Introduction

Research projects that entail collaborations with lay people (citizen scientists) are common in the natural and biological sciences where citizen science has made an enormous and positive contribution to the scaling of data collection efforts [1]. This is especially true when it comes to studies focused on complex or abstract data whose documentation involves advanced reasoning or inferential skills that cannot be reliably outsourced to technology [2]. Abstract data arguably comprise the bulk of that seen in social science research, and yet the adoption of citizen science as a research strategy remains conspicuously absent here [3]. Within the social sciences, some types of linguistic data are spontaneously situational and therefore difficult to capture in systematic fashion (e.g., complaints, apologies, terms of address/reference, greetings, and so on). Collaborations of linguists and other citizen scientists in this domain could help capture greater diversity in factors including scale, location, network, social interaction, and sociocultural variables than is typically seen in the current literature. As stated by Dickinson et al. [4:291] in the context of ecological research, “[d]ispersed data collection and the ability to collect observations and connect with people, in places, and at scales that would otherwise not be possible render citizen science increasingly important.” This is particularly true given the widespread use of smart devices to facilitate standardized, large-scale data collection efforts by researchers in situ and in real time [3, 5]. The relative paucity of citizen social scientists (and associated loss of scientific knowledge) has been noted by several scholars, including linguists Rymes and Leone [6], who have championed for the creation of citizen sociolinguistics, specifically.

The definition of citizen sociolinguistics remains unsettled. While the inclusion of laypeople in research projects is one obvious aspect of such work, opinions differ as to the nature of their involvement. Svendsen [7] states this burgeoning field is defined by the engagement of lay people in the process of sociolinguistic research. Fitting with the approach we apply here, this can range from data documentation to more complex tasks associated with analysis and interpretation. Rymes and Leone [6], however, define citizen sociolinguistics using a more metalinguistic lens, focusing on the secondary analysis of language by everyday people as they reflect publicly (e.g., via social media) on language phenomena (e.g., use of accent/dialect, non-prescriptive grammar forms, monolingualism, etc.). In effect, the layperson becomes the citizen sociolinguist and analyst at the moment they make comments on everyday language uses. Rymes et al. [8] therefore seem to suggest that the citizen sociolinguist need not be included in the execution of research (or even be aware of it).

Despite these differences in perspective on the defining boundaries of practice, both Rymes’ et al. and Svendsen’s approaches entail some form of indirect or direct engagement with local citizens to understand otherwise inaccessible language data. Both would also appreciate that citizen sociolinguists may often be better situated than professional linguists to understand the relevance of language data heard in everyday environs. Moreover, both perspectives contend (albeit to variable degrees) that citizen sociolinguistics could have a democratizing effect on the acquisition of linguistic data by dismantling the artificial boundary between the researcher and researched, facilitating bidirectional dialogue between the two and changing the “traditional way we conceive science and who has the authority to do it” [7:3]. Certainly, the study of language has historical roots in community-based research [9–12], but citizen sociolinguistics can expand on this tradition in a distinct way. Engaging local field-site residents, community members, and students in language studies is common practice in the social sciences [10, 12–14] where people are selected to participate based on their willingness in combination with other characteristics or skills (e.g., being identified lay experts of the language form under investigation). In an era where travel, relocation, and online interaction are the norm [15–16], citizen sociolinguists may increasingly provide a means of critical research support by virtue of the fact they can capture data while navigating multiple daily social networks—networks in which interactions and the relationships that shape them are multimodal and complex [17–18], and for which sociolinguistic data may be more unconscious than that produced via focus groups, prompts, or surveys (what [3] termed hidden data). Citizen sociolinguistics therefore enriches the research goals by scaling-up the project, flattening power asymmetries, and capturing a more diverse array of speakers and speaking styles.

Here we report on the methodological outcomes of a novel citizen sociolinguistics project designed to leverage citizen social scientists to identify and systematically capture spontaneous, verbal interactions, a key form of linguistic data. We follow Svendsen [7] in our interpretation of citizen sociolinguistics in that we include lay people in the actual doing of research. Volunteers from the local community were trained to be citizen sociolinguists and were embedded in research teams, identifying and documenting a specific conversational interaction called “fat talk.” Fat talk consists of verbal exchanges wherein speakers disparage their own bodies; the disparagement often takes the form of a request for evaluation from fellow interlocutors [19]. For example, fat talk frequently includes self-deprecating or body-checking utterances like “I look so fat!” or “Does this make me look fat?” which typically evoke a scripted conversational response (such as “No, you don’t”) [19–21]. The response to fat talk seems to depend on the perceived status, identity, familiarity, and cultural backgrounds among interlocutors. Therefore, fat talk is sociolinguistically important precisely because it is rooted in interaction, making audible the commonly held, deep-seated concerns about physical appearance shaped by broader sociocultural norms [20–21].

