Abstract
We examine the ways in which members of a small group coordinate their memories, bodies and language in a functional and goal-oriented manner when they are co-designing their dream house and then collaborative remembering that previous interactive encounter. Our analyses show the following: (1) participants structured collaborative design and collaborative remembering sessions in different ways (e.g. linear and sequential vs iterative and hierarchically structured, respectively); (2) higher degrees of knowledge building were temporally synchronized with higher degrees of interactivity during both tasks; (3) collaborative remembering did not only follow the spatial structure of successive elements of the dream-house design session, but it was also proceeded by associations between semantic elements of the discourse; and (4) participants collaboratively remember better what initially generated most joint activity during collaborative design. This research thus contributes to understanding of collaborative remembering processes with respect to a knowledge-rich collaborative task.
Introduction
Over the past 20 years, there has been a shift of emphasis in design studies research, from individual to collaborative design (CD). Empirical research has been carried out in several application domains (e.g. Détienne, 2006; Stempfle and Badke-Schaub, 2002) in order to understand relations between the design processes, the collaborative processes and the resulting design product. CD relies on the designers’ interactive reflection on the design in real time, by going backwards and forwards during the design process. Consensus, for example, with respect to the planning of future tasks, often relies on joint remembering of previous CD work and project milestones. The coordination of individual and shared representations of each individual designer (i.e. ‘grounding’: Clark, 1996; Clark and Brennan, 1991) plays a central role in the achievement of shared goals (Badke-Schaub et al., 2007). Such consensus is crucial when elaborating solutions and constructing common ground by design teams (e.g. Détienne et al., 2012; Hébert, 2012). The processes of CD are temporally distributed activities, with new phases of projects being built upon previous ones (e.g. Wiltschnig et al., 2013). This means that at almost any point in the CD process, it will be necessary to successfully recall relevant aspects of previous phases of the project. Even when external memory devices are available for recording and representing CD decisions and their ‘rationales’ (e.g. Vyas et al., 2013), the fact that these representations are necessarily selective means that all future contexts of their use cannot be anticipated. In that case, designers working in teams will sometimes need to reconstruct the contextual meaning of previous states of the project – whether external representations are available or not – in on-the-fly situated interactions with other team members (Bietti and Sutton, 2015; Bietti et al., 2016). Thus, a deeper understanding of how collaborative remembering (CR) occurs within the interactive unfolding of CD projects is of theoretical and practical importance.
This study expands previous research done on interactional and multimodal aspects of CR in small groups (Cienki et al., 2014; Goodwin, 1987; Hutchins, 1995; Wu et al., 2008) by focusing on how it occurs in a simulated work setting, in this case the design studio. So far, little attention has been paid to the influence that the interplay of multiple behavioral channels has in CD and CR creative tasks in small groups. This exploratory laboratory study examines the ways in which members of a small group coordinate body, verbal and material resources, in a functional and goal-oriented manner when they are co-designing their shared ‘dream house’, and this is carried over into CR of the same CD task the day after. Analysis of the two sessions – CD and CR – was oriented towards answering the following two questions, combining discursive and co-verbal behavioral dimensions in each case:
What are the relations between the three dimensions (cognitive, linguistic, behavioral) during CD and CR? Are there regularities in relations between specific clusters of behaviors and CD and CR?
What are the general relations between the second CR session and the first CD session? What is remembered, and to what extent how?
Having a better insight into how CR of previous CD from an interactional and multimodal perspective works in a simulated design setting in the laboratory is the first step towards a better understanding of how social interaction and multimodal communication affect memory processes in collaborative creative work.
In order to look for possible answers for these two main general research questions, the analysis is structured in the following fashion. First, we analyze the temporal relations with regards to the kind of information recalled by comparing CD to CR sessions. Second, we examine the degree of knowledge building during CD and CR, taking into consideration the specificities of each collaborative task. Third, we look at the temporal distribution of knowledge building in relation to (a) degree of knowledge building, (b) rates of coordination of co-verbal behaviors and (c) rates in the degree of interactivity in CD and CR. Then, we select illustrative examples, one taken from CD and other from CR, and provide a micro-analysis of the coordination of verbal, bodily and material resources. Afterwards, we discuss our main findings and present the conclusions. In the following section, we provide a brief review of the literature on the study of CR in conversational contexts from a multimodal perspective.
