Making sense of citizen science: Stories as a hermeneutic resource

Energy Research & Social Science xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Energy Research & Social Science journal homepage: www.elsevier.com/locate/erss

Original research article

Making sense of citizen science: Stories as a hermeneutic resource Gwen Ottinger Department of Politics and Center for Science, Technology, and Society, Drexel University, 3101 Market St., Suite 250, Room 200, Philadelphia, PA 19104, United States

A R T I C L E I N F O

A B S T R A C T

Keywords: Environmental justice Citizen science Big data Storytelling Epistemic injustice

In communities on the front lines of energy facilities, citizen science has been seen as potent tool for environmental justice, giving residents access to quantitative data and, as a result, greater credibility with regulators and other experts. However, as in other realms of energy and environmental policy, greater access to data brings with it increased interpretive challenges—challenges which are especially acute for environmental justice-oriented citizen scientists seeking alternatives to scientists’ frameworks for understanding pollution and environmental health. Drawing on Miranda Fricker’s (2007) theory of epistemic injustice, this paper shows that frontline communities’ struggles to understand air quality data manifest “hermeneutic injustices,” or inequities in resources for meaning-making. Following research showing storytelling as one vehicle for making meaning, it argues that the stories told by frontline communities—stories of harms to health, systemic danger, dissembling, and disrespect—can in some circumstances serve as a crucial hermeneutic resource for making sense of air quality data for which scientific frameworks are inadequate. At the same time, it documents the limits of stories in giving meaning to data, pointing to areas of “narrative mismatch” which call for further hermeneutic invention by community groups working in collaboration with sympathetic scientists.

1. Introduction Access to data is often seen as a key ingredient in environmental protection. Our expanding ability to generate and process data in what some have called the age of “big data” is seen as having the potential, for example, to reduce household energy consumption [1] and address major conservation challenges [2]. Simultaneously, absences of data, theorized as “knowledge gaps” and “undone science,” have been argued to be major obstacles to environmental social movements’ ability to effect change [3,4]. Interest in data is especially intense in U.S. communities on the front lines of energy production, from sites of extraction, such as unconventional natural gas drilling operations, to “midstream” facilities like oil refineries and processing plants for natural gas liquids, to sites of electricity generation, including coal, nuclear, and waste-to-energy plants. In many such communities, concerned citizens groups and allied environmental non-profits have confronted gaps in relevant data by collecting their own, using a combination of standardized and invented instruments [5–7,57]. Frequently, these community-led data-collection efforts have been accompanied by calls for additional monitoring of energy facilities by responsible authorities [6,8]. Across these arenas, the expansion of data brings with it the problem of interpretation. The meaning of data is underdetermined; different conclusions can be drawn from the same data depending on what frameworks or stories are used to interpret it [9]. For communities on

the front lines of energy production, making sense of expanding data—even data they themselves collect—is especially challenging. The added difficulties they face stem from fundamental disconnects between community and expert ways of knowing environmental hazards [10]; see also [11,12]: “frontline communities” ask questions that aren’t being asked by regulatory scientists; they assert the relevance of factors that aren’t represented in standard scientific paradigms; they call for different standards of proof. Their citizen science efforts—which this paper will take to include both data collection and sense-making efforts around publicly available data—thus exist in large part to offer alternatives to hegemonic scientific practices that do not adequately represent community experiences. But citizen scientists who wish not to adopt experts’ interpretive frameworks confront a problem: how should they make meaning of their data, and make it meaningful to the regulators and others whom they want to persuade to take action? In this paper, I ask whether stories told by frontline communities in the United States about their experiences living in close proximity to petrochemical polluters—sometimes mere blocks away; in every case close enough to feel directly affected by pollution—can be a resource for interpreting data in counter-hegemonic ways. Stories have been shown to play a role in making meaning of energy and environmental data in policy settings [9]. Storytelling is also widely recognized as a powerful core strategy of frontline communities in the environmental justice (EJ) movement. Yet scholars studying citizen engagement with data in EJ settings have thus far not focused on the role that story-

E-mail address: [email protected]. http://dx.doi.org/10.1016/j.erss.2017.06.014 Received 17 November 2016; Received in revised form 1 June 2017; Accepted 4 June 2017 2214-6296/ © 2017 Elsevier Ltd. All rights reserved.

Please cite this article as: Ottinger, G., Energy Research & Social Science (2017), http://dx.doi.org/10.1016/j.erss.2017.06.014

Energy Research & Social Science xxx (xxxx) xxx–xxx

G. Ottinger

injustices occur when individuals from structurally marginalized groups are wronged in their capacity as knowers. Fricker suggests that epistemic injustices take two major forms: testimonial and hermeneutic injustice. Testimonial injustice occurs when a person’s statement or account is dismissed because of the person’s (structurally disadvantaged) identity: she is female, for example, or from a racial minority group. Hermeneutic injustice describes inequities in epistemic resources—the concepts, language, and frameworks that we use to understand situated experiences and render them visible and comprehensible in the public discourse [13,17,18]. Epistemic resources are shared throughout a society, but marginalized groups, more often than dominant groups, experience a tension or mismatch between their lived experiences and the resources available to talk about them [18]. Further, their efforts to conceptualize them in a way that can be “heard” in the dominant culture are likely to be met with microaggressive challenges to the categories or experiences of the marginalized group [19]. Fricker [13] gives the example of the difficulty that women had speaking of the coercive nature of sexual advances by supervisors in the workplace as recently as the 1970s. In this as in other cases where frameworks for making experience comprehensible to oneself and others are inadequate, a new category (“sexual harassment”) had to be created before the phenomenon could be taken seriously in the dominant culture. Because of the structural authority afforded to science and scientists in environmental politics, it is useful to think of “layperson” or “nonscientist” as a marginalized identity category in the context of epistemic injustice. Fricker [13] has little to say about the categories of “expert” and “lay” and in fact cites specialized knowledge as a permissible reason for valuing one person’s testimony over another’s. However, research in social studies of science has shown that distinctions made between laypeople and experts hinge on far more than whose knowledge is most accurate [20–22]. Women and people of color face significant challenges in establishing identities as “scientist” or “engineer” (e.g. [23,24]), and social class may also play a significant role in whose testimony is valued in scientific practice [25]. Working class whites may thus also be seen as potential victims of epistemic injustice, to the extent that their status as “laypeople” results in their not being taken seriously as knowers. Theories of environmental justice have recognized that the concept, as articulated by EJ activists, has multiple facets. Schlosberg [26] identifies four: distributive justice, procedural justice, recognition, and capabilities. However, case studies, especially of frontline communities’ engagements with science, show that EJ activists also attack epistemic injustices, which are rife in environmental justice conflicts. Testimonial injustices are most overt. Cole and Foster [27], for example, describe Kettleman City, California, activists’ accounts of odors and illnesses related to pollution from a toxic waste dump being dismissed because they were Latino and, in many cases, not native English speakers. Similarly, former Love Canal, New York, resident and founder of the USbased Center for Health, Environment, and Justice, Lois Gibbs, describes how a public official refused to take her seriously when she alleged that her children were sickened by toxic waste buried under a playground. He branded her a “hysterical housewife”—an explicitly gendered denigration of her testimony [28]. Frontline community members’ testimony is also frequently dismissed on the grounds that their comments are scientifically inaccurate or irrelevant [29,30], a move that pre-emptively denies the possibility that laypeople’s local knowledge or ways of knowing can contribute to collective understanding. Hermeneutic injustices also pervade environmental justice controversies. Perhaps the most striking example are the resources for making meaning of community exposures to chemicals. Quantitative risk assessment, a technique that uses toxicological data to produce probabilistic measures of increased risk of disease as a result of exposure, dominates regulatory responses to hazards in frontline communities. Environmental justice activists and social scientists writing about the movement have critiqued the framework, on two broad

