates a story about one of her team players from her old volleyball team. Participant I's story follows with only a (0.3) pause after T has completed his story, and she initiates it with the words ‘There was a girl from my old volley team it ha :: :: happened on a regular basis that she'. At this point participant I hasn't said or done anything to connect her story to the prior story. It turns out that this story is also about someone whose shoulder went out of place - i.e. the semantic coherence between the two stories can be found in the out-of-place-shoulders. This semantic coherence becomes obvious in the continuation of I's story ‘when she jumped up to smash (0.5), tck- Then her shoulder went out of place'. But the obvious cohesion is constructed before the speech shows it, namely when I (fig.2.1) reproduces T's gesture/body posture simultaneously with ‘when she jumped up to smash' (ll.303-306).

Excerpt 1

Fig.1.1 (link to video clip)

 Fig. 2.1                   Fig.2.2

 (link to video clip)

This article will focus entirely on gestures in tying practices. Common to all these gestures is that from an overall perspective they occur in activities where participants demonstrate an orientation towards creating and securing mutual understanding. From a more narrow and local turn perspective it will be obvious that these gestures are also part of the single turn construction and turn organization, and that they connect different turns or TCUs to each other. Before analyzing the gestures I will say something general about the construction of coherence and cohesion in interaction.

3. Coherence in interaction

One basic assumption about coherence is that a turn, which is placed adjacent to another turn, will be understood in relation to that prior turn (e.g. Sacks, Schegloff & Jefferson, 1974) - i.e. talk is taken to be relevant and coherent even when it is not obviously cohesive (Broe, 2003). The placement of a turn, and the placement of words within the single turn constructional unit (TCU), thus makes out a strong factor in creating coherence in interaction. Still, participants in an interaction use specific techniques to create coherence and to specifically construct cohesion. Some of the most commonly used techniques are the use of pronouns, deictic proterms, con- and disjunctions, (partial) repeats of the turn and the production of increments (e.g. Broe, 2003 ; Lerner, 2004 ; Local, 2004 ; Mazeland & Huiskies, 2001). I shall demonstrate that gestures should be added to this list.

Coherence may be found at different levels. One may talk about coherence at the pragmatic level, where the concern is whether the relationship between the different actions in the interaction is relevant and coherent. The participants' normative orientation to the sequence organization is an example of this : if a participant produces a 1. pair part (e.g. a question) it becomes relevant for other participant(s) to produce a relevant 2. pair part (e.g. an answer). But not only do the participants demonstrate their expectations of an answer by repeating the question if the answer doesn't come, at this point any following action will be treated as either the answer or something leading up to the answer. As Schegloff points out, coherence thus lies in the adjacent placement and the relevance between question and answer : "[t]he sequence structure itself can provide for the organizational coherence of the talk" (Schegloff, 1990:53).

Coherence can also be found at the semantic level. Here we talk about the fact that two turns are understood as coherent because of their semantic content. An example of semantic non-coherence could be that someone answers ‘football' to the question ‘what would you like for dinner ?'. The pragmatic coherence is ensured because an answer follows a question, but semantically there is no coherence between the two turns.

Coherence, at the syntactic level, or at the turn organizational and turn constructional levels, is about grammatical structure. Pronouns, deictic adverbs, conjunctions and disjunctions are typically used to connect and construct concrete observable cohesion between two turns, while verb-, noun- and adverb-endings are typical connectors at the lexical level and create obvious cohesion between single words, typically within the same ongoing turn.

3.1. Some tying techniques

When participants want to connect an ongoing turn to another turn than the just prior one, they use different tying techniques to localize and connect to that turn. Typically they will do one of two things : they will produce an increment, i.e. a syntactic prolongation or continuation of prior turn ; or they will produce a partial repeat of the prior turn.

A turn increment may be produced when a participant is skip-connecting. This is a practice or tying technique participants may use to connect back in time to a previously so far completed turn. When skip-connecting, a participant links his/her current talk to some prior talk, skipping over the intervening talk of other or self (Local, 2004). Turn increment initiators, such as ‘but, and, or, with, because, well' etc., are often used to skip connect to a prior turn. These increment initiators, or continuers, connect the upcoming turn as a continuation of a previously possible completed turn. The choice of increment initiator also establishes the type of connection between the two turns :

In 1. 18 participant S produces the evaluation ‘that's a good one', thus orienting to the fact that participant A's story may be complete after A's punch line and so far possible completed turn ‘each of us has ten clips' (1. 17). But after S' evaluation, participant A skip-connects 1. 19 to his prior turn by initiating the turn with the conjunction and increment initiator ‘and eh ::', thereby demonstrating that the upcoming turn is to be understood as a continuation of A's own prior turn (1. 17). In this excerpt ‘And eh' projects a continuation of prior turn, but one could imagine that if the increment initiator had been ‘but' or ‘because', this would have created a completely different kind of connection between the two turns, and would have projected that the action in the upcoming turn would have been either a modification or an explanation of prior turn.