CR in multimodal social interactions
Remembering together involves re-evoking a shared or partially shared past, distributed among interacting partners (Bietti, 2012b; Hirst and Echterhoff, 2012; Sutton et al., 2010). CR is a form of collective action based on pragmatic goals and social roles, which are dependent on the social and conversational context in which the action occurs (Hirst et al., 1997). CR can be considered as an interactional practice grounded in and bounded to the communicative actions produced by the participants engaged in social interaction, their sequential organization, co-option and pragmatics (Middleton and Brown, 2005: 99).
Conversations about the past experiences are one way in which people develop shared memories of the past (Bietti, 2012b, 2014; Hirst and Echterhoff, 2012; Middleton and Brown, 2005). Conversations are joint activities (Clark, 1996; Clark and Brennan, 1991) in which partners have to ‘share or synchronize aspects of their private mental states and act together in the world’ (Brennan et al., 2010: 304). These types of coordination dynamics can occur without conscious awareness (Dale et al., 2013; Rączaszek-Leonardi and Cowley, 2012). The mechanisms involved when individuals are engaged in conversations often include the coordination of linguistic and bodily resources (e.g. gesture, gaze, posture and facial expressions) in order to achieve shared goals (e.g. to try jointly remember where we left the car parked). For example, studying the interactive functions of head movements like nodding during storytelling, a human activity which joint remembering uses to a great extent, has shown the ways in which head movements act as embodied resources for back channelling, that is as a mode to signal the degree of the listener’s attention to the story being narrated (Bangerter and Clark, 2003). But remembering with other people goes beyond the coordination of verbal and non-verbal resources in synchronized fashion over time. At some point during the joint activity, interactants have to collaboratively create a shared account or expression of the past in order to actually remember together, as distinct from doing something else (e.g. imagining together). This means that CR is not only about how this specific kind of social interaction is coordinated and unfolds over time but also about the collaborative processes that, for example, enable interactants to cue each other’s memories (e.g. Harris et al., 2011; Meade et al., 2009).
Most studies on CR in conversations (Bietti, 2010, 2012a; Brown and Reavey, 2015; Buchanan and Middleton, 1995; Edwards and Middleton, 1986; Middleton, 1997; Middleton and Edwards, 1990) were undertaken using recordings and transcriptions of naturalistic social interactions (e.g. reminiscence groups in residential homes and day-care centres, teamwork in intensive care units and group conversations about the past historical events).It is only relatively recently that research has applied observational methods (e.g. participant observation, interviewing and artefact analysis) to the investigation of CR in multimodal collaborative activities (e.g. O’Hara et al., 2012; Wu et al., 2008).
In order to examine the ways in which older adults cope with cognitive decline in natural settings, cognitive scientists (Dahlbäck et al., 2013) have conducted field work in home healthcare services. These studies show that CR depends on the interanimation ‘of a complex distributed web of processes involving both internal or intracranial and external sources’ (Dahlbäck et al., 2013: 1), which are tightly constrained by the dynamics of the activities in which individuals are involved. Thus, naturally occurring situations should be taken as the loci for the analysis of CR. In a related project, computer scientists (e.g. Wu et al., 2008) have investigated the cognitive strategies that families create and use in order to struggle with amnesia in real-world activities. This study involved the recruitment of 10 families, some of whose members had severe memory problems, and explored the communicative strategies used to compensate for the memory impairment of one of their members, including the use of technological devices (e.g. calendars, personal digital assistants (PDAs) and journals) as well as discursive practices. The general aim of the investigation was to show how by means of ‘distributed cognitive processes across participants and technological devices families may work as cognitive systems coping with amnesia’ (Wu et al., 2008: 833).
The advantages of observational research as a method of inquiry for memory research are multiple due to the fact that it may provide the proper analytical tool for exploring the multimodal and cognitive dimensions of CR in real-world settings.However, despite the fact these ecologically valid studies sustain the view that remembering is an embodied, distributed and situated practice, which should be examined in real-world activities, these studies have not accounted in detail for the central role that embodied resources (e.g. gestures and eye-gaze) play in conversational remembering.
