Microsoft Word - Gagnon corrected proof.docx Corresponding Author: Justin Gagnon, MSc, MA Doctoral Candidate, McGill University Email: justin.gagnon@mail.mcgill.ca Journal of Applied Hermeneutics July 12, 2019 ãThe Author(s) 2019 Using Narrative Construction to Prepare the Ground for Hermeneutic Dialogue Justin Gagnon Abstract Hermeneutic research methodologists have recommended that, when conducting interviews, investigators should minimize the imposition of their pre-understandings on the conversation, as this would restrict the range of possible understandings that dialogue may produce. Instead, the phenomenon under investigation should determine the direction of conversation. This position paper argues that the iterative construction of a narrative enables increasingly focused, in-depth discussions with experts in the clearing where the interlocutors' horizons of understanding converge. Hermeneutic dialogue involves navigating one’s own pre-understanding and exploring new branches of possible understandings that emerge from conversation. This navigation de- mands reflexivity and adaptability, as well as an openness toward the complexity of a living understanding informed by a diversity of perspectives. To illustrate how narrative construction informs an increasingly refined understanding, and how this understanding subsequently frames dialogue with interview subjects, this paper draws on a historical case study of electronic medi- cal records implementation policy. Keywords Hermeneutics, qualitative research, methodology, electronic health records, research design Hermeneutics is a philosophical discipline concerned with interpretation and understanding. Making sense of others’ understandings of the world involves integrating the unfamiliar with that which is familiar to us (Gadamer, 1960/2004). That is, we make sense of our interactions through our own frame of reference. Hans-Georg Gadamer, one of the pioneers of hermeneutic philoso- Gagnon Journal of Applied Hermeneutics 2019 Article 3 2 phy, characterized the interpretive process as a fusion of horizons (Gadamer, 1960/2004). Interpretation of another, upon reading a text or in conversation, involves the meeting of two (or more) worldviews, or horizons of understanding. Our horizons are informed by our traditions, which include our contexts and prior experiences and interactions. Upon fusion of another’s horizon with our own, a shared spaced is negotiated and, through this fusion of horizons and integration of the unfamiliar, our understanding is transformed. In recent years, researchers have made strides in developing what might be called “hermeneutic research methodology” (Curtis, 2010; Moules, McCaffrey, Field, & Laing, 2015). While herme- neutics has its roots in the reading of (religious) texts, the principles of hermeneutics are particu- larly relevant to the conduct of interviews. Whether one is reading texts or dialoguing, interpreta- tion involves a back and forth between what the other has expressed (a reflection of their own understanding, put to language and rooted in their own tradition) and one’s own tradition. Dialogue, however, is bi-directional. In dialogue, a negotiation takes place whereby the interloc- utors work out a common language and frame of reference pertaining to the object of their discussion. Understanding is co-created. The interlocutors each navigate their own understanding and what was said toward a contextually contingent shared understanding. In hermeneutic dialogue, precedence is given to the object of discussion in determining the direction of conversation. Proponents of hermeneutic research recommend caution when con- ducting interviews, as one’s pre-understanding (horizon) can restrict the range of possible understandings that can emerge from conversation (Geanellos, 1999; Moules et al., 2015). However, in research where data is iteratively synthesized into a narrative, one’s pre- understanding is periodically refined and reconfigured, and inevitably influences the conversa- tion. In what follows, I argue that this is still compatible with the principles of hermeneutics, and I provide recommendations for researchers adopting a narrative approach. To provide an illustra- tion of the application of hermeneutics to the conduct of research interviews informed by an increasingly refined narrative, I will draw on a historical case study of electronic medical records (EMR) implementation policy in Quebec. Hermeneutic Interviewing When conducting interviews in a manner consistent with the principles of hermeneutics, meth- odologists recommend that researchers be mindful of how their prejudices or pre-understanding might influence the conversation (Geanellos, 1999; Moules et al., 2015). In traditional qualitative semi-structured interviews, researchers tend to deliberately direct the conversation around established questions. However, when following the principles of hermeneutics, interviews should approximate a “genuine conversation” whereby the object of discussion plays a more prominent role in determining its direction. We say ‘conduct’ a conversation, but the more genuine a conversation is, the less its con- duct lies within the will of either partner. Thus a genuine conversation is never the one we wanted to conduct. Rather it is more generally correct to say that we fall into conver- sation, or even that we become involved in it. The way one word follows another, with the conversation taking its own twists and reaching its own conclusion, may well be con- Gagnon Journal of Applied Hermeneutics 2019 Article 3 3 cluded in some way, but the partners conversing are far less than leaders of it than the led. No one knows in advance what will ‘come out’ of a conversation. (Gadamer, 1960/2004, p. 383) Certain authors, Geanellos (1999) for instance, proposed avoiding pre-determined questions entirely, as these could restrict dialogue around researchers’ pre-understandings (Geanellos, 1999). However, research interviews are inevitably initially directed by things that you know and would like to know. Instead, Moules et al. (2015) suggested having pre-determined questions, but following a more unstructured approach in the conduct of the interview (Moules et al., 2015). Accordingly, the questions are simply a resource to guide the lines of inquiry that the researcher might want to explore, depending on how the conversation goes. Research interviews […] have a structure of sorts, and more importantly a purpose. That is not to imply that the interviewer enters with an immutable list of questions, but that the structure is defined by the topic and the interviewer attempts to keep the topic central to the conversation, while “turning around” with the participant. (Moules et al., 2015, p. 89) In addition to giving precedence to the object of discussion and following the natural flow of conversation, limiting the influence of your pre-understanding also involves keeping open the range of possible understandings that can emerge. The conversation should be conducted in a way that does not simply confirm the researchers’ assumptions. Researchers should remain open to the transformative possibilities of the conversation. It belongs to every true conversation that each person opens himself up to the other, truly accepts his point of view as valid and transposes himself into the other to such an extent that he understands not the particular individual but what he says. (Gadamer, 2004, p. 387) Research interviews are influenced by numerous factors, such as: the research objective, the nature of the phenomenon, the researcher’s interviewing skills and their decision to pursue particular lines inquiry over others as the conversation unfolds. The interlocutors’ knowledge and experience, or pre-understanding, regarding the object of discussion also play an important role. Rather than following a particular formula, I contend that research interviews should be conducted in a manner that best serves the research objective and that reflecting on and acknowl- edging factors that might influence the conduct of research interviews is integral to trustworthy qualitative research. The aim of my historical case study of policy decisions surrounding EMR implementation is to map and understand policy decisions. More specifically, the aim is to provide an explanation of policy decisions in terms of actors, actions and social, historical and organizational contexts. To achieve this aim, I am constructing a narrative, through the triangulation of data from various historical documents and interviews with key experts and policy decision-makers (Creswell, 2012; Langley, 1999). The interviews will tend to involve discussions of policy processes, technological components, information infrastructure, and contextual factors. Knowledge about these aspects related to EMRs is essential to our establishing common ground and having a meaningful dialogue. I will frame my conversations around my prior knowledge about health Gagnon Journal of Applied Hermeneutics 2019 Article 3 4 technology in general and EMRs specifically, as well as institutional and policy processes. Furthermore, I will interpret the new knowledge generated from our discussion according to a pre-established map. Were I conducting conversations about patients’ experience of care, for instance, the discussions would be drastically different. One could posit that discussions around experience of care involve a broader range of possible directions. Every patient’s story involves their own context and experience, and their own way of understanding and communicating it. A common frame of reference would, therefore, need to be re-established with every interview. In all research, the process begins with an extensive review of the literature on the phenomenon under study. An effort is made to understand the available research on the topic. Researchers establish what is known and unknown and thereby refine their research objectives. In research involving interviews, a review of the literature also informs the interview guide and thus the direction of conversation. Both the researcher and interview subject bring their prior knowledge and experience to the conversation. As the subjects approached for interviews should be selected on the basis of their being knowledgeable about the object of study, both the researcher and interview subject are effectively experts on the topic, approaching it from different angles. Research prior to the conduct of interviews, therefore, enables more focused, in-depth discus- sions about the object of study. The scope of conversation is bounded by the shared horizon. In- depth discussion engendered by expertise is achieved at the cost of breadth, and breadth of understanding at the cost of detail. Methodologists have acknowledged the influence of prior research on dialogue, in the context of interviews (Moules et al., 2015). I contend, however, that important consideration should also be given to how data is integrated throughout the research process, as this can have an important impact on the content and structure of one’s pre-understanding. To my knowledge, this has not been discussed extensively in the literature. Narrative Analysis My narrative about the history of policy decisions affecting EMR adoption is iteratively con- structed. Through the reading of historical documents, I constructed an initial narrative that serves to frame my understanding. Subsequently, each instance of data collection influences and is influenced by an evolving understanding. All instances of data collection are inherently transformative. The review of every document and each interview transforms your understanding. They shape the scope and focus of the lines of inquiry pursued in conversation. In addition, each interview influences your pre-understanding for subsequent