Radical Insights: Towards a Critical Hermeneutic

Ethnography and “Epistemic” Violence

In Notes on the History of the Study of Indian Society and Culture (1968), anthropologist Bernard Cohn anticipated both the arguments of Michel Foucault and Edward Said in his work on representations of Indian society during the colonial era—that the vast body of knowledge produced about India was in fact integral to the military triumphs and economic conquests of the British (Lal, 2004). Cohn describes in vivid detail the many mis-translations that occurred as knowledge was gathered, ordered, classified and circulated in published reports, statistical returns and histories. Crucially, Cohn revealed how the investigative procedures deployed in part by colonial anthropologists, tried to make sense of Indian society and culture using a “British” understanding of the world. For example, when attempting to grasp the many belief systems in India, the British looked for canonical texts such as the shastras, in an attempt to construct a parallel of their own experience of religion. From this textual view, the British produced a description of Indian society that was “static, timeless and spaceless…(with) no questioning of the relationship between prescriptive normative statements derived from the texts and the actual behaviour of individuals or groups” (Cohn, 1968:8). Expanding on this argument in a later book using a Foucauldian framework, Cohn refers to this production of knowledge as a form of “epistemic violence” that dominated vernacular ways of understanding, knowing and living (1996). This notion of “epistemic violence,” is arguably the ‘ground floor’ upon which a rich body of critical discourse within postcolonial theory has been built (Nandy, 1983; Spivak, 1988).

The impact of these critiques can be seen in the many debates surrounding the role of the ethnographer as “translator.” The research practitioner is not a neutral observer but an active agent, mediating between cultural frameworks. Further, forms of representation are never naïve or innocent, but emerge from specific historical contexts (Philip, 2015). Within anthropology, there is a significant body of work that has deeply considered the many kinds of translations that occur as part of a research encounter—from the impact of tools, methodologies and interview protocols on the data gathered (Smagorinsky, 1994); to the subjective judgment of the ethnographer as an interpretive agent (Clifford, 1986); to the affordance of different mediums such as film and written text in terms of communicating experiential knowledge (Crawford & Turton, 1992; Loizos, 1993). A central question posed is – “Whose story is it, what is it being used for, what does it promise, and at whose expense?” (quoted in Schrock; Shuman, online).

Inverting the Frame

To return to historian Vinay Lal’s capitulation of Cohn’s legacy, there is a flip side to the research encounter—colonial forms of knowledge also had a tremendous effect on shaping the cultural psyche of the colonizers and their conception of modernity (Mitchell, 2000). Pace Lal on literary theorist Edward Said,

“(the) Europeans created an entire body of knowledge which enabled them to represent the Orient; such representations, moreover, far from having any necessary relationship to the Orient, reveal more about Europeans than they do about the ‘natives’ and their social institutions, cultural practices, and the like.” (2004, online)

This line of thinking shifts the focus entirely from questions of “objectivity,” to deciphering the culture responsible for the production of knowledge. In Orientalism (1978), for example, Said exerts much of his attention on the British, paying little to no attention to Indian history. Similarly, Ashis Nandy’s reading of Rudyard Kipling’s work and life is less about the production of India as a literary space in his novels, and more about the cultural dissociation that Kipling felt as an Anglo-Indian (1988). Nandy’s work in particular raises the interesting point of “hybridity”—that colonial identities are not pure, discrete units, but co-constituted in relation to their perception of the “Other” in a mutually recursive fashion. When applied to ethnographic practice, this insight makes the case that organizations have much to learn about themselves from their interpretation and understanding of cultures that are the subject of inquiry. This counter-intuitively gives research participants a form of agency in that they are able to affect the identity of partner institutions. Even as we consider this notion, it is important to heed feminist scholar Mrinalini Sinha’s note of caution, where such theoretical explorations need to be returned eventually to the “uneven organization of power.” In real world terms, it is ultimately the financial institution that sets up and manages the policies, procedures and products for its customers.