Human cognitive activity is linked to high-level cognitive processes by way of embodied interaction with culturally organized material and social world (Hutchins, 2010: 712), and a detailed description of these embodied interactions during CR in conversations is essential (Bietti, 2015). A better understanding of the multimodal nature of CR activities, therefore, is crucial if we agree with Kendon’s (1986) point regarding the key role that gesticulation plays in everyday communication: I believe gesticulation arises as an integral part of an individual’s communicative effort and that, furthermore, it has a direct role to play in this process. Gesticulation is often an important component of the utterance unit produced, in the sense that the utterance unit cannot be fully comprehended unless its gestural component is taken into consideration. In many instances it can be shown that the gesticulatory component has a complementary relationship to what is encoded in words, so that the full significance of the utterance can only be grasped if both words and gesture are taken into account. (p. 12; emphasis in original)
During CR, deictic gesturing (e.g. hand pointing) allows speakers to establish the focus of visual attention on sources of relevant information, thereby externally grounding their individual and shared memories (Cienki et al., 2014). Gesturing while speaking seems to improve learning (e.g. Goldin-Meadow et al., 2001) and performance in simultaneous memorization tasks (e.g. Stevanoni and Salmon, 2005). Thus, representational gestures, on the other hand, may be deployed for aiding autobiographical memory retrieval but can also be used communicatively to evoke a shared conception of a particular feature of an event among interlocutors (Bietti, 2015). Gaze coordination can be deployed to regulate the distribution of interactional turns, for instance, in order to reinforce the accountability of the recipient to respond to a question acting as a reminder or to go on with the description of the events being collaboratively recalled (Bietti and Galiana Castelló, 2013; Cienki et al., 2014; Goodwin, 1987). Looking away from the interactional participants, furthermore, has been associated with individual semantic retrieval, as well as display of uncertainty (Goodwin, 1987).
We believe that a substantial part of the CR in complex collaborative activities takes place outside the verbal domain. The aim of this study is to examine the interplay between the behavioral, cognitive and discursive dimensions of CR, a knowledge-rich collaborative task in a small group.
Method
Case studies
To our knowledge, little or no research has been carried out to date on multimodal and embodied group processes of CR of groups’ prior complex collaborative activities such as CD. It is therefore at a too early stage of research to be able to isolate relevant variables to be tested and to formulate well-founded hypotheses (precisely, as we discuss at the end of this article, this research has enabled the formulation of such interesting testable hypotheses for future research). We therefore chose the case study method (Yin, 2009) as a means of in-depth study of the complex factors and processes at work in collaborative design and remembering. As Yin (2009) points out, this method is also the most appropriate when considering ‘how?’ and ‘why?’ questions, such as, in our case, ‘how (i.e. by what processes) do groups co-elaborate shared memories of their previous work?’ and ‘why do groups remember certain aspects of their previous work rather than others?’.
The video data used for the analyses come from a European project on CD and CR in small groups, the initial aims of which were to develop methods for capturing data and for analyzing it, within a ‘temporal dynamics of collaborative remembering collaborative design’ problématique.
Procedure
The study was divided into two sessions. In session 1, the group were asked to CD their ‘dream house’ (a holiday home) on the basis of written instructions, without financial limits but which they could only visit together. The house could be located anywhere in the world. Participants had a total time of 45 minutes to perform the joint task, and they had the opportunity to ask clarification questions concerning the task to be carried out to the experimenter. We selected the ‘dream house’ CD as the task for this exploratory study because its topic (e.g. designing a shared-living environment) fitted in a larger research project currently ongoing at Telecom ParisTech on the design of shared housing and sustainability. Participants were asked to write a list of the main features of the house that they agreed on through discussion and interaction (e.g. number of rooms, façade, interior, exterior, location in the world) on a separate sheet of paper while working on the design. They were told to hand the list of the main features of the house to the experimenter at the end of the task. To complete the task, participants were given a sheet of A3 paper and three pens with different colors, depending on each of the participants. In session 2 which occurred the following day, the group was asked to collaboratively remember the joint creativity task in a face-to-face interaction with the design and list of main features that they agreed on the day before.
Participants had a total time of 45 minutes to perform the CR task.The experimenter made available the list with the main features of their dream house that they had agreed on the day before, as well as drawings. With respect to each item in their list, participants were instructed to make an attempt to reconstruct how they had come to make the decisions they did (i.e. insisting on remembering the CD processes as well as the product).
