• No results found

6 D ISCUSSION

6.3 I MPLICATIONS

This work has several implications. First, I have demonstrated how a combination of conversation analysis and Clark’s (1996) theory of common ground can uncover details in a team’s practice, that are difficult to describe or inaccessible to other analytic methods. Studying teamwork ethnographically (Pelikan et al., in press), me and my co-authors noted struggles with common ground due to difficulties with situation awareness but we were unable to describe how this plays out in detail. Using conversation analysis to study how teams build common ground turn-by-turn (using the example of surgeon-initiated requests for action) enabled me to develop a deep understanding of the difficulties that teams face. Prior work noted differences in communication patterns between robotic and laparoscopic surgery (Cunningham et al., 2013; Lai & Entin, 2005; Nyssen & Blavier, 2009) but was struggling to illustrate underlying reasons. For instance, Nyssen and Blavier (2009) counted acts of communication and found significantly more verbal acts in robotic as compared to laparoscopic surgery. However, they remained speculative on why verbal interaction increases during robotic surgery. Employing the method of conversation analysis, I was able to show the kind of verbal utterances that participants produce and it becomes clear that the observed increase in verbal communication is partly due to the fact that participants need to produce more affirmative responses.

Further, trouble with fulfillment that needs to be negotiated verbally significantly adds to the number of verbal acts in robotic surgery; especially when comparing it to scenes such as illustrated in Excerpt 14, in which participants address difficulties through embodied means, only producing a few verbal utterances to underline embodied actions. Conversation analysis proved to be a helpful method to deepen the understanding of why verbal interaction increases in robotic surgery and why participants struggle with common ground. It may be helpful for studying how technology shapes teamwork in other settings, enlarging the understanding that has been developed with the help of quantitative or ethnographic methods. Further, this thesis proves that a combination of conversation analysis and Clark’s (1996) theories on collaborative language use can yield interesting findings. Koschmann and LeBaron (2003) reported difficulties in combining conversation analysis with Clark’s (1996) theories, indicating that categorization of each utterance in a transcript into either presentation or acceptance phase is difficult. I have not applied Clark’s theories at such a detailed level but rather used Clark’s (1996) notion of common ground and action ladders to explain conversation analytic findings. While both approaches use different analytic vocabulary and are distinct in their treatment of mental states, they do not conflict in this analysis but rather produce fruitful results. This thesis shows that a combination of two different approaches that originate in sociology (conversation analysis) and psychology (Clark’s theories) can yield interesting insights when combined carefully on a meta-analytic level.

Second, this work extends the understanding of requests in surgery and more generally, in institutional settings. The body of research on requests in surgery is quite small (Bezemer et al., 2011; Mondada, 2014a, 2014b; Randell et al., 2017) and robotic surgery has not been studied at this level of detail so far. Randell and colleagues (2017) are the only ones who studied requests in robotic surgery and they only provide some initial pointers as part of a more general assessment of robotic surgery. Further, Randell and colleagues (2017) did not compare data from open and robotic surgery with the same surgeon. By doing so, I was able to show that the responses to requests in robotic surgery differ significantly from those in open surgery. Further, this work does not only contribute a systematic comparison of requests in open versus robotic surgery, but it also shows how institutional practice for requesting is mediated by technology. Employing a different strategy of responding to requests in robotic surgery, one may argue that participants adapt their interaction to the particular affordances of the technology. Humans have been shown to adapt to robot interaction partners (Pelikan & Broth, 2016) and employ different requests strategies in different institutional settings (Antaki & Kent, 2012; Bowles, 2006; Curl & Drew, 2008; Kuroshima, 2010; Lee, 2011; Raymond, 2014). This thesis shows that humans do not only employ different strategies to respond to requests in different institutional settings (e.g. doctor-patient interaction, vs. ordering in a restaurant, vs. calling an airline service hotline), but also develop different strategies within the same setting (surgery) as they adapt to the specific technology which mediates interaction in the particular institutional setting. This suggests that request practices may change with technology that shapes the context in which they are occurring. The most important implication lies in the finding that struggles with common ground in robotic surgery are of an extent that is so far only known from collaboration over distance. This has interesting implications for designers of interaction and collaboration technology. The Da Vinci system was initially built for telesurgery, and designers were aiming to enable surgeons to operate from a distance. This initial design goal becomes visible in how teams interact with the robot and it considerably shapes collaboration in the team. As I have shown, the robotic surgery team faces similar struggles as distant teams, even though the surgeon is in the same room. This highlights the importance of the scenario that designers use to imagine how the technology will be used in the future: As my analysis shows, the

initial scenario comes back in the design of technology and dramatically shapes interaction between users of the technology. Writing scenarios for how a technology will be used may sometimes seem like a chore thatkeeps design teams from getting started with the actual work. However, as this thesis shows, it is an important step and designers should carefully think about how the technology will be used in the future, as they have the power to shape how users interact with it. A decade ago, the robot manufacturer Intuitive Surgical Inc was actively engaged in optimizing the Da Vinci robot for telesurgery applications (Garcia et al., 2009; Nguan et al., 2008). Today, the company states on its website “The surgeon performing the procedure is located at a console in the operating room in close proximity to the patient and surgical support staff. Remote surgery, or telesurgery is not a focus of Intuitive’s product design” (Intuitive Surgical, 2018b). If this is true, Intuitive Surgical should adjust to the actual collocated usage of the system, optimizing the design for surgeons controlling the robot in the same room as the patient is located.