interviews. When you process your data iteratively, as in my case study of EMR implementation, rather than after all your data are collected, your pre-understanding is increasingly structured. You have more formal moments in which you integrate your new understanding with the whole. Your understanding becomes increasingly refined throughout the research process, as more data are integrated. Gagnon Journal of Applied Hermeneutics 2019 Article 3 5 When constructing a narrative, your pre-understanding, brought into each interview, is structured in a particular way. Accordingly, a narrative frames and directs the conversation in a particular way as well. Narrative configuration consists of the construction of a story from the data. In narrative con- struction, the researcher chronologically configures and links events and actions within a story, to give meaning to the data, as contributors to a particular end (Polkinghorne, 1995; Sandelowski, 1999). In [narrative configuration], the researcher’s task is to configure the data elements into a story that unites and gives meaning to the data as contributors to a goal or purpose. The analytic task requires the researcher to develop or discover a plot that displays the linkage among the data elements as parts of an unfolding temporal development culminating in the denouement. (Polkinghorne, 1995, p.15) In my research, my narrative comprises events linked to particular actors and their actions. These events will be presented alongside social, cultural, and political contexts. These contexts change over time. I will then propose possible explanations and meanings, linking decisions and actions with changes in conditions surrounding EMR adoption, in an attempt to make sense of why Quebec has had such limited success in this regard. We often make sense of happenings in the world narratively (Bruner, 1991). We tell stories. We even make sense of our own lives in a narrative fashion. Yet, narrative construction, in the context of research, has particular qualities that set it apart from narrative understanding. Our thoughts can be rather messy and are often quite abstract. Things are not always organized chronologically. When probed, we sometimes find a great number of inconsistencies in our thoughts. We have not worked everything out. The act of thinking and communicating ideas to others transforms our thoughts. When called upon to tell a story, we put our thoughts to language and give it structure. Our understanding is transformed by our telling a story. In working out our stories, inconsistencies may be made apparent, and only then do we tend to try to resolve them. Over time, after multiple iterations of re-working our narratives, our stories take on a more logical structure. They become more coherent and refined. The formal act of writing has an even greater structuring effect on our understanding. Writing is also transformative. Through writing, you produce something tangible that you can return to and reinterpret in light of having distanced yourself in time, thought, and experience. Structure is necessary as a point in a cycle of interpretation and understanding in which meaning comes together as a unit of articulation an in doing so, the newly articulated meaning becomes available in itself for questioning, reappraisal, deconstruction, and in- terpretation. (Moules et al., 2015, p. 51) Gagnon Journal of Applied Hermeneutics 2019 Article 3 6 A written narrative permits the integration of vastly more information. It allows for greater complexity, with multiple layers and divergent paths. At the same time, having something tangible permits your identifying and resolving leaps in logic, contradictions and inconsistencies. Writing better permits your working out a logic linking the data to events and actions in your story, and interpreting their meaning in relation to the plot that you have devised. When the act of writing is performed iteratively, the narrative may be increasingly refined. It enables greater coherence and complexity. In the conduct of interviews, this evolving formal narrative informs your pre-understanding. For my historical case study, I constructed an initial narrative with a reading of news articles, scientific papers, strategic plans, auditor general reports, ministry of health reports and commis- sion reports. I used the information contained in these documents to construct a timeline of factual events. These events consist primarily of policy actions and decisions, such as the crea- tion of a governing body tasked with driving and coordinating EMR infrastructure projects, or the publication of policy that sets the context for EMR adoption. At present, I have only begun conducting interviews. In my interviews, at this stage, I am primar- ily interested in establishing the key actors and organizations, and understanding their influence on EMR implementation. I am beginning to get a sense of the organizational cultures of the implicated institutions and the different political contexts, as they pertain to key policy decisions. I have also begun delineating phases in the history of EMR implementation, in terms of distin- guishable groups of actors, organizations and contexts – in terms of major shifts in preoccupa- tions and priorities. As I conduct more interviews, I will make more and more links between context, actors and actions. I will develop and put forth explanations about how and why EMR adoption in Quebec has been so challenging. The more I learn, the more questions will inevitably arise as new gaps in understanding are revealed. I will cease data collection when I will have constructed a suffi- ciently coherent narrative without any apparent, seemingly relevant gaps. I have begun the interview component of my research prioritizing experts familiar with the bigger picture, so to inform my initial narrative. Then, as I locate particularly meaningful events in the history