Participants
For this exploratory study, one group of three participants was recruited. The group consisted of three doctoral students at the Department of Economic and Social Sciences, Telecom ParisTech (two males, aged 26–28 years, and a female, aged 26 years). Participants were native speakers of French, and their participation in the study was voluntary.
Materials
We obtained multi-angle video recordings of face-to-face interactions, including static (a fixed camera in front and the participants, CANON LEGRIA HF M46) as well as dynamic perspectives (head-mounted wearable video cameras, MOUVICAM). The audio and video recordings were transcribed in detail using specialized software (ChronoViZ: see Fouse et al., 2011). Audio recordings were also transcribed in accordance with the Jefferson Transcription System 1 (Jefferson, 2004).
Analysis
CD session was longer than the CR: the former [00: 26:36.14] whereas the latter [00:20:41.15]. The difference was related to the fact that the participants did not have time constraints and in addition they were allowed to interact freely without interventions from the researchers, asking to stop when they thought they had finished. For both sessions, when the participants believed that they had completed the task, they informed the researcher in charge of this, and the sessions were over.
The first type of analysis we conducted was into the temporal relations between elements in CD and CR. In order to do so, we measured what type of information was introduced and elaborated in CD and how it was recalled in CR. This was based on a post hoc identification of the main elements of the dream house actually imagined and their properties. Looking at the temporal relations between elements of CD and CR enabled us to gain a better insight into how the participants associated elements of the design over time and how the different structures and dynamics of each task (e.g. writing a list of the main features of the house that they agreed on during CD vs relatively unconstrained CR the day after) affected such associations.
The second kind of analysis we performed was about the degree of knowledge building (e.g. Scardamalia and Bereiter, 2003) during CD and CR, taking into consideration the specificities of each collaborative task. Scardamalia and Bereiter (2003) define ‘knowledge building’ as ‘the production and continual improvement of ideas of value to a community, through means that increase the likelihood that what the community accomplishes will be greater than the sum of individual contributions and part of broader cultural efforts’ (p. 1371). In the field of collaborative learning that focuses on detailed qualitative analysis of interactions between students, Baker (1994, 1995) has described fine-grained ‘knowledge elaboration operators’ that function in dialogue (on the ‘content elaboration’ level) as a means for co-constructing problem solutions. They describe the way in which each participant in the collaboration elaborates the solution elements proposed by others.
In order to measure the degree of knowledge building in CD and CR, we used the following procedure: based on the categories that we had identified for the previous analysis, we counted the number of attributes added to each of the categories during CD as well as CR. Table 1 provides an illustrative example of the procedure applied to measure the degree of knowledge building in both tasks.
Example: degree of knowledge building.
We based our work to measure the degree of knowledge building in CD and CR on this procedure. In addition, we analyzed how categories were elaborated and remembered in temporal terms during both tasks. Paying particular attention to how participants build knowledge in CR about the previous CD task enabled us to shed light on the semantic structure of group reminiscing. Succinctly, this is what type of semantic association characterized CR in comparison with CD.
In the third kind of analysis we conducted, we compared (a) the temporal distribution of knowledge building in relation to the number of attributed added to each of the categories examined above with (b) the rates of use of co-verbal behaviors and (c) the rates in the degree of interactivity in CD and CR. The minimum unit of time used for coding of the video and audio recording was 100 ms. For all time points, a binary value was assigned to each behavior for each participant, indicating whether he or she performed that behavior at that particular moment in time. The co-verbal behaviours that we coded were four: (1) representational gestures, (2) pointing gesture, (3) nodding and (4) manual tracing.Making correlations between (a), (b) and (c) was crucial to investigate how social interaction influenced group reminiscing in CR.
Based on possible temporal correlations between (a), (b) and (c), in the fourth kind of analysis we performed, we analyzed two examples from a multimodal perspective, one from CD and the other from CR. The examples correspond to the sequences in which the synchronization between degrees of content elaboration, rates of interactivity and behavioral use reach their highest points. For these particular micro-analyses, we included gaze direction as co-verbal behavior as it plays a central role to establish a shared focus of visual attention between participants. The micro-analyses illustrated how multimodal social interactions affect CD and CR in real time as these rich and complex collaborative processes unfold.