(Partial) repeat is another important tying technique. Where skip connecting has to be done fairly close to the abandoned turn, repetition is a tying technique that is relatively independent from the turn's actual position (Sacks, 1996, p. 733). A (partial) repeat of a prior turn localizes the turn as the product-item of ongoing turn (Jefferson, 1972). By exploiting the identity of turn-constructional materials in order to enable the recipient to identify the source turn (Lerner, 2004), repeats lend a turn the capacity to tie it to another one than the one it is next-positioned to. Often repeats are done in repair activities (e.g. Schegloff, 2004 ; Stivers, 2005), when hearing or understanding problems can be localized. But the technique can also be observed when participants simply connect an upcoming turn to some prior turn, thus re-activating the prior turn.

In this paper I shall demonstrate how gesture makes out a thus far ignored, but very important part of the tying techniques and of constructing obvious cohesion between two turns. This will be explored further in the following.

4. Gestures ensuring coherence - constructing obvious cohesion

I shall present another six different excerpts from my collection of data, to demonstrate and exemplify how gestures are deployed as tying techniques in three different sequential and turn constructional environments :

1. Gesture as connector between single words within a TCU after an insertion of same participant.

2. Gesture as connector between turn units of same participant after other participant's turn, or side sequence allocated by other participant.

3. Gesture as connector between different participants' TCUs.

Participants' orientation to the well-described synchronic production of gesture and related speech is of vital importance when describing how gestures make out a systematic part of turns in interaction. McNeill, et al (McNeill & Duncan, 2000 ; McNeill et al 2002) described synchronicity in connection with cognition while developing the notions of growth points and catchments. But this well described synchronic production of gesture and related speech is not only a cognitive concept or a theoretical frame, it is about units that the participants demonstrate an observable ongoing normative orientation towards throughout the interaction (e.g. Laursen, 2002a-c). Before demonstrating how gestures are parts of the interactional tying techniques I shall exemplify how the participants in an interaction demonstrate an orientation towards the well-described synchronicity :

Fig.4 (link to video clip)

The gesture in excerpt 2 relates to the lexical component ‘au point'. It is initiated just before ‘au point', and realized and completed simultaneously with this lexical component. This synchronicity is, as mentioned before, very well described, and can be observed everywhere throughout an interaction. Synchronicity is observable even with restarts, repetitions and insertions, which the following excerpt will illustrate. The next excerpt takes place just prior to excerpt 4 and participant A is producing the same gesture as in excerpt 4.

Excerpt 4a

A interrupts his own speech production in l. 5, ‘It must have been ri :: ::ght (.) au- au-', after N's speech onset in l. 6 - ‘It was almost over the edge'. A also interrupts his own gesture production, but he keeps his arm/hand in a so-far-rest-position, and the gesture so to say freezes in space. At the first possible completion point of N's turn, l. 7, ‘It was almost over the edge to be on target', A restarts his speech and produces the abandoned word ‘au point' in overlap with N's turn increment ‘I think'. The gesture is also resumed and completed synchronically with its related lexical component ‘au point'.

It is worth noticing that the resumption is said in overlap with N's turn. Often a turn said in overlap is repeated in the open afterwards, and often it looses some of the features that positioned it in its original sequential environment (e.g. Schegloff, 2004). In this excerpt we can observe the repeat in the open, which we have already seen in excerpt 4, with a typical change in prosodic markers and in tense. But we cannot observe any change in gesture. It is a common feature, and recurrently observable that the gesture remains a part of the repeated turn. I emphasize this because many scholars regard gestures on equal terms with paralinguistic markers such as intonation and rhythm, which are typically changed when repeated in another local environment. This paper will argue further for seeing gestures as linguistic markers, non the least, because the participants treat gestures as they would treat speech.

The following examples will further demonstrate how the participants in the simultaneous production of gestures and related speech orient to an interactional structure where gesture is an integrated part of the single turn constructional unit.