telling might play in helping to make meaning—including meanings that challenge expert interpretations—of that data. Using a framework of “epistemic injustice” [13] to conceptualize the work done by citizen science in environmental justice campaigns, I show that stories can be a powerful interpretive resource for frontline communities looking to leverage data into advocacy, even in cases where robust scientific frameworks for understanding the data are elusive. Drawing on ethnographic research among community air monitoring advocates, I show how residents of frontline communities have used their pre-existing narratives of living next to refineries and other petrochemical facilities to make sense of, and make claims about, the results of citizen science projects. In a number of cases, I show, combining stories and data has been a very successful strategy: stories supported with data—or, to think of it another way, data situated within stories—have in some cases compelled action to address local environmental hazards. At the same time, however, I demonstrate that pre-existing stories are not a sufficient means of overcoming the interpretive gaps faced by frontline communities. I point to a series of “narrative mismatches,” in which stories told about local conditions don’t map on to available data. In situations of narrative mismatch, I argue, communities may be unable to mobilize information that could help to demonstrate the harms they suffer. Overcoming interpretive gaps in this area will require more active building of interpretive resources, including through collaboration with sympathetic scientists. In the next (second) section, I introduce the idea of epistemic injustice and its major variants: testimonial injustice, or unfair attributions of credibility to speakers, and hermeneutic injustice, or structural disadvantages in access to resources for sense-making. I explain how epistemic injustices affect communities suffering from environmental injustices, and how frontline communities’ citizen science efforts represent a strategy both for bolstering residents’ credibility in the eyes of regulators and other experts and—more tenuously—for making sense of environmental hazards in a way that offers an alternative to expert interpretations. In Section 3, I discuss narrative as a tool of meaningmaking, both for frontline communities trying to understand and communicate about their experiences, and for policy actors trying to contextualize complex, incomplete, and/or uncertain science. The use of story in science-rich policy environments, I suggest, suggests a likely parallel in the EJ movement, although the EJ literature has not made explicit connections between stories and data. Sections 4–6 respectively introduce the case of community-based air monitoring at energy and petrochemical facilities, describe my ethnographic methods for studying EJ-based advocacy for increased air monitoring, and catalog four kinds of stories common in frontline communities: stories of harms to health, stories of systemic danger, stories of dissembling, and stories of disrespect. Section 7 documents the power of these stories as an interpretive resource—and partial remedy for hermeneutic injustice—through examples of communities using them to make sense of, and make politically powerful, data that resisted interpretation using scientific frameworks. Section 8 shows their limits, describing a number of narrative mismatches that have stymied community groups trying to mobilize potentially powerful data. I conclude by suggesting that the problem of narrative mismatch is not limited to frontline communities but extends across energy and climate policy arenas. Making sense of increasingly voluminous data may require innovations in the stories we tell about them (c.f. [9,14]—innovations that, I argue, are best pursued in collaborations that include community members and/or other so-called “lay” citizens working alongside scientists willing to go beyond existing interpretive frames to look for meaning that better represents the concerns and experiences of frontline communities (c.f. [15,16]). 2. Epistemic injustice and citizen science As described by philosopher Miranda Fricker [13], epistemic 2

Energy Research & Social Science xxx (xxxx) xxx–xxx

G. Ottinger

much longer term exposures. This practice has long been criticized by regulators and industry scientists for being an “apples to oranges” comparison. However, I argue [12] that activists’ approach should instead be seen as a strategic refusal of experts’ distinctions between chronic and acute exposure, to call attention to harms suffered by residents as a result of semi-regular, sub-acute releases that fit neither category. In this one intervention, we can clearly see activists trying to use citizen science data to address hermeneutic injustices. At the same time, the fact that activists continue to (mis)use experts’ screening levels to contextualize their air monitoring data points to the larger interpretive challenges they face. Experts’ frameworks tend not to do justice to their experience, but few other resources for interpreting quantitative data exist, making it difficult for frontline communities to communicate their data in a way that is meaningful in the public discourse.

grounds. Risk assessments often rely on assumptions about environmental exposures that only hold for white, middle-class lifestyles, neglecting, for example, exposure to mercury as a result of subsistence fishing in immigrant communities or ceremonial uses of waterways by Native American tribes [31,16], and thus resulting in consequential inaccuracies. In addition, EJ scholars and activists find risk assessment’s underlying worldview to be problematic, in that it reduces health harms to probabilistic measures of disease incidence, presumes that risks can be known and controlled, and institutionalizes “acceptable” levels of risk in place of taking a precautionary approach [32]. Risk assessment thus tends to exclude communities’ understandings and experiences of hazards and harms, and silence alternative ways of conceiving of health and quality of life [33,34]. As a resource for meaning-making, risk assessment does at times work to the advantage of frontline communities by enabling regulators to assemble complicated, heterogeneous data into a persuasive argument that environmental remediation is necessary to protect community health (e.g. [16]). The hermeneutic injustice lies in the fact that risk assessment, with all of its biases and partiality, so dominates efforts to understand the relationship between pollution and health. Communities challenging risk assessments have been successful, to a large extent, in getting experts to reconsider their assumptions and incorporate communities’ local knowledge—e.g. of fish consumption rates—into the existing frameworks, resulting in more accurate risk assessments [16]. However, attempts to create alternative conceptions of environmental hazards and their effects on communities, such as shifting to a “health and well-being” paradigm, have met significant institutional resistance [34]. Further, many communities continue to struggle to articulate alternative ways of understanding the harms of pollution [58], and may find themselves falling back on expert-derived frameworks that fail to fully express their concerns [35]. The resources available for persuasively interpreting data are thus not attuned to the perspectives and experience of frontline communities; I argue that this mismatch constitutes a hermeneutic injustice. Citizen science can be seen as a response to epistemic injustice in frontline communities. Community-driven data-collection efforts in particular, which have included air monitoring with homemade “buckets” [7,12], water monitoring in areas affected by natural gas drilling [5], and health surveys designed and conducted by community members [10], confront testimonial injustice by giving marginalized speakers quantitative information to bolster their credibility when talking to experts who might otherwise be quick to dismiss what they have to say on the basis of who they are. While data alone is no guarantee that regulators will accept community members’ accounts, community-generated data does appear to make U.S. government agencies more likely to collect additional data or take enforcement action [10,36]. One striking example was recounted by former Global Community Monitor (GCM) organizer Ruth Breech at the Community-based Science for Action (CBSfA) conference in New Orleans in 2014. Some years before, Breech had worked with a group of women in Cincinnati, Ohio. They claimed that a plastics plant, located in close proximity to an elementary school, was releasing dangerous levels of toxic chemicals. Company officials and regulators at first labeled them “crazy grandmothers” and “crazy mothers,” according to Breech, and refused to act. Then the women took air samples and found high levels of chemicals manufactured by the plant that they had been complaining about. After seeing their results, regulators agreed to conduct additional monitoring and, ultimately, helped see to it that the elementary school was relocated (author’s fieldnotes, 15 November 2014). Whereas citizen science meets testimonial injustices head-on, its relationship to hermeneutic injustice is more complicated. The very nature of the data collected by frontline communities may represent a challenge to experts’ frameworks for conceptualizing environmental hazards. For example, activists from frontline communities routinely compare results from short-term air samples to screening levels for