Temporal relations and type of information recalled
When comparing temporal relations between categories of the dream house in CD and CR, we found important differences. During CR, our analysis showed that participants structured CD activity in a linear and sequential fashion, that is, they presented new categories (characteristics of their dream house), elaborated with a good degree of detail, in terms of the number of attributes added to a specific category, and then moved on in order to introduce a new category and go on with a similar pattern (e.g. {cat. 1} => {cat. 2} => {cat. 3}). On the other hand, during CR, participants structured the activity in a spiral and iterative fashion. In CR, we could observe that participants often recalled specific features of one category (e.g. sea side {location}) that were semantically associated with the features of another category (e.g. sea side {location}=> kitchen with sea view {rooms}), meaning that the addition of new attributes for each of the specific categories the participant collaboratively recalled occurred across features of different categories not within the same categories as we had observed during CD (e.g. {cat. 1}/{cat. 3} => {cat. 2}/{cat. 1}/{cat. 3}). This suggests that CR, in this case, does not only follow the spatial structure of successive elements, and sub-elements of the dream-house design, but that it also proceeds in a cognitive and linguistic manner, that is, by association between semantic elements of the discourse (van Dijk and Kintsch, 1983).
Nonetheless, despite these differences in the ways of structuring both activities, we found that the most elaborated categories during CD, for example, {rooms}, {location} and {swimming pool}, were the best-remembered ones during CR, considering the amount of information remembered and not memory accuracy. The analysis also showed that some of the categories less elaborated in CD were omitted during CR, for example, {WIFI} and {sauna}. However, this did not occur for all the less discussed categories during CD, for example, {tennis court} and {garden}. This could possibly be explained by the associative process mentioned above: for example, {swimming pool} = {place for sport} => {tennis court}.
Degree of knowledge building
This analysis indicated that some of the categories omitted during collaborative recall did not appear in the list (see Figure 1), for example, {WIFI}, and that to large degree the order of the categories recalled followed the spatial organization of the written categories on the sheet of A3 paper (see Figure 1). Here, we found that the temporal dynamics of CD, as well as of CR, were determined by the specific tasks the participants were asked to perform and by the physical affordances provided by the setting (e.g. participants seated around a table) from a lesser to a greater extent. In contrast to what occurs in CD, in which writing patterns shape the temporal dynamics of the creative task, the dynamics of CR seemed to be partly determined by the spatial organization of the categories along with their attributes on the sheet of A3 paper that participant C used to write them down throughout the CD session, as well as the participants’ reading patterns.
List of categories written by participant C and the order in which they were recalled during CR.
These factors seemed to have determined the structure as well as the degree of knowledge building (or group remising) in CR. Temporal sequential organization of CD as well as CR can be observed in the written list of categories and attributes above, and the order in which they were remembered the day after.
Temporal distribution
We observed that the interactional sequences in which the participants elaborated the categories they most proposed (as features of the dream house), in terms of the numbers of attributes added to each of the specific categories in CD and CR, were correlated to those sequences in which both more bodily resources (e.g. manual gesture and nodding) were used by the participants, and the exchanges of speech turns across participants occurred at a faster pace. This suggests participants collaboratively remembered better those creativity moments when they were more jointly involved in elaborating the features of their dream house. Based on these first indications between possible temporal correlations between (a) the temporal distribution of knowledge building in relation to the number of attributed added to each of the categories examined above with (b) the rates of use of co-verbal behaviors and (c) the rates in the degree of interactivity in CD and CR, we analyzed two examples from a multimodal perspective, one from CD and the other from CR. These examples correspond to the sequences in which the degress of synchronization between (a), (b) and (c) reached their highest points. For these particular micro-analyses, we included gaze direction as co-verbal behavior as it plays a central role to establish a shared focus of visual attention between participants. 2
Session 1: CD
In the next interactional sequence, participants discussed about the kitchen that their dream house should have:
Qualitative micro-analysis during collaborative design.