4.1. Gesture as connector between single words within the same TCU

In excerpt 5 it is observable not only how gesture and related speech are repeated together, but also how gesture is tying the repeats of the turn to the original not completed turn. If we have a look at the excerpt without gestures we can observe how the participant ties back to her non-completed turn after having produced a self initiated insertion. Verbally she produces a partial repeat of the turn and the turn initiator ‘and' as tying technique :

Excerpt 5

Participant B talks about a snake she saw on the beach. She abandons her ongoing turn (l. 38) after a hesitation marker ‘in eh :: :' and produces an insertion, giving a possible explanation for pushing the snake and giving an explanation for the snake's presence on the beach. After the insertion and a minimal response token from participant D (l. 42), B resumes her abandoned turn (l. 43) by producing an increment initiator ‘and' followed by a pronoun, followed by a partial repetition of the abandoned turn ‘And they p ::pushed it up into the ::eh'. Apparently she thus uses three different devices for tying back to her prior non-completed TCU (l. 39) ‘pushed it up in eh ::' : an increment, a pronoun, and a repeat.

But together with the abandoned turn (ll. 36-39) ‘who then- (0.2) pushed it up in eh :: :', B has also produced a gesture that illustrates a sweeping movement and relates to the lexical component ‘pushed up' (see pictures 5.1-5.3). This gesture is reproduced together with the resumption and repetition of the abandoned turn :

 pict.5.1        pict.5.2            pict.5.3

(Link to video clip)

The gesture is resumed and repeated twice, synchronically with the restarts of the talk - l. 43 and l. 45 : as already noted gesturally, an insertion ll. 38-39 ‘It was on its way to the ocean. It wa ::s (0.2) °confused° or something like that'. First, she points in another direction indicating the location of the ocean, and secondly, she produces a gesture that may express uncertainty or not knowing : she opens her hand and moves it sideward. Together with her restart in l. 43, she also reproduces the gesture of the interrupted turn. She stops the gesture together with the pause in the turn after the hesitation markers. And she again reproduces the gesture together with the second resumption of the turn. Speech and gesture are completed together at turn completion.

We can here observe the well-described synchronicity between gesture and its related lexical unit, and we can observe how the participant in the gesture production demonstrates an orientation to the single turn constructional units, where gestures are an integrated part of the single TCU, i.e. not something dispensable when the turn is repeated (Schegloff, 2004). The sequence has many of the typical marks of a word search : sound stretch, hesitation markers and repetitions. And the participant completes the sequence when she produces the word ‘grass'. So this is also an example of gesture in a word search activity, but my main reason for choosing exactly this excerpt is that it is a clear example of gesture being a part of the syntax : even though the gesture is modified when repeated, or rather de-tensified, it remains recognizable as the same gesture. The speech produced together with the gesture is by the first repeat also produced in a ‘slimmer' version than by the first production (pronoun instead of the noun), but by the second repeat both the verb and the pronoun are omitted from the speech. So the turn units look like this :

Original turn ‘people who then- (0.2) pushed it up' + synchronic gesture

1. repeat ‘and they pushed it up in the eh :: :' + synchronic gesture

2. repeat ‘into the grass up behind' + synchronic gesture

First, the participant thus demonstrates an orientation towards not just the relevant lexical item in the production of the repeats, but also towards the relevant ongoing turn constructional unit. Secondly, the repetition of the gesture maintains the syntax of the turn even though essential parts are left out of the speech : Verbally the second repetition only contains the locative, but gesturally the sweeping movement still contains both the action (verb) and the agent (subject, noun) - see also Armstrong et al (1995). This means that even though the verb, the subject and the object are omitted from the speech, the utterance has not lost its syntactic structure or its semantic content, because they are still present gesturally. Thirdly, the gesture in the second repetition ties back to the original turn : embodying the action and the agent of pushing this gesture does the same work as the pronouns did in the first repetition, it connects the turn back to ‘they'/‘people' and to ‘pushed it' in the original turn. This gesture thus, can be seen, not just as a catchment (e.g. McNeill et al 2002) creating coherence, but also as an integrated part of the syntactic structure of the turn constructional unit, and as an observable part of constructing obvious cohesion between two turns across participants' own insertion.

This is also the case in the next excerpt where the gesture again ties an ongoing turn to a prior turn after a participant's own insertion. The example demonstrates how a participant by repeating her gesture ties back to her own prior abandoned turn after having produced a self-initiated insertion, or side sequence. In this example the gesture connects specific identifiable parts of a participant's two turn constructional units. In the example the gesture locates exactly what the increment initiator ‘but' ties back to in the prior turn, i.e. the gesture locates what is being modified. At a first glance, though, there seems to be no insertion in this sequence, but simply just TCUs being produced adjacently placed.