3. Storytelling in environmental advocacy Quantitative data, including that generated by community members themselves, thus often fall into a kind of hermeneutic gap for frontline communities. The frameworks for making sense of the data offered by scientists tend to be inadequate to community members’ experiences, but communities often find it challenging to invent new categories and languages to make data meaningful, especially ones that are recognizable to experts. Social science research on narrative practices in both environmental policy and the environmental justice movement suggest that storytelling may be one way that environmental justice activists can fill the hermeneutic gap around data. Storytelling has been shown to be an important tool for meaningmaking across a wide range of issues [37–39], including environmental policy [9,14] and environmental justice [40,41]. The practice of telling stories involves selecting characters, describing a setting, communicating causal relationships between events, and choosing starting and ending points [42]. Through their choices, storytellers implicitly create a moral landscape, distinguishing heroes and villains, good actions and bad ones, desirable and undesirable futures [43]. Stories thus assign responsibility in particular ways, enabling some actors to be called to account and others to escape sanction [39,44]. They also justify past actions [38] and make the case for next steps [9,42], in part by warning of the consequences if appropriate actions are not taken [14]. In environmental justice realm, storytelling has been an especially important tool for meaning-making. One goal of EJ activism has been to enhance the ability of marginalized communities, including frontline communities, to speak for themselves [45]. In that context, storytelling is not only a powerful way for community members’ voices to be heard [27] but also a mechanism for weaving together individuals’ heterogeneous experiences into collective knowledge about a place, its history, and the threats posed to it by pollution or industrial development [40]. Further, to a much greater extent than in policy realms, where elites control the stories, EJ storytelling gives community groups a way to refuse dominant narratives about them, and advance their own, alternative understandings of their communities, how they’ve been treated, what is owed to them by regulators and other dominant groups, and what their future should look like [40,41]. Although the goals of citizen science in frontline communities are similar—to give voice to local knowledge and offer alternatives to experts’ claims about environmental quality—the literature on the EJ movement has thus far not theorized the potential relationship between storytelling and engagements with quantitative data, in keeping with larger cultural assumptions that stories and science belong to different realms [43]. Most studies of frontline communities’ citizen science projects have examined them as interventions in technoscientific practice, comparing their standards, purposes, and epistemologies to experts’ science (e.g. [10,16,46,12]), with a few acknowledging that there may be “plural logics” for citizen science, including education and organizing [5,7]. Questions of how frotline communities may be 3

Energy Research & Social Science xxx (xxxx) xxx–xxx

G. Ottinger

around the world [36], in no small part due to the dissemination efforts of two northern California-based non-profits: Communities for a Better Environment (CBE) and Global Community Monitor (GCM). The buckets spread along with an organizing model (the “bucket brigade”) and guidelines for using the devices: the several-minute samples are taken when chemical odors in the community reach an intensity that residents judge to be a 7 on a 10-point scale. Over time, these organizations, along with others in the network, began to experiment with other community-friendly sampling methods and devices, including MiniVol™ monitors for measuring ambient air levels of particulate matter and “wipe sampling,” a technique to collect soot deposited on windowsills or other surfaces so that it can be analyzed by a laboratory for heavy metals and polyaromatic hydrocarbons (PAHs). The second technological development to come out of the 1994 Unocal release was the creation of a continuous, real-time fenceline monitoring system that was installed at Unocal’s borders with Rodeo (the South Fenceline) and Crockett (the North Fenceline) in 1996. Each Fenceline system monitors some two dozen chemicals using several different monitoring devices, including “open-path” monitors that send a laser beam down the length of the refinery’s fenceline to detect any chemicals that may be crossing into the community. Crockett and Rodeo residents, working with engineers from CBE, were instrumental in selecting and testing the monitoring equipment, and have remained involved in overseeing the system ever since [8]. For other air monitoring activists, the Crockett-Rodeo Fenceline has since served as an exemplar of the kind of monitoring that could and, they argue, should be happening at all refineries. Although bucket monitoring is frequently accompanied by calls for continuous, real-time monitoring, very few communities have been successful in persuading refineries or regulatory agencies to install similar systems. Fenceline air monitoring remains concentrated in the Bay area, where the Chevron refinery in Richmond, California, has a system similar the Crockett-Rodeo Fenceline, installed in 2013. In addition, the regional air regulator, the Bay Area Air Quality Management District (BAAQMD), in 2016 passed a rule requiring similar systems to be installed at all 5 of the area’s refineries. The existing systems are required by the terms of their local land use permits, paid for by the refineries, and operated by an independent contracting company. Data are made publicly available in real time through a website (www. fenceline.org).

narrating their citizen science data, particularly in practices of storytelling in EJ campaigns more generally, have received little explicit attention. Yet in policy settings not unlike those addressed by EJ activists, storytelling has been shown to exist in complex relation with science. Scientific research cannot, in itself, determine what policy actions ought to be taken; further, on the most pressing policy issues, science is usually incomplete, uncertain, an/or contested [47]. Stories can help to envision policy directions in the absence of conclusive information or to assemble existing data into a case for particular policies [9,48–50]. Stories are powerful, in that similar data can be used to make conflicting cases, depending on, for example, where the story starts and ends [9]. They can also open up the policy-making process, by engaging diverse participants in playful exercises that tell stories about energy transformations [51]. At the same time, stories are constraining: they are most compelling when they do not depart too greatly from existing narratives [43], and policy realms may be defined by a small handful of stories into which data must be fit in order to be recognized [9]; see also [14]. As I will show below, similar patterns are evident in frontline communities’ practices of citizen science. Like policy experts, activists narrate their data and, at the same time, are constrained in the stories they can tell by pre-existing narratives. Examining the intertwining of data and stories in the EJ context helps us to better understand both the function that stories serve as a resource for meaning-making, as well as the limitations of storytelling in addressing the hermeneutic injustices suffered by frontline communities. 4. Community-driven air monitoring in frontline communities One arena in which to examine the relationship between data and story-telling in frontline communities—and, specifically, to understand whether and how stories can become a hermeneutic resource for communities trying to make sense of data—is community-based activism around ambient air toxics monitoring. Communities in close proximity to fossil fuel facilities, especially oil refineries and natural gas fracking sites, are routinely exposed to toxic air pollutants with known health effects, including carcinogens benzene, toluene, ethylbenzene, and xylene (BTEX), respiratory irritants like sulfur dioxide, and neurotoxins like hydrogen sulfide and carbon disulfide. Historically, information about the ambient levels of these and other toxins in residential communities has been scarce. Regulatory monitoring in the United States focuses on just six “criteria pollutants,” of which only sulfur dioxide is released in significant quantity by petrochemical plants, and monitoring stations are set up in areas considered representative of the airshed, not heavily industrialized areas where emissions are high. Further, standard procedures for monitoring air toxins involve sampling over a 24-h period which, residents of frontline communities argue, is too long to capture the peak periods of pollution that particularly concern them [12]. Beginning in the mid-1990s, a network of community activists and allied non-profit organizations with a special interest in air monitoring emerged. The network’s emergence was catalyzed by a 16-day release of a neurotoxin called catacarb from the Unocal oil refinery in Rodeo, California in 1994, which sickened residents in the downwind town of Crockett. Community outrage at the unreported (by Unocal) and undetected (by regulators) release resulted in two innovations in monitoring for toxic gases. The first was the development of the “bucket” air sampler, an inexpensive, community-friendly “grab sampling” device, with which community members collect air samples in non-reactive plastic bags and send them to a laboratory for analysis. Buckets enabled community groups to generate data about ambient air levels of toxic chemicals—without waiting for regulators to arrive with their monitoring equipment. First used in Crockett, Rodeo, and communities near the four other San Francisco Bay area refineries, buckets have since been adopted by communities near petrochemical facilities across the United States and