As Figure 2 shows, we divided the interactional sequence into four different phases, each of which corresponds to a different time (1–4) moving through the sequence description. In t (1), participant C was writing down the list of the features that the dream house should have, whereas A and B were focusing their visual attention on C’s act of writing (Figure 2a). C’s two-handed manual gesture in Figure 2b was temporally coordinated with C’s polar question (yes or no) as to whether the kitchen must have an ilot central (‘central island’) (line 8). C’s polar question operated as a reformulation of A’s original proposal the line before (line 6) regarding the fact that their dream house should have a cuisine americaine (‘American kitchen’). C’s two-handed manual gesture was behaviorally synchronized with a redirection of eye-gaze towards A and B. As for C, their polar question, manual gesture and eye-gaze were bodily resources that provoked a response from the listeners in the interaction (Rossano et al., 2009). C’s eye-gaze reinforced the accountability of A and B to respond to his question and confirms that they were all referring to the same kind of kitchen design. Between lines 6 and 8, we can observe how A and C entered into common ground as to the kitchen type that they would like to have in their dream house. However, in t (2), this sort of common ground between C and A seemed to be undermined by B’s proposal (line 9). In Figure 2c, we can notice participant A changing eye-gaze direction towards B, when she claimed that the kitchen could have a grand table au milieu (‘a big table in the middle’) which ruled out the proposal to have an American kitchen (line 6), which implies having a kitchen with a central island (line 8).
In the next line, A re-used the lexical items and syntactic structure that B employed the line before. In doing so, C seemed to agree with B’s proposal. However, his redirection of eye-gaze towards C may have indicated the seeking for his agreement too (Figure 2d) as C was the person responsible for taking notes about the features of their dream house. A’s redirection of gaze towards C was temporally coordinated with C’s facial expression of hesitation as he seemed not completely convinced about the fact that the kitchen could have both, a central island and a table in the middle. Hence, he asked for a clarification while stressing the lexical item milieu (‘middle’) at the end. C’s facial expression of hesitation along with the clarification question led to a new phase of the interactional sequence. In t (3) (Figure 2e and f), C and A tried to reach an agreement about the final design of the kitchen, considering the possibility of it having both a central island and a table in the middle. First, A emphasized that his idea was to have an open kitchen (line 13). Second, and based on the assumption of having an open space assigned for the kitchen, he further elaborated the design using both hands to perform a representational tracing gesture and employing the surface of the table as a surrogate for the large space that the kitchen would occupy in their dream house (Figure 2e). Having the kitchen located in an open space gave the possibility of counting on both a central island and big table in the middle. In line 14, C performed a full-handed pointing gesture temporally synchronized with the adverb deictic là (‘there’) the reference of which laid within the imaginary space created by A using the table as surrogate surface (Figure 2f). In the interactional sequence depicted in t (3), participants A, B and C focused their visual attention on the imaginary space created by A when he began explaining the idea of an open space kitchen that he had in mind. Despite A’s efforts to clarify by means of verbal and co-verbal resources his idea of having both a central island and a table in the middle, and the sharedness of a joint focus of visual attention on the surrogate and imaginary space, C did not seem to be fully persuaded by A’s explanation. Hence, in t (4), he handed over a sheet of A3 paper to C in order for him to draw the design of the open kitchen (Line 17, Figure 2g) that he had in mind and had tried to bodily explain using the table as surrogate surface less than 6 seconds before. The final outcome of A’s efforts to explain the design and the idea of an open space kitchen that could include both a central island and table in the middle are reflected in his sketch.
Manual gestures operated to bodily represent the space that the kitchen (see Figure 3) would occupy in their dream house. However, these manual gestures did not occur in a vacuum. They were environmentally coupled (Goodwin, 2007) using the table, around which the participants were seated and having it as the centre of their interactional space, a surrogate surface that acted as a reference point and target of a shared focus of visual attention between the participants. However, we noticed that the environmentally coupled gestures performed by A did not seem completely effective to convince C about the design of the kitchen. At this point, the strategy that C found to get a better and more comprehensive explanation of the possible design of the house was to give a sheet of A3 paper to A and request him to draw the design. The sketch of the kitchen design drawn by A appeared to have solved the problem as the participants moved on to discuss a new feature of their dream house. This example illustrated how the persistence of a graphical representation, namely, the sketch, in time and space may have provided better affordances for explaining the kitchen design compared to the environmentally coupled gestures performed a few seconds before.
Sketch of the kitchen.
The participants used the sketch in the CR session the following day. In the next extract taken from the CR session, we analyzed the ways in which participants A, B and C used the sketch drawn by C to collaboratively recall the main features of the dream house that they had agreed on the day before.
Session 2: CR
Qualitative micro-analysis during collaborative remembering.