Towards a Critical Hermeneutic

Hermeneutics is a system of interpretation by which we derive meaning from a text. As a field, hermeneutics emerged from the exegesis of religious texts and can be thought of as a series of rules or principles that govern meaning. For example, a hermeneutic principle might state that the meaning of a passage is linked to another passage in which an important phrase or word appears. The field has widened significantly over time to encompass different kinds of interpretive positions and reading strategies, where the meaning of the text changes in relation to the approach. Aesop’s fable about the fox and the crow, for example, can be read literally as an encounter between an animal and a bird (which makes little sense), or more commonly using a moral hermeneutic that highlights the lesson learned from the encounter. Within market research, a similar hermeneutical approach has been applied by researchers such as Thompson and Haykto to analyze the meanings that consumers use to interpret their “experiences and conceptions of lived experience” (1997).

In contrast to literary hermeneutics outlined above, where the attention is placed on textual interpretation, critical hermeneutics focuses on a critique of the rules and principles that are used for interpretation. It argues that “uncritical acceptance of [pre-]understanding could perpetuate the exclusion of past suppressed interests” (Arnold and Fischer, 1994: 56). Simply put, hermeneutics stops at the point of applying rules and saying that knowledge and understanding is historically and socially produced. Critical hermeneutics focuses on uncovering and articulating the social and historical conditions and structures that produce meaning. The implication is that language itself is not a value neutral container—the goal is to expose power relations that produce meaning as opposed to pure interpretation (Radford, 1991).

But how does this concept apply to ethnographic research? To return to a post-colonial framework, if the reflexivity gained from the concept of “epistemic violence” alerts us to the fact that forms and representations of knowledge are never naïve or innocent, the Saidian framework provides a counter-intuitive lens to read knowledge systems as extensions of institutional epistemologies. Personas, mental models and user journeys are artifacts that signal an institution’s structure of interpretation. They also represent an idealized relationship with a particular user base—who are we building this product for, how will they use it and what kind of relationship between institution and end use will the product foster. In producing research artifacts, the researcher as “interlocutor” performs two types of interventions: firstly, in the field, the researcher recognizes the subjective voice of participants as collaborators in defining problems and co-designing outcomes, while acknowledging the asymmetrical power dynamic of the relationship. And secondly, the researcher is able to provide a critical intervention that shifts the burden of ethnographic engagements from constructing an objective description of the world, to helping institutions come to terms with their own structures of interpretation that can be adjusted. It is by putting both parts together that we arrive at a critical hermeneutic.

Institutional Myths

The start of a client engagement is a crucial time for a project, from building a rapport with internal teams, to getting a deeper understanding (and sometimes even modifying) core objectives, to parsing the specific interests of individual stakeholders. From the perspective of developing a critical hermeneutic, it is an opportunity to surface the epistemological assumptions that are driving research. Drawing on the work of sociologist Michel Callon and anthropologist Michael Silverstein, Rita Denny’s work on the relationship between theory and practice makes the important point that assumptions are embedded within an institution’s linguistic practices where “terms reflect the theoretical stances of those involved” (2013:142). As Denny illustrates with her cases studies, the inability to fully grasp the linguistic practices of an institution or discipline, can have catastrophic results. Her point is not to regard language as a proxy for the world, but to critically examine the interplay between terminology and practice. To return to our case study, the financial institution’s existing personas were a valuable starting point to learn about the institution’s assumptions through their vocabulary.

Shared Value System – The personas were organized based on demographic segmentation, but on a more abstract level, were largely aligned to a common financial value system (e.g., savings is important, managing my money in the present will pay off in the future). In effect, the financial institution was projecting a value system across the persona set that supported institutionally approved goals. Just as in the case of Barthes’ conception of a “myth,” the inherent value system of the personas had become naturalized to the extent that some of our clients were only marginally aware of the embedded ideology.

Data Driven Behavior – The primary mechanism for learning about each persona was a brief introductory paragraph that comprised a few biographical background details. A majority of the content was dedicated to data-driven lists: financial products that the persona either currently had or would potentially be interested in (e.g., loans/credit cards) and technology/devices used for transactions (e.g., computer vs. mobile). These data modules were interspersed with call-outs for opportunities to provide advice and guidance related to specific life events, e.g., a student loan.