5. An ethnographic approach to data and storytelling In the past 20 years, then, the air quality data available to communities in close proximity to petrochemical facilities has grown dramatically, and is only continuing to grow. But how do they make sense of the data, and what role does storytelling play? The answers I offer in this paper derive from ethnographic engagement with air monitoring activists since 2001. My research has included an extended period of ethnographic fieldwork in New Sarpy and Norco, Louisiana, where buckets were being used as part of community campaigns (2002–2003); and two rounds of semi-structured interviews with individuals involved in community-based air monitoring efforts, one focused on the use of real-time technology in Chalmette, Louisiana (2007) and the other on the history of the Crockett-Rodeo Fenceline and the development of the bucket (2014–2016). I have also done archival research on the origins of buckets and the Fenceline, and I have observed a variety of events significant for air monitoring advocates, including the CommunityBased Science for Action Conference in New Orleans in November 2014, cosponsored by GCM and the Louisiana Bucket Brigade; GCM and US EPA Region 9’s “Workshop to Strengthen Community-Based Air Monitoring Protocols” in July 2015; and several BAAQMD public hearings on proposed refinery air monitoring rules in 2014 and 2015. In addition to these traditional research activities, I have actively participated in non-profit groups’ efforts to give meaning to air monitoring results. As a “science intern” at the Northern California EJ non4

Energy Research & Social Science xxx (xxxx) xxx–xxx

G. Ottinger

that the villain is not personified: the threat is, respectively, the pollution from the plant, or the accident-prone plant itself. The other two kinds of stories, stories of dissembling and stories of disrespect, feature human antagonists. In both kinds of story, the antagonists may be industry representatives, including the engineers who run the plants and the spokespeople and managers who attempt to defend it against criticism and manufacture community support; they may be the environmental regulators ostensibly charged with ensuring that the plants are not operating to the detriment of their neighbors; or community stories may include antagonists from both industry personnel and regulatory agencies. In stories of dissembling, residents who wish to know how they are being affected by petrochemical pollution are thwarted by plant personnel and/or regulators, who are portrayed as hiding or withholding relevant information. In one common version of this story, community members call the regulatory agency when they notice something awry at the neighboring facility—a particularly bad smell, for example—but the regulatory agency waits to investigate until the situation has passed. In another, a facility subject to surveillance (e.g. fenceline monitoring) takes steps to make sure its emissions evade detection. These stories tell of facilities flaring at night or on weekends, when inspectors aren’t looking, or of plant personnel disabling monitors when they know they’re going to have a release. In stories of disrespect, a concerned community member confronts an industry or environmental agency official about the impact of pollution on his community. The official dismisses his concern, insists that his observations are somehow wrong, questions his motives, or otherwise refuses to acknowledge the legitimacy of the community members’ intervention. Many stories of disrespect go on to describe some additional evidence that the official’s position could not possibly be true; here they overlap with stories of dissembling. None of these kinds of stories are dependent on quantitative data. They derive from community members’ observations and experiences. At the same time, however, they can be bolstered by data. One 2001 Louisiana Bucket Brigade publication tells the story of systemic danger with a table showing the number of pounds of sulfur dioxide released each month in accidents at the Orion refinery, showing that the facility averaged more than two accidents and 32,000 pounds of SO2 per week. Conversely, these stories can provide a context for making sense of quantitative data that is otherwise difficult to interpret, as the next section will show.

profit Communities for a Better Environment in 2001 and 2002, I compiled health-based standards and screening levels to which bucket results could be compared. As a volunteer at the Louisiana Bucket Brigade in 2002 and 2003, I helped develop protocols for taking wipe samples and tried to assemble comparable “levels of concern” for wipe sampling results. As a member of GCM’s advisory board from 2014 to 2016, I participated in the organization’s process of developing a new quality assurance plan. Finally, I am the Principal Investigator on a National Science Foundation grant to work with Bay area communities to design a web-based interface that renders data from fenceline monitors more accessible, interactive, and usable. Through these various activities, I have observed instances of community members and EJ activists telling retrospective stories that give meaning to air monitoring data, watched as they made sense of new data in the context of on-going campaigns, and struggled with them to interpret data in ways that are both sensible and significant. My cases have been based primarily in the San Francisco Bay area, California, and the New Orleans, Louisiana, area; however, the stories I have heard from other areas, combined with the Bay area’s status as a hub of activity on air monitoring issues, suggest that the practices I have observed are reasonably representative of practices in other U.S. communities. Because of the participatory, ethnographic nature of my research, incidents reported here are reconstructed from my field notes. Speech is paraphrased except where otherwise noted. 6. Air quality stories This paper asks to what extent the stories of frontline communities can be a resource for interpreting air quality data—which in turn requires asking what stories are being told by community members about the experience of living on the borders of an oil refinery or petrochemical plant. Across the communities with which I have interacted, several kinds of stories are told and retold: stories of harms to health, stories of systemic danger, stories of dissembling, and stories of disrespect. In stories of harms to health, community members tell how they have been sickened by exposures to pollution. Usually the protagonist—possibly the narrator, but just as likely a family member—starts out in relatively good health. But after breathing refinery emissions, they begin to experience health effects, or are diagnosed with cancer or some other disease. Sometimes these stories include a major chemical release from the facility that results in a severe worsening of symptoms. In another kind of harms-to-health story, a protagonist who suffers from chronic ill health—usually respiratory symptoms—goes away from the community for a period and finds that her symptoms vanish, only to come back when she returns home. Stories of systemic danger dovetail with stories of harms to health, but they focus on the disruption caused by releases and accidents. In these stories, an otherwise unremarkable moment is shattered by the sound of an explosion, an overpowering smell, or the ground shaking. The event plunges the protagonist into a state of panic and confusion—what’s going on? Should they shelter in place? Should they evacuate?—that may also include acute physical symptoms like burning eyes or shortness of breath. These stories continue past the moment of crisis, to situate the accident as one of many, and/or to describe how the fear it produced has changed the protagonist’s life. One example of a story of systemic danger comes from Steve Lerner’s [52] book, Diamond, which devotes an entire chapter (“Dangerous Neighbor”) to a chronology of the many accidents at a Shell chemical plant and refinery that shaped life in Norco, Louisiana’s, African-American community. Of a 1988 explosion that killed 7 workers, Lerner writes, “Many residents have not slept well since the explosion. 'We slept in our clothes for a long time,' said Jenny Taylor. ‘Tell you the truth, I still have my clothes laid out on a chair so I can jump into them….We don't know what they [the plants] are going to do, so you can't ever really be relaxed.'" (34) Harms-to-health stories and systemic-danger stories are similar in