Line 44, ‘But then it got too hot for my shoulders', is constructed as a syntactic and prosodic continuation of l. 42, ‘There wasn't a single soul but-'. When including the gesture though, the grammatical construction of the turn reveals itself to be different from what the speech alone reveals, and it becomes an example of skip-connecting back to a prior turn after using the increment initiator ‘but' and resuming the prior turn.

Fig. 6.1, l.41d                   Fig.6.2, l.42c

(Link to video clip)

Participant B produces a gesture together with the speech (l .41) while moving her arms to chest height as if she was holding on to the oars in the boat. Thereafter she tips them down fast and moves them alongside her hips exactly as she produces the word ‘slips'. Not very surprisingly, the gesture is produced synchronically with its related lexical component. In the following turn unit she produces another gesture synchronically with the related speech from l. 42 : ‘There was no one. There wasn't a single soul'. But now something interesting happens : synchronically with the word ‘but' (l. 42), she reproduces the gesture she originally produced synchronically with ‘slips' (l. 41). This gesture demonstrates a connection to the prior production of the gesture and of the prior turn. And, when accepting the synchronic production of speech and gesture as something the participants generally demonstrate a normative orientation to (Laursen, 2002a, 2002b), the gesture very clearly demonstrates to other participants exactly which prior turn ‘but' is to be understood as a modification of and thus also as an increment initiator of. The gesture is the element in the TCU that, not only as a catchment creates coherence across speech, but also constructs obvious cohesion between the two turn units.

The participant clearly shows how the TCUs are connected. The coherence doesn't just rely on the fact that the participants semantically will understand the connection between the turns : The reproduced gesture ties the turn increment initiator ‘but' to the prior turn where the gesture was originally produced, i.e. to ‘In the beginning I was just rowing in my slips' ; and ‘but-' thus projects an upcoming modification of the prior turn, and not of the turn ‘but-' is prosodically and semantically integrated in. The reproduced gesture constructs obvious cohesion between the turn units in l. 41 and l. 44, and ensures the coherence before it is verbally clear that an increment of the turn in question will follow.

So far it has been possible to observe that a repeated gesture is partly constructing obvious cohesion to prior turn, and also that a continued gesture projects that the turn is not completed yet, and that there is more to come. The difference between a "frozen" gesture and a continuously repeated gesture seems to be that the frozen gesture indicates interactional trouble, where as the continuously repeated gesture may indicate something repairable, e.g. a word search, but not interactional trouble (Laursen, 2002a ; Laursen & Stax under preparation).

4.2. Gesture as connector between turns across other participants' turn

We have seen how gesture is part of tying single words together within the same turn constructional unit, and across own insertions. The following excerpts illustrate first, how gestures is part of the tying practice of skip-connecting back to a participants own prior turn after an other-initiated side-sequence, and secondly, how the gestures project continuation, and connect to an upcoming turn. Excerpt 3 was an illustration of the tying technique skip-connecting, i.e. how a participant ties back to his/her prior so far completed turn by producing an increment. A closer examination of this excerpt reveals that we may conclude that the turn in ll. 17-19, ‘each of us has ten - has ten clips', actually couldn't have been completed at that point : Participant A has produced a gesture that is kept in position. This non-completed gesture already projects at the so-far-turn-completion that there is more to come, and it is resumed when participant A produces the turn increment in l. 19.

Fig. 3, ll. 15a-19a (link to video clip)

Participant A produces in ll. 12-15, a gesture related to the lexical component ‘bus ticket cutter' : he moves his hand up and down as if stamping a bus ticket card in the automat [cultural information : in Denmark the ticket automats on the busses cut a piece of the bus card while stamping it at the same time]. He holds the gesture throughout participant S's evaluation and resumes the gesture right after he has produced the increment initiator ‘A :nd eh ::'. The gesture is then repeated together with the continuation of the turn talk, and isn't completed until the completion of the story- ‘and when we get to the top of the ticket we die' (l. 20). Again we can see that ‘and', as increment initiator, ties the turn back to the prior turn to which it is connected (see also Local, 2004). But the gesture had already done the tying work by projecting continuation, thus constructing obvious cohesion between the so far abandoned turn and its continuation across turns and talk and time - and space ! This excerpt further exemplifies that the syntactic structure of the turn is maintained even though the lexical component is never realized : in the turn increment l. 19 there is no subject, object or verb in the speech, but they are fully present through the repeated gesture.