Fat talk was first identified by Nichter and Vukovich [19] who investigated the pervasiveness of diet-focused culture in American adolescents (see also [20–21]). Self-deprecating or self-evaluating statements such as “My thighs are so fat!” or “Do these jeans make me look fat?!” were frequently expressed by the young people in Nichter’s focus groups [20]. Since then, expanded investigations of fat talk have confirmed it to be common among women of all ages [22–23], and indeed the focus of most fat talk studies has been on women (e.g., [24–25]). However, reports of fat talk among men has increased in recent years ([21, 26–27]), particularly in reference to body shame and weight stigma (e.g., [28–30]). While there is some variation in the data collection tools used in these studies, most utilize structured means of data collection that include self-report/recall exercises, lab-based experiments with confederates, discourse completion tasks and, as in the case of Becker et al [31], a free list response to a fat talk prompt. While these reveal important dimensions about fat talk cultural scripts, their structured nature masks whether these scripts are evident in fat talk when it arises organically and spontaneously in different social interactions.

While outwardly mundane on its surface, fat talk can have potentially important impacts on interlocutors beyond its interactional significance. At the group level, there may be some positive outcomes, such as greater group cohesion and an inclusive recognition of broader social norms [20]. However, outcomes appear largely negative at the individual level, and can increase negative health effects including body dissatisfaction and disordered eating [32–36]. Despite this clear connection to psychosocial health, fat talk has remained virtually undocumented in its naturally-occurring forms precisely because it is typically expressed in ways that are brief and spontaneous, being commonly woven into or in between other conversations (much like small talk) [37]. For this reason, most prior academic work on fat talk relies upon written prompts in which participants are asked to respond to artificial interactions or to listen to a fat talk statement and then complete a body dissatisfaction survey [25,38]. Fat talk was an ideal candidate for this nascent study in citizen sociolinguistics for multiple reasons, both in terms of advancing sociolinguistic understanding on this understudied topic and in terms of citizen sociolinguist engagement. These include its status as a verbal interaction shaped by sociocultural norms, its clear connection to psychosocial health and group cohesion, the brief nature of its expression being an accessible data target for new citizen science researchers, and the public’s general fascination with this form of speech (see [39–41]).

One consistent theme across researchers who look at citizen science with a critical eye is that it should do more to democratize the process, specifically that it should alter “the relation between the researcher and the researched” [7:139, 42] and increase the scope of practitioners who have the authority to “do science.” At a basic level, this requires that such projects go beyond seeing the citizen scientist as a “mere” data collection assistant [43–46] and instead create opportunities for meaningful inclusion and sustained dialogue between academics and lay folk. Thus, an additional challenge we undertook in this work was to ensure that citizen sociolinguists were integrated into the interpretation and analysis of sociolinguistic data. In this way, our project is designed to assess whether citizen sociolinguistics can simultaneously increase public knowledge while advancing scientific process and understanding in novel ways [6–7].

The primary aim of this project was to test the feasibility of citizen sociolinguistic methods for capturing large numbers of public fat talk utterances. This entailed assessing fat talk captured by citizen sociolinguistic research teams for its adherence to already understood forms of fat talk reported in the literature. We assess the methodological outcomes of our work in three domains: (1) locational diversity of data collected, (2) gender bias in data collection, (3) age bias in data collection. In each case, we compare the diversity of data collected by a citizen sociolinguistic method to that typically reported in the literature to determine if the former can significantly widen our understanding of fat talk. Another aim was to assess efficacy of citizen sociolinguistics in creating meaningful inclusion and dialogue between academics and lay people. As such, we assessed citizen sociolinguists’ (4a) participation in the scientific process and (4b) willingness and ability to act as informal science educators to promote the study’s findings. While detailed data analysis was beyond the scope of this methodology-focused paper (see [47]), we conclude with a brief nod to the sociolinguistic data produced by citizen sociolinguistic research teams (hereafter ‘research teams’), highlighting the ways these contextually rich observations can expand our understanding of fat talk from its primarily psychosocial implications into a solidly sociolinguistic domain.