7. The sense-making power of stories Even the most extensive air monitoring undertaken, or advocated for, by the frontline communities described here does not lend itself to ‘proving’ in any scientifically defensible way that residents’ health is being harmed by facility emissions. Environmental and health agencies (both in the U.S. and internationally) have published reference concentrations and screening levels for the chemicals that are being measured. However, comparing those to measured values is less than straightforward. The process of setting those levels involves a good deal of uncertainty [53]. As a result, the levels set by different agencies can differ dramatically, leaving open the question of which value monitoring results should be compared to. In addition, screening levels, reference concentrations, and regulatory standards (where they exist at all) are largely incommensurate with the data collected by or available to communities. Where the former speak to averages over periods ranging from an hour to a year, results from bucket samples represent averages over only a few minutes, and measurements from continuous monitors like the Crockett-Rodeo Fenceline are reported as instantaneous measurements, without relevant averages being calculated (or data being made available for download to enable community groups to calculate them). Viewed as a potential resource for making sense of air quality data collected by fenceline communities, then, scientific and administrative 5

Energy Research & Social Science xxx (xxxx) xxx–xxx

G. Ottinger

(LABB), an EJ non-profit supporting New Sarpy residents in their campaign, I helped Rushing take samples of the deposits on the truck and vegetation and send them to a laboratory for analysis. The results showed measurable levels of a number of heavy metals that are emitted by refineries. But in looking for scientific frameworks that could help us interpret the results, I quickly learned that, without much more sophisticated analysis, there was no way to know whether the chemicals we measured posed a threat to residents’ health. Although we had measurements, we couldn’t say what they meant. Or so I thought. For Rushing the meaning of the results was immediately clear: because it contained heavy metals that the refinery was releasing, that substance was clearly no ordinary dirt. It had come from the refinery, just as he’d been saying, and the refinery inspector had been wrong, not to mention condescending, in trying to convince him otherwise. When Rushing subsequently told his story—which he did with even greater vehemence—he included the wipe sampling results, enhancing his portrayal of the inspector as disrespectful, and possibly even deceitful. Again, this portrayal depended only on the fact that heavy metals, linked to refinery operations, had been measured in the sample. Whether the levels measured exceeded thresholds for health effects was not important to the story of disrespect. In making meaning of their data, community groups did not always have to find alternatives to harms-to-health stories. In at least one case, stories of harms to health did become a resource for reinterpreting air monitoring data—and shifting experts’ frameworks in the process. In June 2012, the former Unocal refinery, now owned by Phillips 66, in Rodeo, California, had an accidental release from a sour water tank. A foul smell pervaded the area, and the Crockett-Rodeo fenceline monitoring system measured high levels of hydrogen sulfide (H2S), around 12 parts per million (ppm). However, the release did not trigger any sirens or a “shelter in place” advisory from the refinery or the county government, because measured levels of H2S remained below the trigger point for the Community Warning System (CWS), which was set at 15 ppm. According to Crockett resident and Crockett-Rodeo Fenceline oversight committee member Frank Brosnan, stories of harms to health were prominent at a subsequent community meeting about the release. Residents described being sickened by exposure to the high levels of H2S, and many even went to the emergency room. In keeping with stories of systemic danger, they also questioned the adequacy of the CWS, which had not warned them of the hazard. Brosnan and others began asking how the CWS level had been set, concluded that the bases for it were not especially robust (see [53]), and successfully argued for it to be changed from 15 ppm to 10 ppm, given the clear evidence of health impacts at 12 ppm. In this case, then, harms-to-health stories gave air monitoring data an alternate meaning. The official framework for interpreting the data said, in effect, that 12 ppm was nothing to worry about. But residents’ stories of how their health had been harmed as a result of the release cast the 12 ppm in a new light: it became a worrisome level, and one on which action needed to be taken. Importantly, residents’ ability to change official interpretive frameworks by mobilizing monitoring data—something that activists have not been able to do with bucket sampling—was enabled in no small part by a basic commensurability: stories of acute health effects were combined with short-term data to influence short-term or acute levels of concern. Mobilizing fenceline data to make arguments about longer-term or systemic issues has proven more difficult, as the next section suggests, highlighting the limitations of story-telling as a strategy for meaning-making.

frameworks are quite limited. In particular, they do not easily facilitate the incorporation of data into the harms-to-health story, where data would seem most relevant. Both other stories, especially stories of dissembling and stories of disrespect, can give data meaning independent of its ability to prove that community health is being harmed by pollution. For example, in June 2001, a lightning strike set fire to a gasoline storage tank at the Valero refinery in New Sarpy, Louisiana. Although photos of the event show a large plume of black smoke emanating from the tank, refinery officials attempted to reassure the neighboring community that, in the 13 h that the fire burned, they were not exposed to any dangerous chemicals because the wind blew emissions from the fire away from the community. This claim was met with incredulity by residents who had seen and smelled the smoke, and in the days following the fire, refinery officials’ statement was offered as one more episode in an on-going story of dissembling, just as the fire itself was one more episode in residents’ story of systemic danger. At the time of the fire, however, a bucket air sample had been taken and sent away for laboratory analysis. When the sample results arrived (a process that at the time took a couple of weeks), the story of dissembling gave them immediate significance. Sample analysis showed several toxic chemicals, including carbon disulfide and carbonyl sulfide, present in the air at high levels. Refinery officials’ statements that the fire had no impact on the community were already being painted as absurd. These data—showing that refinery-related chemicals had been in the air at high levels at the time of the fire—bolstered the story that refinery officials had been making a clumsy attempt to hide the true impact of their accident. The high chemical levels also became part of the story of systemic danger, suggesting what residents had to fear from the many accidents at the refinery. Importantly, the story of systemic danger and, especially, the story of dissembling gave residents’ data meaning independent of scientific frameworks for interpreting air monitoring data. The simple presence of measurable levels of carbon disulfide and carbonyl sulfide showed that refinery officials had been less than truthful when they claimed their massive fire had no impact on the community. The demonstrated impacts, in turn, added gravity to the story that accidents at the refinery put residents in danger. Although the levels of chemicals measured during the fire were compared to health-based screening levels, attacks on their commensurability with the screening levels or questions about significance for health in the longer term could not detract from their meaning in the context of stories of systemic danger and dissembling. The inadequacies of scientific frameworks for interpreting data in ways that represent residents’ experience—the hermeneutic injustice described above—were thus countered by using stories to make sense of air sampling data. Another instance of data taking on meaning in the context of community stories occurred in New Sarpy, Louisiana, in the summer of 2002. Throughout the summer, at events associated with Concerned Citizens of New Sarpy’s campaign for relocation, resident Harlon Rushing was telling his own story of disrespect. He had noticed black soot on his truck and the bushes around his home, three streets over from the Orion refinery. He washed the truck and put up an open-sided tent as a shelter for it, but within days soot once again coated his truck and his new canvas carport. Convinced that the soot had been emitted from the refinery, he called Orion to complain. The company sent an inspector out, but the inspector informed Rushing that the black, slightly sticky substance dotting his property was just dirt. Rushing found the inspector’s response insulting, because—in his telling—it expected him to buy into an explanation that defied common sense: the substance was on the tops of the leaves, not the undersides, he pointed out, and “dirt don’t just jump up on stuff.” Rushing’s irritation with the soot provided an occasion to try out “wipe sampling” for deposited particulates. As an intern at Communities for a Better Environment (CBE) in California, I had researched the procedure; as a volunteer for the Louisiana Bucket Brigade

8. Narrative mismatches In keeping with research showing storytelling to be a vehicle for meaning making by EJ communities (e.g. [41]), the examples above show how the stories of communities on the front lines of energy and petrochemical facilities serve as powerful resources for making sense of 6