This gesture does even more work in the sequence. If we move a little back in the sequence we can in excerpt 7 observe how the gesture in excerpt 3 actually places this sequence and the single turn initiated l. 12 in the prior sequential environment in which the story is to be understood. As we can see in the next excerpt 7, the first time this gesture is produced, it constructs semantic cohesion with other participants' relevant prior turn long before the related lexical component and the punch line of the story is produced (excerpt 3 ll. 12-20).

Participant A initiates a gesture at l. 4, in a pause right after S has completed a turn. He moves his right hand up in front of him before he starts to say l. 5 ‘·hh So'. Exactly at the same time S starts to speak again (l.6), ‘There's a small paper on the :', and A interrupts his speech production again. It is worth noticing though, that he does not interrupt the gesture. He continues the movement and produces a gesture, moving his hand up and down. The gesture is thus realized without speech. And, more importantly we could say, it projects more to come.

It is also worth noticing that participant S treats this gesture in the same way that overlapping speech would normally be treated : in l. 8 she prolongs the sound ‘on the :', and she restarts the turn ‘On the-' exactly when she can see that A abandons the gesture and moves his hand to so-far-rest-position - i.e., this gesture is treated by other participant in the same way as she would have treated a verbally overlapping TCU.

At this point it is not obvious what the gesture relates to semantically in the speech. But because of the adjacent placement, we can assume that A's turn should be understood in connection to S' prior turn. The gesture thus relates semantically to something in the prior turn, and is at the same time projecting that the lexical component related to the gesture will also relate to the prior turn. The production of the gesture, thus, not only projects the production of a related lexical component, but, because of the gesture's relation to the prior turn, it also projects the semantic content of the upcoming turn. When participant S has completed the turn, she also demonstrates an orientation to the fact that a turn was projected : she directs her gaze from D to A in the following (0.5) pause, and A restarts his turn ‘So it's like the bus ticket-', still holding his hand in waiting position. In the following 0.3 pause he resumes the movement (see Laursen, 2001 and 2002a for gestures realized in speech pauses within turns), and reproduces the gesture at exactly the same time as he produces the word ‘cutter'. This reproduced gesture thus ties the turn and the lexical component to its original placement in the sequence, thereby showing in what connection it should be understood, and it works in the same way as (partially) repeated speech normally works as tying technique.

First, this gesture is an example of how realization of a gesture projects the production of a related lexical component and turn. And secondly, and in this context more important, the gesture demonstrates that the repeat of the gesture does turn organizational work by tying the ongoing turn back to the prior interrupted turn and to the original context in which the turn was to be understood - again across time and another participant's turn.

In the next, and final, excerpt 8, which is a continuation of excerpt 6, participant S tells how she got her suntan on a boat trip into the Australian bush. In ll. 45-47 E and S are talking simultaneously, aligning each other's turns both verbally and gesturally (I shall return to this part later). Participant R produces a compliment (l. 48) - ‘Your hands they're also just comple :tely dark' - in overlap with participant S' turn - ‘But they :: : ( ?)'. Right after turn completion she re-verbalizes the turn in the opening of l. 50, ‘You just have such a tan on your hands'. But participant S already shows an orientation to the compliment made in the overlap by looking at R and replying, in overlap, ‘you should see my back' (l. 50). This other-initiated side sequence results in participant S showing her beautifully dark brown back, before she produces a turn increment (l. 56), initiated by an increment initiator ‘·hh But my shoulders and arms and thighs'.

After participant S has shown her suntanned back, she skip-connects back to a prior turn. On behalf of the speech alone this increment is syntactically, semantically and pragmatically understood as a continuation of her reply to participant R (ll. 50-51) ‘then you should see my back'. But if we include the gestures, it turns out that the increment l. 56 ‘But my shoulders and arms and thighs' is NOT a continuation of l. 51 but of the prior interrupted turn l. 47 ‘But they :: :( ?)'.

 Fig.8.1,l.46b-52       Fig. 8.2, l.58

(link to video clip)

Participant S skip-connects, not just by reproducing the increment initiator ‘but', but also by reproducing the gesture she originally produced together with the completion of the turn l. 46, ‘then I got burnt at the s ::', and throughout the following turn, ‘I got really burnt not so much, but they :: :', before the overlap from participant R and the following side-sequence. Further, when including the gestures it turns out that throughout the insertion sequence she never completes the gesture - i.e. she already demonstrates at this point that she has NOT abandoned her ongoing turn, but will return to it at completion of the side sequence.