Materials and methods

Recruitment and retention

For this pilot study, ten citizen sociolinguists, 6 women and 4 men ranging in age from 23–60 (mean = 43.4) years old, were recruited in-person from the research team’s networks, primarily from work, religious, community or personal networks (see Table 1). There was wide variation in employment and educational background, and none had any prior training in social science methods. The citizen science literature notes that high attrition is a common problem across citizen science projects [48]. To address this, our research design required routine, sustained interaction between the citizen sociolinguists and the academic research team. This was done to ensure that citizen sociolinguists had some control over the research as well as feeling involved in the project and its execution [49]. Raddick [45] demonstrated that citizen scientists report higher motivation when they personally identify with the project goals or content. Over the six-week course of this study, only one citizen sociolinguist withdrew. As described below, these elements of success–feeling integrated into and in control of the project–were explicitly built into our research design.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Citizen sociolinguists (CITSLX) profiles.

https://doi.org/10.1371/journal.pone.0217618.t001

Results and discussion

Assessment of data collected by citizen sociolinguists

Citizen sociolinguistics data show fat talk occurs in many locations.

Of 374 conversations invoking physical appearance, 326 included location information. Results show that fat talk occurs among individuals in many day-to-day public interactions across many different locations. Data collected by research teams show remarkable diversity in the locations where people engage in such talk, ranging from spaces where more personal interactions are likely (e.g., residence communities) to more formal locations (professional offices) to highly populated public venues (parks and sporting arenas or holiday gatherings) (see Table 2). Research team members collected data in all of these various locations. Of these many public settings, over 20% of all interactions were collected from restaurants, bars, and cafes. Interactions from residential locations (e.g., neighborhood spaces) were also common (17%) as were interactions from malls, stores, and shopping centers (16%). These locations are far more diverse than those included in most conventional studies of fat talk, which typically use lab or interview settings [12,25]. In short, it appears that fat talk is pervasive in day-to-day life and is an acceptable topic of conversation across many different contexts.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Top 5 locations where appearance-related talk was documented.

https://doi.org/10.1371/journal.pone.0217618.t002

While intuitive on its surface, the ways in which location affects fat talk have not been thoroughly investigated in prior sociolinguistic studies. This topic is particularly important given that different forms of fat talk highlight anxieties about the body and are associated with specific psychosocial disorders, as seen in the association of “muscle talk” with muscle dysmorphia or fat talk with disordered eating or depression [32,36,56–59]. Therefore, understanding the role that location plays in fat talk interactions is an area of future research that would benefit from partnership with citizen sociolinguists and/or citizen scientists generally.

Citizen sociolinguistics data show fat talk occurs among and between men and women.

Data collected by our research teams show that men do engage in discussions about appearance, both with other men and with women (either as initiator or respondent) (see Table 3). In contrast, the literature indicates that fat talk falls primarily in the domain of women, particularly young adult women [23, 60]. Of the 374 conversations that involved physical appearance, the majority were initiated by women (69%). Twenty-seven percent of such observations were initiated by men. Mixed-gender conversations were recorded 20% of the time. To our knowledge, this is a novel discovery. Only 3.5% of observations failed to document the gender of the initial speaker. These results indicate that citizen sociolinguists capture more gender-diverse data than do typical fat talk studies, which, by design, tend to focus exclusively on women’s talk with other women or men’s recognition of it [37, 56, 61–63].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Frequency of appearance-related talk by gender.

https://doi.org/10.1371/journal.pone.0217618.t003

Interestingly, there was no difference by gender of the citizen sociolinguist observer on the odds of recording a male-male interaction (p-value > 0.05, Odds Ratio (OR) contained 1.0). Male citizen sociolinguists were, however, statistically less likely than women citizen sociolinguists to record women-women interactions (OR = 0.113-.691, B = .279, p-value = 0.006), and more likely to observe men-women (mixed) interactions (OR = 1.09–9.09, B = 3.155, p-value = 0.33).

Our findings may stem, in part, from the fact that the research teams had more women than men. Nonetheless, it is clear that fat talk can be initiated by men or women to engage women or men. An area of fruitful research would be to investigate the content and structure of fat talk among men, as it (1) happens consistently in public settings and (2) is a traditionally understudied area in social science research [56, 64].