Energy Research & Social Science xxx (xxxx) xxx–xxx

G. Ottinger

professionals leading the study design understood that, were they to ask about cancer deaths, they would have to distinguish brain cancer from lung cancer from lymphoma, and that the survey would have to find a large number of deaths in any one of those categories to constitute a statistically significant finding for the fewer than 100 households surveyed. These methodological constraints might be seen as a kind of hermeneutic injustice: the available resources—epidemiological approaches to understanding environmental health—make it difficult to understand deaths in a small community as potentially connected and to relate them to environmental conditions and current illnesses among living community members (c.f. [54]). The case of the New Sarpy health study could also be seen as a problem stemming from the constraints imposed by existing narratives. The story of the study results was not a difficult one to tell: refinery releases were noticeable; samples had been taken to show that people were being exposed to the chemicals released; the survey showed that they were sick with respiratory ailments caused by those same chemicals. Yet incorporating it into residents’ campaign would have required New Sarpy residents to shift the focus of their harms-to-health story from cancer deaths to respiratory illness. The example thus points to another kind of hermeneutic invention that may be required to make sense of citizen science: the creation of stories that can fit new and old information together. To be fair, had the community not settled with Orion so soon after the study results were released, it seems possible that residents may have slowly updated their stories to take account of the new results.1 The fenceline air monitoring systems in Crockett-Rodeo and Richmond, California, highlight another form of narrative mismatch. The monitoring systems and the online interfaces that display monitoring data to the public are oriented to accidental releases of high levels of chemicals and the acute health harms that they might cause: they report current measured levels of chemicals and allow users to view the past 24 h of measurements. Community members experiencing acute harms to health can thus make connections to potential air quality issues in real time, as did Crockett and Rodeo residents during the H2S release described above. The data generated by the monitoring systems could also potentially speak to long-term health and environmental issues, but the web interfaces do not support exploration, queries, or calculations that would help residents build a picture of air quality over months or years. Hypothesizing that a different sort of interface would enable community members to make more meaning of the data, especially in the context of concerns about long-term and systemic issues, in 2016 I initiated a participatory design process—in collaboration with engineers from the CREATE Lab at Carnegie Mellon University and community activists from Crockett, Rodeo, Richmond, and Benicia, California—to create a new website to make data more accessible and interpretable to residents of Bay area refinery communities. Through that process, it quickly became evident that the inaccessibility of data online was a relatively superficial obstacle to sense-making. Once it was possible to plot and explore months’ worth of data, we ran into the same interpretive problems that other community-based monitoring efforts have. Health-based screening levels that distinguish safe from unsafe levels are uncertain and highly variable. Regulatory limits, where they exist at all, share with screening levels the problem of being expressed in terms of time frames that do not map well onto community members’ everyday experience of living with petrochemical pollution. And detection limits for the monitors themselves were in many cases too high to

citizen science data that would be difficult to interpret—or likely to be misinterpreted—using scientific frameworks alone. They arguably help overcome epistemic injustices by addressing the tensions between dominant hermeneutic resources—namely concepts and frameworks used by scientists to make sense of chemical exposures—and the lived experiences of frontline communities. Yet stories also have limitations as resources for making sense of complex data. As the examples below demonstrate, data do not always fit easily with stories of harms to health, systemic danger, dissembling, and disrespect. These situations of narrative mismatch require further invention of hermeneutic resources, which frontline communities may or may not have the capacity to do. They thus highlight the continuing problem of epistemic injustice and the need for expert help in inventing modes of making data meaningful that are faithful to residents’ experience. In New Sarpy, Louisiana, in the summer of 2002, residents and their allies were trying to put data behind their story that emissions from the adjacent Orion refinery harmed their health. An occupational medicine specialist (who had previous experience working with frontline communities) and a Master’s of Public Health student (who did not), led New Sarpy residents in designing and conducting a community health survey that was both scientifically robust and highly attuned to local conditions. The survey asked members of randomly selected households whether they experienced a list of symptoms (e.g. coughing, wheezing) or had been diagnosed with illnesses (e.g. asthma, cancer) that were known to be associated with exposures to chemicals that had been detected in the community and that were also queried in the U.S. National Health and Nutrition Examination Survey (NHANES). By using NHANES questions, the study design team ensured that data about the rates of symptoms in New Sarpy could be compared to a nation-wide benchmark—solving, survey designers imagined, the interpretive problem in advance. To avoid making the study methodology overly complex, survey respondents were asked only about illnesses (morbidity) and not deaths (mortality) of family members or neighbors. Data from the survey showed to a high degree of statistical significance that New Sarpy residents experienced elevated rates of asthma, chronic bronchitis, and number of respiratory symptoms. There was additional evidence of elevated cancer rates, but these findings were not statistically significant. Despite the clear evidence provided by the survey that residents of the refinery-adjacent community suffered more illness than other Americans, the study results were not incorporated into community members’ stories or used in their campaign. In fact, a number of residents told me that the study “didn’t show anything.” To no small extent, residents’ dismissal of what was arguably a powerful study stemmed from a narrative mismatch. The harms-tohealth story that New Sarpy residents had been telling had a specific flavor: it was almost exclusively about the many New Sarpy residents who had died of cancer, including a number of individuals whose properties had been adjacent to one another. Their narrative focus on cancer took up a larger story about the petrochemical industry in the region, which environmental activists have dubbed “Cancer Alley.” Indeed, one man whose wife was among those who had died of cancer regularly concluded his harms-to-health story by saying that they called the area “Cancer Alley” for a reason. In the context of New Sarpy residents’ particular story, study results showing elevated rates of respiratory problems but not of cancer incidence or cancer death had little meaning. The narrative mismatch between the health study’s story of a community gasping for breath and residents’ story of neighbors and loved ones dropping dead of cancer could be seen as a disconnect between the hermeneutic resources offered by science and the lived experience of the community. The decision not to pursue questions about cancer deaths was made on the basis of scientific models for understanding morbidity and mortality—models which look at the two separately, divide illnesses and death into narrow disease categories, and demand high levels of statistical significance. The public health

1 Houston’s [40] (1) account of storytelling as a vehicle for accommodating and weaving together multiple perspectives points to this as a strong possibility; on the other hand, Polletta et al. [43] (1) suggest the difficulty of advancing new stories that depart from existing narratives, although in this case it is not the narrative arc (refineries emit chemicals that make people sick) but the details of its substance that would have changed. Studies of both storytelling and EJ-oriented citizen science would benefit from research on how and whether new data can shift existing narratives.