of EMR implementation in Quebec, I will progressively narrow my focus around these key moments. I will therefore select subjects more specifically knowledgeable about these events. The data derived from each interview will fill in gaps in my understanding that emerge throughout the research process. They will thus provide greater depth to my narrative. Before discussing strategies that favour hermeneutic dialogue in the conduct of research inter- views, I will briefly summarize my argument so far. I have noted that data collection informs your understanding, and processing that data further refines your understanding. When narrative- ly configured, your understanding takes a particular form. Furthermore, formally writing a narrative makes it something tangible that favours greater complexity and coherence. This then informs your pre-understanding in the conduct of interviews. Subsequently, each interview becomes increasingly focused around filling specific gaps in understanding, answering specific questions, and corroborating or refuting specific assumptions. Gagnon Journal of Applied Hermeneutics 2019 Article 3 7 Following the principles of hermeneutics, is it a problem for one’s pre-understanding to expand and one’s line of questioning narrow throughout the research process? Is it a problem that I have a timeline of events and an increasingly coherent and complex historical narrative that delineates the space where my discussions take place? I argue that, given my research objective of con- structing a historical narrative, it is not. We always already have a horizon. We never approach a text, experience, or topic as a completely blank slate - we already have a fabric of meaning into which we accommodate, with more or less difficulty, the next new event. (Moules et al., 2015, p. 43) Your horizon directs the conversation. Following the principles of hermeneutics, however, means ensuring that you avoid, as much as possible, restricting the range of possible understand- ings that can emerge. You want to ensure that you are not simply confirming your beliefs. You want to remain open to what the other has to say and not ignore information that does not fit with what you already think you know. You want to remain open to transformation of your under- standing. I argue that to achieve this, the following should be considered. First, the conduct of interviews requires skill as well as reflexivity and creativity. Approaching a genuine conversation in the prepared ground established by your narrative pre-understanding involves allowing the object of discussion to lead and following the flow. To obtain good data […] requires a discernment and instinct around choices of which di- rection to take the conversation, which leads to follow and which to divert, which state- ments to probe further, when to engage the participants in interpreting, and ultimately keeping the topic as the focus while respecting the participants’ needs to tell their narra- tives of experience. (Moules et al., 2015, p. 89) It demands an awareness of the potential impact of your words on both the direction of conversa- tion and the boundaries of understanding. It is important to recognize when you may be imposing your understanding in a way that restricts the range of possible understandings that can emerge. Reflection upon your prejudices is recommended, and researchers should be open and transpar- ent about them in their writing. Recognition of the potential impact of one’s words is generally cultivated with experience. With each interview, through trial and error, and the practice of reflexivity, you cultivate a better sense of how your pre-understanding might influence the direction of conversation. As every situated understanding action both reveals and conceals (you cannot know what you do not know), the ideal of self-awareness is an unattainable state. Aware- ness is a continuous process that demands practice and adaptability. In addition to creativity, reflexivity and skill, I contend that researchers should embrace the complexity and ambiguity of an understanding informed by a diversity of interpretations. As more data are collected, the range of interpretation grows rather than shrinks (Curtis, 2010). The number of possible interpretive paths and interlinkages increases. The story becomes increasing- Gagnon Journal of Applied Hermeneutics 2019 Article 3 8 ly complex. The advantage of the narrative structure is that is embraces complexity and ambigui- ty by permitting the inclusion of multiple, potentially contradictory accounts. If two people have different interpretations about the factors that contributed to a certain outcome, both may be included. Then, one can reflect on the meaning of each interpretation in the context of the whole. A “true” story is not the aim. The aim is to tell a good story. As Weick (1995) suggested, a good story should be reasonable, coherent, and plausible (Weick, 1995). Understanding is finite. Research follows a certain path, limited in time, setting and scope. One cannot include every- thing of potential relevance nor reconcile differing accounts. In my case study of EMR imple- mentation in Quebec, I will not be reconciling, aggregating, averaging, or reducing people’s accounts when I consider them to meaningfully contribute to the plot. I will include contradic- tions in people’s interpretation as bifurcating branches of my narrative understanding. My research presents a living understanding rather than a timeless verity. Triangulation of data sources is valuable in this regard. Inclusion of complementary viewpoints can simultaneously strengthen particular propositions when different accounts are concordant and broaden the range of possible interpretations when accounts are irreconcilable. Each inter- pretation has value. Every interview should therefore be approached with curiosity and humility. I will now provide a concrete example from