Just before the speech ‘I got really burnt not so much' S moves her right hand to her left shoulder and moves the hand back and forth over the shoulder. She continues the movement until the completion of R's compliment ‘your hands they're also just completely dark', where she stops the movement, but without completing the gesture. Instead she keeps her hand on the shoulder, and the gesture so to say freezes in a so-far-rest position, just as the gesture in excerpt 2. It remains frozen until participant S, occasioned by the side sequence, turns around and pulls her sweater up to show her tanned back. Immediately after this S resumes the gesture by returning the hand to the shoulder, thus returning to and resuming her previously abandoned turn both in speech and gesture.

Several things are worth noticing about this gesture. First, the gesture is not completed throughout the first side sequence. It freezes instead. And the participant only abandons the gesture when initiating what turns out to be the next side sequence, with a question ‘do you wanna see my back'. Immediately after having completed this sequence, she resumes the gesture and reproduces it together with the increment-initiator ‘but'. In this way the increment ‘but my shoulders' is tied back to the prior turn where the gesture was originally produced. Again the gesture is not only a catchment ensuring coherence, but also a part of the TCU constructing obvious cohesion with the proper prior turn.

Secondly, it is also noticeable that the frozen gesture, just like in excerpt 3, had already projected continuation (see Schegloff 1984 for gestures and projection spaces). By freezing the gesture participant S demonstrates that her turn so far isn't at all complete despite of her responding to R's evaluation, and that there's still more to come. The gesture thus ties forward to an upcoming turn, projects continuation of the verbally abandoned turn, and demonstrates that this is a side sequence. Based on the syntax, semantics and pragmatics of the speech, or turn at TALK, alone this would have been an example of participant S unproblematically abandons her ongoing turn occasioned by a comment by participant R. But when including the gestural part of the turn, it becomes clear that S treats R's comment as an interruption.

4.3. Gesture as connector between different participants' TCUs

A third noticeable thing about the gesture in excerpt 8 is that it not only ties back to a prior turn, it actually ties back to a prior third participant : if we in excerpt 8b examine what goes on just prior to the first production of this gesture, we can see that the gesture was produced as an aligning action to a third participant's gesture in the overlapping turn :

 Fig.8b,l.44-46  Fig.8b,l.46a Fig.8b,l.46b-47a

(link to video clip)

Simultaneously with the speech l.44, ‘hot for my shoulders', participant S produces a gesture that relates semantically to the shoulders where she moves both her hands back and forth over each shoulder. Participant E produces an aligning evaluation (l.45), ‘no you get burnt', together with a gesture, where she caresses her left shoulder with her right hand in the same way as we shall see participant S do right after : when E has completed her turn, speech and gesture, S interrupts her own ongoing turn, speech and gesture, ‘then I got burnt at the s :', and produces an aligning upgrade of E's aligning evaluation ‘I got really burnt'. The alignment is not only shown in the talk alone, she also repeats, or reproduces, the gesture E has just completed, thus orienting to E as recipient.

This means that not only does the secondly repeated gesture l. 58 excerpt 8 tie back to participant S' turn l. 47 excerpt 8b, it also ties back to the turn, and context, where the gesture was originally produced, i.e. the aligning action where participants S and E were co-constructing the story. When resuming the turn after the insertion sequence ll. 54-57 excerpt 8, S could have chosen to reproduce the gesture she was producing prior to E's aligning evaluation, thereby demonstrating that her aligning turn to E was also to be understood as an insertion occasioned by E's comment. But by reproducing the gesture of her aligning comment to E, and originally produced by E, participant S demonstrates that she doesn't treat E's overlapping comment as an interruption (as she does later with R's comment), she treats E's comment as a contribution to her ongoing story. Verbally S produces aligning turns to both E's and R's overlapping comments, but gesturally she treats E's comment as co-authoring the story and R's comment as an interruption of the storyline.

Excerpts 3-8 demonstrated that gestures work as tying technique at the turn constructional and turn organizational level to construct obvious cohesion between two turns of the same participant. Excerpt 8, though, was also a demonstration of the fact that these repeated gestures should not only be understood cognitively in connection with the single participants language production (McNeill et al 2002), nor only in connection with the projection space of one speaker's single TCU (Schegloff 1984). Instead they should also be seen as a device the participants in an interaction use to demonstrate connections and thereby also to "do" different kind of interactional work.