7

Energy Research & Social Science xxx (xxxx) xxx–xxx

G. Ottinger

“sexual harassment” related by Fricker [13], have yet to be well documented. Moreover, the era of big data seems still to call for additional experimentation in methods for making new, socially situated meaning of complex data. What seems clear is that these processes will need to be broadly collaborative. The situated knowledge and experience of ordinary citizens, especially residents of frontline communities, is clearly necessary to the process of making new meaning, but so too is the expertise of individuals with deep knowledge of relevant scientific disciplines (e.g. chemistry, toxicology, meteorology) and individuals accustomed to working with data. Social scientists, community organizers, and other “middle actors” [15] (see also [55]) who can speak across the various levels and communities involved in the making of policy, also have a role to play (c.f. [51]). Key to the success of these collaborations, however, will be the willingness of all participants to look beyond existing scientific frameworks and stories to create alternative ways of making sense that are grounded in multiple, diverse ways of knowing. This vision of hermeneutic invention underscores the importance of cultivating curious, humble, open-minded experts [56] willing to engage in collaborative practices of meaning-making (see also [16]), and suggests the need for policies aimed at furthering environmental and epistemic justice to go beyond promoting access to data and attend to fostering processes of meaning-making as well by, for example, providing resources for community groups to work with NGOs, regulators, and experts from a variety of disciplinary backgrounds on analyzing public data and, hopefully, developing modes of interpretation adequate to the experiences of frontline communities.

register ambient air levels of chemicals until they were far above standards and screening levels. Looking to community stories as resources for sense-making only led us to more mismatches. Stories of dissembling were especially prominent. Community members involved in our participatory design process often claim, for example, that air quality is worse at night and on the weekends (when no one is watching) and that the monitors are taken down when there are releases. Yet with readings of two dozen chemicals and meteorological data every five minutes at each of several monitoring stations, it has proven difficult to create algorithms that could show these kinds of trends—if indeed they do exist. Stories of systemic danger led us to look in the data for “incidents,” or moments of especially high chemical concentrations, but we struggled to define what an incident looked like. How high? For how long? For which chemical or chemical(s)? Having had limited success with these approaches, we have begun to experiment with the collection of other kinds of data to give meaning to measurements of chemical concentrations, hoping that real-time health data from commercial fitness trackers and residents' reports of odors or symptoms can help bridge the gap between extant stories and available data. The struggle to find broader meaning in real-time air quality data underscores the limitations of stories as a hermeneutic resource for frontline communities. The data can be quite helpful when they line up with the stories being told, as when the immediacy of fenceline monitoring data match the immediacy of the kind of harms-to-health stories that arise in the context of an accident, or when data effectively catches refinery officials in a lie. But stories of long-term or systemic harm are not easily told with instantaneous data, and instantaneous data are hard to make sense of when the stories emphasize systemic harm.

References [1] Kaile Zhou, Shanlin Yang, Understanding household energy consumption behavior: the contribution of energy big data analytics, Renew. Sustain. Energy Rev. 56 (2016) 810–819. [2] C. Duncan McKinley, Abraham Miller-Rushing, L. Heidi Ballard, Rick Bonney, Hutch Brown, C. Susan Cook-Patton, M. Daniel Evans, A. Rebecca French, K. Julia Parrish, B. Tina Phillips, F. Sean Ryan, A. Lea Shanley, Jennifer Shirk, F. Kristine Stepenuck, F. Jake Weltzin, Andrea Wiggins, D. Owen Boyle, D. Russell Briggs, F. Stuart Chapin III, A. David Hewitt, W. Peter Preuss, A. Michael Soukup, Citizen science can improve conservation science, natural resource management, and environmental protection, Biol. Conserv. 208 (2017) 15–28. [3] Scott Frickel, Sahra Gibbon, Jeff Howard, Joanna Kempner, Gwen Ottinger, David J. Hess, Undone science: charting social movement and civil society challenges to research agenda setting, Sci. Technol. Hum. Values 35 (4) (2010) 444–473. [4] Scott Frickel, Michelle Edwards, Untangling ignorance in environmental risk assessment, in: Soraya Boudia, Nathalie Jas (Eds.), Powerless Science?: Science and Politics in a Toxic World, Berghahn Press, New York, 2014, pp. 215–233. [5] Abby Kinchy, Kirk Jalbert, Jessica Lyons, What is volunteer water monitoring good for?: Fracking and the plural logics of participatory science, Pol. Power Soc. Theory 27 (2014) 259–289. [6] Gregg P. Macey, Ruth Breech, Mark Chernaik, Carolyn Cox, Denny Larson, Deb Thomas, David O. Carpenter, Air concentrations of volatile compounds near oil and gas production: a community-based exploratory study, Environ. Health 13 (1) (2014) 82. [7] Christine Overdevest, Brian Mayer, Harnessing the power of information through community monitoring: insights from social science, Texas Law Rev. 86 (7) (2008) 1493–1526. [8] Gwen Ottinger, Citizen engineers at the fenceline, Issues Sci. Technol. 32 (2) (2016) 72–78. [9] Kathryn B Janda, Marina Topouzi, Telling tales: using stories to remake energy policy, Build. Res. Inf. 43 (4) (2015) 516–533. [10] Phil Brown, Popular epidemiology and toxic waste contamination: lay and professional ways of knowing, J. Health Soc. Behav. 33 (3) (1992) 267–281. [11] Jason Corburn, Environmental justice, local knowledge, and risk: the discourse of a community-based cumulative exposure assessment, Environ. Manage. 29 (4) (2002) 451–466. [12] Gwen Ottinger, Buckets of resistance: standards and the effectiveness of citizen science, Sci. Technol. Hum. Values 35 (2) (2010) 244–270. [13] Miranda Fricker, Epistemic Injustice: Power and the Ethics of Knowing, Oxford University Press, Oxford, 2007. [14] Rosemary Randall, Loss and climate change: the cost of parallel narratives, Ecopsychology 1 (3) (2009) 118–129. [15] Mithra Moezzi, Kathryn B. Janda, From if only to social potential in schemes to reduce building energy use, Energy Res. Soc. Sci. 1 (2014) 30–40. [16] Jason Corburn, Street science: community knowledge and environmental health justice, in: Robert Gottlieb (Ed.), Urban and Industrial Environments, The MIT Press, Cambridge, MA, 2005. [17] Rebecca Mason, Two kinds of unknowing, Hypatia 26 (2) (2011) 294–307. [18] Gaille Pohlhaus, Relational knowing and epistemic injustice: toward a theory of

9. Conclusion: building hermeneutic resources Communities on the front lines of oil refineries and petrochemical plants in the United States regularly experience epistemic justices: their testimony is not given due credibility; resources for shared-sense making favor dominant, especially experts’, ways of looking at the world. Collecting and engaging with ambient air monitoring data has been one way that communities adjacent to refineries have attempted to address testimonial injustices, by giving their testimony quantitative backing, but their efforts have made hermeneutic injustices all the more visible. Scientific frameworks for making sense of the data are not adequate to the communities’ experiences. Building on research that shows storytelling to be a key way of making meaning in environmental justice and environmental policy settings, I have argued that frontline communities’ stories of harms to health, systemic danger, dissembling, and disrespect can often serve as resources for making (alternative) sense of air monitoring data, but that their effectiveness is limited in situations where the stories are not well matched to the available data. For citizen scientists in frontline communities—certainly in the U.S. but possibly also in other countries were fossil fuels are extracted, transported, and processed—stories are important but not adequate means of addressing hermeneutic injustices. The problem of narrative mismatch is not unique to situations of environmental or epistemic injustice [9], and further research to understand patterns in where narrative mismatch does and does not occur would be valuable. But noting the occurrence and consequences of mismatch points to the continuing need, across spheres of energy and environmental policy, for hermeneutic invention, or the creation of new concepts and frameworks that help to make visible understandings and experiences of marginalized groups. Especially as the data sets available to policy makers and frontline communities grow, those pushing for greater environmental justice will need to find ways to go beyond both current scientific frameworks and existing stories in order to enable communities to use data to better communicate their experience and push for meaningful change. The processes of hermeneutic invention, such as the case of naming 8