In Figure 4, from t (1) to t (3), B used multiple verbal and co-verbal resources to locate the terrace of her dream house. In doing so, she employed the sketch drawn by A as a reference point the day before. In Figure 4a, B performed a full hand pointing gesture that ended up with her touching the sketch, temporally coordinated with the deictic of space ici (‘here’). B’s pointing’s target was the item drawn on the left side of the sketch marked by the letter ‘T’, which was closer to her sitting position around the interactional space, created by the physical affordances provided by the table and the interacting participants. According to Mondada (2009), ‘interactional spaces are actively and constantly shaped and sustained by the participants’ bodies, glances and gestures during the interaction’ (p. 1979). In the sketch, the letter ‘T’ stood for ‘terrace’, that is, the feature of the house they were trying to collaboratively remember the location of which they had agreed on the day before. In t (1) (Figure 4b and c), C aimed at correcting B’s memories regarding the location of the terrace using the deictic adverb ici (‘here’), which was then followed by the conjunction ou (‘or’) and another deictic adverb là (‘there’) (Line 3). C’s use of spatial deictics was temporally coordinated with full-handed pointing gestures touching and using the sketch as reference point. As one may expect, the target of C’s first pointing gesture synchronized with ici (‘here’) was the side of the sketch closer to his position in the interactional space, whereas the target of the almost simultaneous second pointing gesture, touching the sketch, was the right side of the drawing, closer to participant A. It is important to note how B’s and C’s pointing gesture were temporally coordinated in Figure 4b. In t (1), B’s and C’s pointing gestures touching the drawing could be thought of as ‘directing-to’ action (Clark, 2003), the aim of which was to try to move the addressees’ attention to the sketch in order to create a shared focus of visual attention upon which the grounding of collaborative recall could occur. In t (2), B’s pointing gestures (Figure 4d and e) had a different function within the interactional space and the joint activity of CR. B’s pointing gestures were temporally synchronized with personal pronouns in the second person to refer to the other participants in the group. These manual gestures were also coordinated with B’s redirection of eye-gaze towards A and C, instead of the sketch. These two bodily behaviors (pointing and eye-gaze) along with the personal pronouns in the second person singular form were embodied resources that B employed to recall which of the participants had proposed specific features of their dream house during CD the day before.
In t (3), B performed a tracing gesture, touching the sketch, in order to show a possible flaw in the kitchen design (Line 5, Figure 4f). B’s tracing gesture also worked to redirect participants A’s and C’s gazes towards the sketch. In the next line, C claimed that the possible flaw that B had noticed before had already been solved by the design itself. He did that while performing a pointing gesture and touching the part of the sketch that solved the possible flaw pointed out by B (Figure 4g). Almost immediately, B nodded while redirecting her gaze towards the sketch, and thus, she agreed with C’s resolution of the potential conflict (Line 7). Thus, the interplay of C’s memories of the previous CD activity along with the sketch acting as an external representation supporting remembering seemed to have provided sufficient arguments to solve the design problem posed by B in line 5.
Gestures environmentally coupled with the sketch operated a common point of reference for collaboratively remembering the location of other features of the house (e.g. terrace). Such common points of reference, along with the use of pointing gestures and personal pronouns in the second person singular form, allowed the participants to recall who had been responsible for proposing several of the features that were listed in the list that they had to write down the day before. In addition, participants’ pointing to the sketch of the kitchen helped them to create a shared focus of visual of attention, that enabled them to collaboratively remember the complete design of the their dream house beyond the kitchen. Thus, the sketch of the kitchen operated as an external representation that grounded the overall collaborative activity and acted as a memory cue for collaborative recall through the interplay of verbal, bodily and material resources in the design of the house as a whole.
Discussion
Although this exploratory study was not intended to be representative of a population and was used as illustration for understanding the phenomenon in question, our results suggest that a naturalistic way to organize CR of previous CD work in actual teams of designers may be structured not in the temporal relations regarding previous design decision-making processes but rather on semantic associations between the items involved in those processes. In our exploratory study, participants structured CR in an iterative and hierarchical fashion as a sort of network of sematically associated categories, whereas in CD they organized the collaborative activity in a linear and sequential fashion by introducing new categories, elaborated them and then moved on in order to present a new category and continue with a similar pattern.