5. Discussion

Participants generally produce many gestures in story telling. Often the storytellers are producing the gestures when embodying the actions and activities of the original context. It is worth noticing how the gestures here are part of footing (see also Goodwin, 2001) - i.e. the storyteller embodies different participants from the story in the original context and is thus, through the gestures and the body posture, making the perspective of the story and who-is-acting-now apparent to the other participants. Even though these gestures also work as connectors in the local situation, the limits of this paper doesn't allow me going into further details. This paper has shown how gestures do essential work in constructing obvious cohesion between turn constructional units. The participants show an orientation to gestures being an integrated systematic part of the single turn constructional unit and of the interactional tying techniques.

Mazeland & Huiskies (2001) conclude in their paper that "[r]esumptions do non-obvious returns to what its speaker then proposes to have been a prematurely abandoned line of interaction". The examples in this paper have thoroughly demonstrated that NONE of the participants propose that their return to a prior turn is a return to a prematurely abandoned line of interaction. Through the originally continued or frozen gesture, they demonstrate, on the contrary, that this line of interaction was in no way abandoned. Even though they may have had to leave it for a while, they would definitely return to it. The gesture thus connects forward in time to a later point where a turn increment will be produced. This fact further expands Schegloff's Projection Space (Schegloff, 1984) and McNeill's Growth Points (McNeill, 2000). And just as important, we saw how the gestures tied back in time to a prior turn, thus placing the ongoing turn in its original context. And it was demonstrated that analysis without gestures provided a different grammatical analysis than when the gestures were included. The gestures were constructing obvious cohesion in all the examples. Had one not included gestures, the analysis would, for better, have been less precise, for worse, wrong. It is important to emphasize however, that the speech and the gestures do different kinds of work in the tying techniques. A frozen or repeated gesture projects THAT which an increment follows, and that to WHICH ELEMENT the increment is tied. Whereas the verbal increment initiator projects WHAT KIND of acting or relation to prior is to be expected in the upcoming increment.

Of course it will not always be the case that the participants produce gestures in their TCUs. But whether they do or not, the opposition between the citation of Mazeland & Huiskies and the analyses in this paper has demonstrated that it can be premature to conclude something general about both the interactional organization and about grammar and language on behalf of analyses of speech alone. When dealing with coherence, gestures often turn out to be the elements constructing the obvious cohesion at the local level.

The findings in this article show that gestures :

• Connect back in time across turns ;

• Project what is to come at some later point, e.g. an increment - and thus show that an apparently completed turn isn't complete at all !

• Connect OVER TIME and IN SPACE ;

• Project that an increment is to be expected - whereas the verbal increment initiator, e.g. ‘but', projects which action to be expected. Or to put this in other words : The verbal increment initiator makes the turn and action RELEVANT AND COHERENT, where as the gesture makes the turn OBVIOUS COHESIVE

I hope that a continued closer look into gesture in interaction may reveal more about the linguistic structure than any grammatical theory has provided so far.

References

Armstrong, D. F., Stokoe, W. C., & Wilcox, S. E. (1995). Gesture and the nature of language. Cambridge University Press, Cambridge.

Broe, V. (2003). Skip-connecting as a means for maintaining coherence - an aspect of the sequential organisation of talk. Acta Linguistica Hafniensia, 35, 160-185. Copenhagen : C.A. Reitzel.

Goodwin, C. (2000). Gesture, aphasia, and interaction. In D. McNeill (Ed.), Language and Gesture (pp.84-98). Cambridge : Cambridge University Press.

Goodwin, C. (2001). Tripping over footing : Stories as multi-modal interactive fields. Workshop on Narrative Cognition and Culture. Nyborg, Denmark

Goodwin, C., Goodwin, M. H., & Olsher, D. (2002). Producing sense with nonsense syllables : Turn and sequence in conversations with a man with severe aphasia. In C. E. Ford, B. A. Fox, & S. Thompson (Eds.), The Language of Turn and Sequence (Oxford Studies in Sociolinguistics) (56-80). Oxford : Oxford University Press.

Goodwin, M. H., & Goodwin, C. (1986). Gesture and coparticipation in the activity of searching for a word. Semiotica, 62-1/2, 51-75.

Ford, C. E., Fox, B. A., & Thompson, S. A. (Eds.). (2002). The Language of Turn and Sequence (Oxford Studies in Sociolinguistics). Oxford : Oxford University Press.