my research to illustrate how I navigated the flow of conversation, while utilizing a graphic timeline of events to prepare the ground for our discus- sion. In one of my first interviews, I spoke with a physician/researcher who had been involved in a number of health information technology initiatives over the last 30 years. I had done my research on him and had an idea about the specific moments about which he might be particular- ly knowledgeable. I adapted my interview guide to include more directed questions about his relevant experience. In my interview, after explaining my project, I began by asking my subject to speak broadly about the basis of his expertise, such as: what initiatives was he involved in? This already began generating lines of inquiry that I had not anticipated. For instance, he mentioned that he was involved in a chip card program prior to decision-makers’ formal discussions about strategic plans to digitalize the health system. I had known a bit about the chip card program, but I had not discovered in my research that he was involved. I decided to put on pause his description of his experience and focused the discussion around how the program went. Later in the interview, after we had discussed other similar initiatives, I asked my interview subject to reflect on how the projects differed in terms of how they were managed. I asked him about the different organi- zational cultures of the actors involved, and what this impact this might have had on the outcome. None of these questions had been specifically planned; they emerged from the conversation. Throughout my interview, as a prompt for our discussion, I used a graphic timeline depicting the major events that I considered relevant to the narrative. I felt that this was extremely valuable in quickly establishing common ground. I could point to two distinct events and ask if he felt there was any connection between them. For instance, in 2006, Quebec proposed a formal plan for digitalizing the health system. I asked: “what contributed to the development of this plan?” He looked at the timeline and located “Canada Health Infoway” - a federal organization charged with driving health information technology interoperability across the country. He said: “Que- bec’s plan was devised in such a way as to meet Canada Health Infoway’s conditions for receiv- Gagnon Journal of Applied Hermeneutics 2019 Article 3 9 ing funding. They didn’t devise a comprehensive digitalization plan until much later.” Thus, we uncovered an important influence explaining the state of EMRs in Quebec. Most of the remain- der of the conversation involved a deeper probe into this. Had I taken a more traditional approach to the conduct of my interviews, I suspect that I would not have pursued the new direction that emerged from our conversation. Additionally, had I not prompted the subject with my visual timeline, and established common ground and my expertise on the topic, we may not have had such a rich, focused discussion. Conclusion In sum, the iterative construction of a narrative involves the development of an increasingly coherent and complex, formally written, understanding. I began with an initial timeline, informed by the reading of historical documents. Then, in the conduct of interviews, my pre-understanding became increasingly refined and formally structured into a narrative. The conversations I had with my interview subjects were situated in the clearing where our horizons met. I presented to the interview subject what I knew, and as two experts on a topic, we had a focused discussion. As I conduct more interviews, they will become increasingly focused and directed around filling in gaps, addressing remaining questions, and reinforcing or refuting events and interpretations. I have argued that this does not inhibit my following the principles of hermeneutics. With each interview, there is a back and forth between the discussion (the part) and the grand narrative (the whole); between following the flow of conversation and making sense of what was said in relation to your pre-understanding. I concur with fellow methodologists in positing that navigating this back and forth requires reflexivity and adaptation. In addition to this, I contend that embracing the divergent accounts and interpretations that emerge in conversation is vital to conducting trustworthy hermeneutic research. References Bruner, J. (1991). The narrative construction of reality. Critical Inquiry, 18(1), 1-21. https://doi.org/10.1086/448619 Creswell, J. W. (2012). Five qualitative approaches to inquiry. Qualitative inquiry and research design: Choosing among five approaches. Thousand Oaks, CA: Sage. Curtis, B. (2010). Baudrillard and Gadamer are dead : Exploring play and the virtual through hermeneutic conversation. (Master of Arts Thesis), Montreal, QC, Canada: Concordia University. Gadamer, H-G. (1960/2004). Truth and method (3rd rev. ed; J. Weinsheimer & D.G. Marshall, Trans.). New York, NY: Continuum. Geanellos, R. (1999). Hermeneutic interviewing: An example of its development and use as research method. Contemporary Nurse, 8(2), 39-45. https://doi.org/10.5172/conu.1999.8.2.39 Gagnon Journal of Applied Hermeneutics 2019 Article 3 10 Langley, A. (1999). Strategies for theorizing from process data. Academy of Management Review, 24(4), 691-710. https://doi.org/10.5465/amr.1999.2553248 Moules, N. J., McCaffrey, G., Field, J. C., & Laing, C. M. (2015). Conducting hermeneutic research: From philosophy to practice. New York, NY: Peter Lang. Polkinghorne, D. E. (1995). Narrative configuration in qualitative analysis. International Journal of Qualitative Studies in Education, 8(1), 5-23. doi:10.1080/0951839950080103 Sandelowski, M. (1999). Time and qualitative research. Research in Nursing & Health, 22(1), 79-87. https://doi.org/10.1080/0951839950080103 Weick, K. E. (1995). Sensemaking in organizations. Thousand Oaks, CA: Sage.