The differences in the ways of elaborating contents in CD and CR may have been related to the tasks that the participants were asked to carry out during the study. In CD, writing patterns while making the list shaped the temporal dynamics of the task by determining how participants built up their knowledge on their dream house. Another question is that the spatial organization of the categories along with their attributes on the sheet of A3 paper as well as the participants’ reading patterns partly affected group reminiscing in CR. The list with the main features of the dream house as well as the sketch of the kitchen operated, first, as tools for problem-solving and explanation during CD and, second, as a reduced and imperfect model of the entire design acting as a memory cue to ground CR of the entire design, not just the design part that they depicted. Because of the methodological limitations of the study design, it would not be possible to claim that the participants would have remembered less if the list and/or sketch were not in front of them or available to them.
Inter-discursive operations (Baker, 2015) produced by the participants in dialogue along with the turn-taking structure given by the reciprocity of roles (e.g. speaker-addressee), their alternation over time and the expectation of an immediate response (Levinson, 2006) enabled them to transform, refine, extend and elaborate the solutions elements proposed by others in both CD and subsequent CR. As we could observe in the illustrative examples we selected for the detailed micro-analyses, inter-discursive operations realized by semantic connectives (e.g. opposition and difference: mais (‘but’) and en fait (‘in fact’); explanation: c’est-à-dire (‘which means to say’); causation: c’est pour ça (‘that’s why’); addition: et (‘and’); and correction: ou (‘or’)) shaped discourse coherence in both tasks. In CD, however, discourse coherence was also determined by the structure of the task itself, that is, how the participants jointly agreed on the design of the dream house, and wrote a list of the main features of the house and through discussion and interaction arrived at an agreement (e.g. number of rooms, façade, interior, exterior, location in the world) on a separate sheet of paper while working on the design. This means that CD was governed by syntagmatic relations between components of the design (e.g. {cat. 1, attributes 1, 2, 3} => {cat. 2, attributes 1, 2, 3} => {cat. 3, attributes 1, 2, 3}), given by the spatial structure of successive elements and sub-elements of the dream-house design.
This is in contrast with CR, in which discourse coherence was in part facilitated by external resources such as the list and the sketch of the kitchen but mostly relied on the turn-taking structure of conversations and relatively unconstrained semantic relations among elements of utterances at a local level. Not having to follow instructions set down beforehand (apart from remembering together the design they had created the day before) allowed participants to organize group remembering by means of paradigmatic relations among elements of the design (e.g. {cat. 1, attribute 3}/{cat. 3, attribute 3, 2} => {cat. 2, attribute 2, 3} /{cat. 1, attribute 3}/{cat. 3, attribute 3, 1}). This preliminary finding suggests that CR of a simulated complex-rich collaborative activity such as CD may not only rely on linear semantic associations over time, as CR unfold in conversation, but also on complex webs of mutual dependencies among associated concepts for which interaction and recurrence are crucial. These – the latter – are of great importance for establishing common ground between interacting partners during collaborative recall (Bietti and Galiana Castelló, 2013; Bietti et al., 2016; Harris et al., 2011; Meade et al., 2009).
The fact that higher degrees of content elaboration co-occurred with higher degrees of interactivity during both tasks suggests that the participants’ behavioral engagement was crucial for CD and subsequent CR. Differences in the participants’ behavioral engagement with the task may have been motivated by various reasons, from moments of low and high joint attention to instances in which they were more emotionally engaged with the task (it could have been more fun to CD the swimming pool rather than the bathrooms). What is crucial and should be taken into consideration for future studies was that participants collaboratively remember better what initially generated most joint activity during CD. Future studies on CR and CD tasks should include further behavioral measurements such as changes in body posture (e.g. leaning back vs leaning forward) and laughter as a means to estimate attentional spans and amusement during CD and how they may affect subsequent CR.
The results of this exploratory study, should they be validated on a broader corpus, would have important implications for extended creative CD projects. Such projects involve successive meetings, with the necessity to collaboratively remember and co-construct meanings of previous CD decisions. Finally, the work presented here opens new avenues of research in memory studies, given that, to our knowledge, the role of collaborative creativity and CD has been little studied to date.
Footnotes
Acknowledgements
The authors thank Chloé Le Bail, Samuel Goeta and François Huguet for their participation in our study. They thank two anonymous reviewers for their invaluable feedback on an earlier version of this manuscript.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: We gratefully acknowledge the support of the European Commission (P7-PEOPLE-2012-IEFMC-IEF 326885).
Notes
Author biographies
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