Gullberg, M. (1999). Gesture as a communication strategy in second language discourse. Lund : Lund University Press.

Heritage, J. (1984). Garfinkel and ethnomethodology. Cambridge, MA : Polity.

Jefferson, G. (1972). Side sequences. In : D. Sudnow (Ed.), Studies in social interaction (294-338). New York : Free Press.

Kuehn, C. (2002). Reception of gestures : Coherence and correction. Paper presented at ‘The living medium' ISGS, Austin. Retrieved, from http://www.utexas.edu/coc/cms/International_House_of_Gestures/Conferences/Proceedings/Contributions/Kuhn/Kuhn.html

Laursen, L. (2001). Gestures produced in mid-turn pauses. Paper presented at TalkBank Gesture Workshop. Carnegie Mellon, Pittsburg.

Laursen, L. (2002a). Kodeskift, Gestik og Sproglig Identitet. Unpublished doctoral dissertation, University of Southern Denmark. (English summary)

Laursen, L. (2002b). Relations between utterances. Gestures Connecting Utterances. Paper presented at Nordic Conference on Connections. Copenhagen Business School, Copenhagen.

Laursen, L. (2002c). When gestures are continued or produced after the related lexical component : Pursuit of response, and how participants orient to a gesture's completion point as the completion of the ongoing activity. Paper presented at ‘The living medium' ISGS, Austin.

Lerner, G. (2004). On the place of linguistic resources in the organization of talk-in-interaction : Grammar as action in prompting a speaker to elaborate. Research on Language and Social Interaction, 37(2), 151-184.

Local, J. (2004). Getting back to prior talk : and-u(h)m as a skip-connecting device. In : E. Couper-Kuhlen & C. Ford (Eds.), Sound patterns in interaction : cross-linguistic studies from conversation. Studies in Linguistic Typology. Amsterdam : Benjamins

Mazeland, H., & Huiskes, M. (2001). Dutch ‘BUT' as a sequential conjunction. Its use as a resumption marker. In : M. Selting & E. Couper-Kuhlen (Eds.), Studies in Interactional Linguistics (pp.141-169). Amsterdam/Philidelphia : Benjamins.

McNeill, D., Quek, F., McCullough, K.E., Duncan, S., Furuyama, N., Bryll, R., et al. (2002). Catchments, Prosody and Discourse. Gesture, 1(1), 9-33.

McNeill, D., & Duncan, S. (2000). Growth points in thinking-for-speaking. In D. McNeill (Ed.), Language and Gesture (pp. 141-161). Cambridge : Cambridge University Press.

Ochs, E., Schegloff, E. A., & Thompson, S. A. (Eds.). (1996). Interaction and grammar. Cambridge : Cambridge University Press.

Sacks, H. (1996). Lectures on conversation. Oxford : Blackwell.

Sacks, H., Schegloff, E., & Jefferson, G. (1974). A simplest systematics for the organization of turn-taking for conversation. Language, 50, 696-735.

Schegloff, E. E. (1984). On some gestures' relation to talk. In M. J. Atkinson & J. Heritage (Eds.), Structures of Social Action. Studies in Conversation Analysis (pp. 266-296). Cambridge University Press.

Schegloff, E. A. (1990). On the Organization of Sequences as a Source of "Coherence" in Talk-in-Interaction. In B. Dorval (Ed.), Conversational Organization and its Development (Vol. XXXVIII in the series "Advances in Discourse Processes"). Norwood, N. J. : Ablex Publishing Corporation.

Schegloff, E. E. (1996). Turn organization : one intersection of grammar and interaction. In E. Ochs, E. A. Schegloff & S. A. Thompson. Interaction and grammar. Cambridge : Cambridge University Press.

Schegloff, E. E. (2004). On dispensability. Research on Language and Social Interaction, 37(2), 95-149.

Selting, M., & Couper-Kuhlen, E. (2001). Studies in Interactional Linguistics. John Benjamins Publishing Company.

Steensig, J. (2001). Sprog i virkeligheden. Bidrag til en interaktionel lingvistik . Århus, Aarhus University Press.

Stivers, T. (2005). Modified repeats : One method for asserting primary rights from second position. Research on Language and Social Interaction, 38(2), 131-158.

Streeck, J. (2002). Grammars, words, and embodied meanings : On the uses and evolution of so and like. Journal of Communication, 52(3), 581-596.