Energy Research & Social Science xxx (xxxx) xxx–xxx

G. Ottinger

[41] Phaedra C. Pezzullo, Performing critical interruptions: stories, rhetorical intervention, and the environmental justice movement, West. J. Commun. 65 (1) (2001) 1–25. [42] James A. Throgmorton, Planning as persuasive storytelling and global-scale web of relationships, Plann. Theory 2 (2) (2003) 125–151. [43] Francesca Polletta, Pang Ching Bobby Chen, Beth Gharrity Gardner, Alice Motes, The sociology of storytelling, Ann. Rev. Soc. 37 (2011) 109–130. [44] Juliane Reinecke, Shaz Ansari, Taming wicked problems: the role of framing in the construction of corporate social responsibility, J. Manage. Stud. 53 (3) (2016) 299–329. [45] David Schlosberg, Environmental Justice and the New Pluralism, Oxford University Press, Oxford, 1999. [46] Raoul S. Liévanos, Jonathan K. London, Julie Sze, Uneven transformations and environmental justice: regulatory science, street science, and pesticide regulation in California, in: Gwen Ottinger, Benjamin R. Cohen (Eds.), Technoscience and Environmental Justice: Expert Cultures in a Grassroots Movement, MIT Press, Cambridge, MA, 2011, pp. 201–228. [47] Daniel Sarewitz, How science makes environmental controversies worse, Environ. Sci. Policy 7 (2004) 385–403. [48] Raul Lejano, Mrill Ingram, Helen Ingram, The Power of Narrative in Environmental Networks, MIT Press, Cambridge, MA, 2013. [49] Noam Bergman, Stories of the future: personal mobility innovation in the UK, Energy Res. Social Sci. 31 (2017), http://dx.doi.org/10.1016/j.erss.2017.06.028 (forthcoming). [50] Sachiko Muto, From Laissez-Faire to intervention: a narrative analysis of policy discourse on standardization for smart grid in the United States, Energy Res. Soc. Sci. (2017), http://dx.doi.org/10.1016/j.erss.2017.05.036 (forthcoming). [51] Joe Smith, Robert Butler, Rosie J. Day, Axel H. Goodbody, David H. Llewellyn, Melanie Rohse, Bradon T. Smith, Renata A. Tyszczuk, Julia Udall, Nicola M. White, Gathering around stories: interdisciplinary experiments in support of energy system transitions, Energy Res. Soc. Sci. (2017), http://dx.doi.org/10.1016/j.erss.2017.06. 026 (forthcoming). [52] Steve Lerner, Diamond: a struggle for justice in louisiana's chemical corridor, in: Robert Gottlieb (Ed.), Urban and Industrial Environments, The MIT Press, Cambridge, MA, 2005. [53] Sylvia Noble Tesh, Uncertain Hazards: Environmental Activists and Scientific Proof, Cornell University Press, Ithaca, 2000. [54] Barbara L. Allen, The popular geography of illness in the industrial corridor, in: E. Craig Colten (Ed.), Transforming New Orleans and Its Environs: Centuries of Change, University of Pittsburgh Press, Pittsburgh, 2000, pp. 178–201. [55] Dean Nieusma, Middle-out social change: expert-led development interventions in Sri Lanka’s energy sector, in: Gwen Ottinger, Benjamin R. Cohen (Eds.), Technoscience and Environmental Justice: Expert Cultures in a Grassroots Movement, MIT Press, Cambridge, MA, 2011, pp. 119–145. [56] José Medina, The Epistemology of Resistance: Gender and Racial Oppression, Epistemic Injustice, and Resistant Imaginations, Oxford University Press, Oxford, 2012. [57] Sara Ann Wylie, Kirk Jalbert, Shannon Dosemagen, Matt Ratto, Institutions for Civic Technoscience: how critical making is transforming environmental research, Inf. Soc. 30 (2) (2014) 116–126. [58] Ottinger Gwen, Epistemic fencelines: air monitoring instruments and expert-resident boundaries, Spontaneous Gener. 3 (1) (2009) 55–67.

willful hermeneutic ignorance, Hypatia 27 (4) (2012) 715–735. [19] Kristie Dotson, Tracking epistemic violence, tracking practices of silencing, Hypatia 26 (2) (2011) 236–257. [20] H.M. Collins, Robert Evans, The third wave of science studies: studies of expertise and experience, Soc. Stud. Sci. 32 (2) (2002) 235–296. [21] Sheila Jasanoff, Breaking the waves in science studies: comment on H.M. Collins and Robert Evans, the third wave of science studies, Soc. Stud. Sci. 33 (3) (2003) 389–400. [22] Brian Wynne, Seasick on the third wave? Subverting the hegemony of propositionalism: response to Collins & Evans (2002), Soc. Stud. Sci. 33 (3) (2003) 401–417. [23] Wendy Faulkner, Nuts and bolts and people: gender-troubled engineering identities, Soc. Stud. Sci. 37 (3) (2007) 331–356. [24] Mia Ong, Body projects of young women of color in physics: intersections of gender, race, and science, Soc. Probl. 52 (4) (2005) 593–617. [25] Steven Shapin, A Social History of Truth: Civility and Science in SeventeenthCentury England, University of Chicago Press, Chicago, 1994. [26] David Schlosberg, Defining Environmental Justice: Theories, Movements, and Nature, Oxford University Press, Oxford, 2007. [27] Luke W. Cole, Sheila R. Foster, From the Ground Up: Environmental Racism and the Rise of the Environmental Justice Movement, New York University Press, New York, 2001. [28] Ronnie Greene, From Homemaker to Hell-Raiser in Love Canal, Center for Public Integrity, 2013. [29] Mhairi Aitken, Wind Power planning controversies and the construction of ‘expert’ and ‘lay’ knowledges, Sci. Cult. 18 (1) (2009) 47–64. [30] Gwen Ottinger, Refining Expertise: How Responsible Engineers Subvert Environmental Justice Challenges, New York University Press, New York, 2013. [31] Robert R. Kuehn, The environmental justice implications of quantitative risk assessment, Univ. Ill. Law Rev. 38 (1996) 103–172. [32] Nancy Langston, Precaution and the history of endocrine disruptors, in: Soraya Boudia, Nathalie Jas (Eds.), Powerless Science?: Science and Politics in a Toxic World, Berghahn Press, New York, 2014, pp. 29–45. [33] Melissa Checker, ‘But I know it’s true’: environmental risk assessment, justice, and anthropology, Hum. Organ. 66 (2) (2007) 112–124. [34] Jaclyn R. Johnson, Darren J. Ranco, Risk assessment and native Americans at the cultural crossroads: making better science or redefining health? in: Gwen Ottinger, Benjamin R. Cohen (Eds.), Technoscience and Environmental Justice: Expert Cultures in a Grassroots Movement, MIT Press, Cambridge, MA, 2011, pp. 179–199. [35] Gwen Ottinger, Elisa Sarantschin, Exposing infrastructure: how activists and experts connect ambient air monitoring and environmental health, Environ. Soc. 3 (2) (2017) 155–165. [36] Dara O'Rourke, Gregg P. Macey, Community environmental policing: assessing new strategies of public participation in environmental regulation, J. Policy Anal. Manage. 22 (3) (2003) 383–414. [37] Merlijn van Hulst, Storytelling, a model of, and a model for, planning, Plann. Theory 11 (3) (2012) 299–318. [38] Gary Allen Fine, Barry O'Neill, Policy legends and folklists: traditional beliefs in the public sphere, J. Am. Folkl. 123 (488) (2010) 150–178. [39] Andrea Whittle, Frank Mueller, Bankers in the dock: moral storytelling in action, Hum. Relat. 65 (1) (2011) 111–139. [40] Donna Houston, Environmental justice storytelling: angels and isotopes at Yucca Mountain, Nevada, Antipode 45 (2) (2013) 417–435.

9