The dangers of data-driven personas in terms of caricaturing and over-simplifying the idiosyncratic behaviors and needs of people has been well established (Rode, Baur & Cox, 2009). But it is John Sherry’s critique that pinpoints our most pressing concern with the data modules, where collapsing personas and market segments results in the representation of “consumer(s) whose only dimension of real interest is their relationship to our brand” (2007:22). Sherry’s observation intersects with Bernard Cohn’s insight regarding a crucial epistemological assumption on the part of the British: that Indian society could be known and represented as a series of facts and administrative power stemmed from the efficient use of these facts (Cohn, 1996). A simple substitution of key terms reveals the persistence of the same epistemological assumption on the part of our stakeholders: that culture can ultimately be represented as “data,” which in turn can be leveraged to ensure marketability and profitability.

The problem with this particular conception of data is that it severely impairs the scope of questions that internal teams can ask when developing a product/interface. There is no need to understand or dive deeper into user motivations, because all action can be represented as a decision tree that typically targets the most profitable segments and behaviors. The presence of “data” exerts an “objective” frame that precludes the act of thinking about and truly connecting with the experience of end users. In a very practical sense, this means missed opportunities both in terms of developing a better product for unanticipated needs as well as missing segments that do not neatly fit the quantified criteria. In a highly competitive and crowded domain, the cost of getting things wrong could very well be catastrophic. My point is not that quantitative data is inherently bad. Rather, quantitative data should supplement a deeper understanding of customer behavior, not detract from it by focusing exclusively on profitability. Despite the many shortcomings that have been identified in personas, from the tendency to focus on individuals over collectives (Sherry, 2007; Flynn 2009), fetishize consumers (Arnould & Cayla, 2013) and their inability to develop new approaches (Regan & Revels, 2007), as design tools, personas can help product teams to gain insight into why people think, feel and behave in different ways.

Our challenge in developing the new foundational persona set would be threefold: firstly, to test the institutional financial value system; secondly, to potentially develop an epistemological core that would serve as the basis of the new persona set; and finally, to produce richer, more complicated representations that require a deeper engagement with end-users as people (instead of data points).

Designing the Research

The study was conducted with participants who were customers of the financial institution across 3 geographic markets. The participant criteria were defined as part of a workshop and spanned several financial behaviors, channel usage, account and other product related characteristics and a broad demographic mix in terms of age, gender, occupation and education. The study was completed in multiple phases:

Phase I – Diarists used a paper/mobile diary to report their financial activities over the course of a month. In addition to capturing their financial moments, the diarists also completed a workbook with exercises that explored key dimensions of their financial lives. The research team scheduled weekly calls with the participants to reduce the risk of attrition over the course of the month. These calls had the benefit of building rapport and trust with the cohort over time—we gradually became intimately involved with details of our participant’s lives beyond just their finances.

Phase II – A third of the diarists were selected for the next round of research comprising a 2-hour contextual interview at a location that was most relevant to their financial lives. We met most of the participants at their homes, but several interviews were conducted at their work place. The research interviews were divided between two teams, each led by a researcher with two accompanying stakeholders. This level of involvement in the research interviews was a key determinant in terms of the acceptance of the insights as stakeholders were able to narrate user stories from firsthand experience. As part of the interviews, we encouraged our stakeholders to ask open-ended questions to make the process more inclusive. The interviews were conducted using a semi-structured protocol that explored aspects of the workbook that were not covered during the weekly research calls. In addition to discussing their habits and routines, the interviews also included:

• A long history of each participant’s financial life, where they described a narrative arc of financial events or milestones that changed their behavior or outlook. Although this exercise was peripheral to the client’s objectives, the goal was to arrive at a historical and contextual understanding of difference around financial behaviors.
• A co-creation exercise where participants mapped out their financial ecosystem including accounts, institutions, people and technologies.

Disrupting Myths, Building “Empathy”

After each field visit we conducted a daylong workshop with our stakeholders to share research data and insights. As an agency, much to the credit of the internal researchers, we were fortunate to have a range of participants across roles and positions across the organization’s hierarchy. The workshops also served as interim milestones for iterations of the personas and allowed us to test epistemological groupings across markets. One of the main objectives during this time was to build “empathy” by both exposing and disrupting the embedded mythology of the existing foundational personas. Empathy is generally understood as “a holistic understanding about the users” (quoted in Loi, 2009; Mattelmäki 2003) or the “capacity to understand and represent the fuller complexity of human beings” (Burton, Henry & McColgin, 2012). The ability to understand or feel the end user’s condition leads to the development of more meaningful experiences and products. As Sohrab Vossoughi, the President and Founder of the design agency Ziba, puts it:

“(W)hat makes a design good is not a beautiful image or an intricate model, but the ability to put yourself in the shoes of the person who’s going to use it, and the organization that’s going to produce it…Handling the technical details of design is now relatively easy, but delivering an integrated, meaningful experience is exceptionally hard.” (online, 2014)

The emphasis on the end user experience is a step in the right direction, but to think about “empathy” within the framework of critical hermeneutics changes the conception of the tem considerably. Rather than focus exclusively outwards on the end user, empathy is an opportunity for self-awareness and reflexivity in terms of better understanding our own epistemological moorings (and those of the stakeholder). The “humanizing” process works both ways—an encounter with difference ultimately reveals our own biases and motivations.

Research Data Walls – Using large foam core boards, we designed a science fair approach to prominently display participant data (workbooks, images, quotes, etc.) throughout the day. At the start of each workshop, stakeholders were given dialectical prompts to explore the boards, for example, which participant were you most like, growing up? which one would you like to be in the future? The prompts were extremely effective in terms of motivating and synchronizing the focus of our stakeholders, while producing a moment of suture and identification with the research participants. At the start of each workshop, attendees would take turns to respond to the prompts with the names of two participants the reason why they chose them. The process was both confessional and therapeutic with some uncomfortable moments as stakeholders began to narrate financial values that were clearly not aligned with institutional expectations. The end effect was the humanizing of both our research participants as well as the workshop participants who were able to step outside of their roles and “see” the institutional value systems that they were embedded within. The surfacing of these values was arguably a pivotal step towards recognizing different financial outlooks.

Participant Deep Dives – To deepen the connection with the research subjects and expose the epistemological assumptions embedded in the old personas, we spent a considerable amount of time on “deep dives” of individual participants. These were strategically chosen after each round of research to expose the group to the broadest variety of financial circumstances and behaviors possible. For each deep dive, we began with a long history of the participant’s financial history using video clips to convey specific life events or behaviors that were particularly poignant, often calling on stakeholders to tell part of the story. In addition to photographic and video documentation, we also represented a core financial characteristic of each participant as a visual model.

The “immersive” first-hand encounter with the data produced similar effect on our client’s as described by Cotton (2012), with an emotional arc that gradually moved from discomfort to self-reflection as the research data eroded the logic of the previous model. Language played an important role as we began to introduce key insights that directly challenges the epistemic beliefs of our clients using their own vocabulary as a set up to our alternative framework. But it was the experience of having the conversations off-site in a space that was not their regular work environment that aided with the free-flowing dialog and epiphanies. A unifying characteristic across all of our stakeholders regardless of their role was the high level of financial knowledge that comes with the domain of working at a financial institution. Because our research emphasized on the long history of each participant’s financial lives, stakeholders began to pick up and discuss the detailed narratives that clearly illustrated the impact of socio-economics and other factors such as role models in terms of shaping an individual’s financial outlook and knowledge. It was unequivocally clear from participant ecosystem maps that there were no clear patterns in terms of technology preferences. Even though this insight directly contradicted the initial hypothesis of our stakeholders, it the widespread visibility of multiple media forms across presentations became a pivotal point of discussion. Even as we led the workshop and framed the research data, the space for conversation between participants was crucial in terms of both surfacing and working through epistemological assumptions.

(Re)Building the Personas

Having disrupted initial assumptions about the structural basis of the new personas, the research team turned to the materials collected during field research to identify other dimensions that might serve as epistemic cores. We began by first making a list of financial attributes and dimensions that were expressed within the existing persona set. This was followed by parsing the research data to validate the dimensions from the existing persona set, while adding other dimensions that emerged from research. As part of the process of building and vetting the list, we experimented with several clusters to see if any patterns emerged in terms of grouping participants together.

Beyond developing the epistemological cores, much of the latter stage of our engagement was dedicated to creating representations of each persona that would effectively communicate the internal value system while avoiding the trappings of reification. Instead of using a data driven model that validated the overall demand for a particular device/product, our goal was to give teams enough information to understand how a persona might use a product if they chose to. By “complicating” the representation of data, our objective was to turn questions of marketability (is this a segment we should be building this product for?) into a space for speculation and exploration (what need would this product solve?).