Citizen sociolinguistics data show fat talk occurs among people of diverse ages.

Results show that young adult, middle-aged, and older people all engage in fat talk (Table 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Frequency of appearance-related talk by age category.

https://doi.org/10.1371/journal.pone.0217618.t004

Conversations involving physical appearance were overwhelmingly initiated by individuals who appeared to be age 30 or under (60%), with 46% of documented interactions being between two younger individuals. Thirty-two percent of observations were initiated by people estimated between the ages of 31–50 (middle aged), with 15% of observations documented as being made toward a younger interlocutor and 10% to another middle-aged individual. Just 7% were initiated by an individual in the older age category (age 51+), rarely toward other older individuals (2%). Only 2% of observations failed to have a documented age estimate for speaker 1. Like the data collected by research teams, the literature also tends to focus on fat talk among younger people. That said, the research teams did capture data from middle-aged and older people—a phenomenon rarely documented in the literature [57–58, but see 61 as an exception]. Notably, older citizen scientists were less likely to observe younger conversants (OR = 0.51–0.442, B = 1.50, p-value = 0.001).

The focus on high school and university students in previous fat talk investigations has given the impression that it is more common among young people [see 20–21, 23, 25). However, the data collected by research teams demonstrate that while it may be more frequent among younger adults, everyone, regardless of age, engages in it. Additionally, our data show that age status is not a barrier to initiating fat talk–young people can engage someone older and vice-versa. Therefore, a fruitful next step would be to investigate responses to fat talk across and among age categories, as this would reveal how the relative ages of interlocutors can have consequences for expressions of politeness and how variation in responses to self-deprecation impacts the function of fat talk among diverse speakers.

Conclusion and broader implications for the future of sociolinguistic research

Citizen sociolinguistics is an emerging field of scientific investigation that offers the potential of scaling up empirical examinations of everyday language use. It additionally offers the opportunity for inclusivity across multiple domains (languages, ages, genders, race/ethnicity, and residences) which are at the heart of everyday language interaction. The benefit of collaborating with citizen sociolinguists is that data collection moves organically with the researcher as s/he moves through day-to-day social environments. To date, there has been a pervasive bias in fat talk research in that it focuses on samples of convenience, usually drawing from undergraduate researcher populations or very specific/targeted groups (e.g., high school students) (e.g., [19, 26, 64]. While such work is deeply informative, it has limited interpretive power in broader contexts. In our study, citizen sociolinguistic research teams captured natural variation in fat talk that revealed greater sample diversity (e.g., influence of gender and age) and environmental contexts (e.g., how location affects fat talk) than has been documented in the past. We found data collected by citizen sociolinguists to be reliable and valid, mirroring that collected by the trained researchers [see 47]. More than reliable and valid data collection, however, the finding that citizen sociolinguists engaged in informal science education in their local communities suggests that this methodology can have effects far beyond the original research question. This finding demands that we ask the following question: Could citizen sociolinguists (and citizen scientists) become informal community science educators that bridge gaps in the realm of public (psychosocial) health outreach programs? It is an intriguing idea.

Beyond implications for the scope and variation of scientific output produced by the research teams, our study showed that citizen sociolinguistics can make unique contributions to the field of sociocultural linguistics. In our case, citizen sociolinguists contributed data and interpretations that broadened our understanding of fat talk in terms of who engages in it and where. The data produced by citizen sociolinguistic research teams confirmed that it is not just important for understanding psychosocial health, but that it is an interaction with clear sociolinguistic importance. We also suggest that this methodology could prove important for other areas of study, including the growing field of medical humanities [67]. The form of the interaction, like other similar interactions (e.g., compliments, apologies, requests, and so forth), is embedded in the perceived social distance and situated relationships of the interlocutors [e.g., 13, 55, 68]. These interactional dynamics expressed as patterns in conversational form not only reflect the relationships among interlocutors, but mediate broader perceptions about an individual’s perceived status as defined by social norms related to factors such as body, gender, or age. With more expansive and targeted studies of spontaneous, naturally-situated verbal interactions in the future, citizen sociolinguists can further advance our understandings of discursive interactions in ways that were previously impossible.

Supporting information

Acknowledgments

We thank the citizen sociolinguists, undergraduate and graduate researchers who participated in this research. We are grateful for support provided by the Virginia G. Piper Charitable Trust for this work.