Enhancing Intersectional Analyses with Polyvocality: Making and Illustrating the Model

Abstract

Since the inception of the intersectionality framework by feminists over three decades ago, scholars have advanced the analysis and subsequent understanding of peoples’ social locations, identity constructions, and systems of oppression involving gender, ethnicity, religion, class, and caste, to name a few. Considering these axes of differentiation as mutually constitutive rather than only as individual factors has been the single most important innovation. However, intersectionality has yet to reach its potential theoretically, methodologically, and practically. For instance, the framework is rarely applied to social phenomena that extend beyond the confines of a given nation-state. In previous publications, we have addressed this shortcoming by arguing for applying intersectionality across multiple social scales (intimate, regional, national, and transnational). We have shown how any given person’s intersectionality can and often does shift according to the scale of analysis. In this article, we address another important way to strengthen intersectionality—bringing in polyvocality. That is, and drawing upon arguments originally made in postmodern critiques of “writing culture”, publications tend to reflect partial and/or limited perspectives, typically those reflecting researchers’ privileged, authoritative accounts. In this article, in contrast, we include different insider (ego) and outsider (ego’s relatives’ and the researchers’) perspectives. The article includes the theoretical and methodological argument for adding polyvocality to intersectionality and then applies the proposed model to an ethnographic case. We illustrate how intersectional constellations shift from voiced interpretation to voiced interpretation and, in so doing, deepen, expand, and problematize these same analyses.

1. Introduction

Research that engages gender with other axes of differentiation such as race, ethnicity, religion, caste, and class has been immensely enhanced since the arrival of the intersectional analytical approach in the 1980s [1,2,3,4,5,6,7,8]. Considering these axes of differentiation as mutually constitutive rather than as individual, unconnected factors has been, arguably, the framework’s single most important advance. That is, gender could be the only criterion applied to analyses of people’s lives but it would produce a much less rich understanding than an analysis that assesses how gender interacts with and also inflects race, class, religion, and other axes of differentiation. Similarly, these additional axes of differentiation interact with and inflect gender and each other.

For all its analytical advances, however, the intersectional framework has had limitations that have thwarted its ability to reach its potential theoretically, methodologically, and practically. One problem that affects most intersectional research to date stems from involving only a limited set of intersections [9]. No study can realistically consider every “identity” in relation to every other “identity.” By “identity”, we refer to notions of belonging and identifications—either personal or collective—that are not naturally given, but are culturally defined and constituted. Constructions of “identities” are dependent upon socio-historical and political conditions as well as ideological connotations and describe a dynamic concept. This is why we and other scholars [10] prefer the active term “identification” over the static noun “identity.” Whatever the term used, however, what we reference are social practices of differentiation that create and are created by inclusions and exclusions [11]. Scholars initiated intersectionality, specifically arguing that gender should be analyzed in relation to other axes of differentiation. They did not specify which axes to include or at what scale—local, for example, or national. Implicitly, they limited such studies to how these constellations of identifications operate inside but not across nation-states—what has come to be called “methodological nationalism” [12]. Given that we live in an increasingly transnational, globalized world, such analytical limitations seem artificial and without sufficient justification. Thus, in previous publications we have argued for how intersectionality can and should be applied across multiple social scales (intimate, regional, national, and transnational) simultaneously. We have argued for this by documenting how any given person’s intersectionality can and often does shift according to the social scale of analysis (e.g., [2]). We term this innovation “scaling intersectionality” [13].

In this article we continue this intellectual trajectory for strengthening intersectionality, only we argue for another enhancement—polyvocality. We argue that there is an unstated and likely unintended propensity in intersectional analyses to include only the perspective of researchers/analysts. That is, publications tend to reflect partial and/or limited perspectives, typically those reflecting researchers’ privileged, authoritative accounts while excluding, or only minimally including, the perspectives of research subjects. In this article, we seek to issue a corrective by drawing upon arguments originally made in postmodern critiques of “writing culture” (e.g., [14,15]). We include in our intersectional discussions both insider (ego) and different outsider (ego’s relatives’ and the researchers’) perspectives. We illustrate how intersectional constellations shift from voiced interpretation to voiced interpretation and, in so doing, deepen, expand and problematize these same analyses. The article not only makes the argument for bringing polyvocality to scaled intersectionality analyses, but also demonstrates the model by applying it to an ethnographic case.

Developments in Intersectionality Theory, Method and Applications

Critiques of gendered analysis as implicitly reflecting white feminists’ perspectives date to the 1980s. Feminist women of color argued that gender should not be examined without including other axes of differentiation, given that, at least for them, gender could not be extracted from race (e.g., [6,7]). This key insight was revolutionary and served as the foundation for intersectionality. Since that time, the framework and its critique have become standard scholarly practice. Indeed, intersectionality has not only been taken as foundational to much if not most of feminist scholarship, it also has been employed in various ways to expand and extend feminist approaches worldwide [8]. While we and many others have welcomed these developments, our recent work has addressed a number of ways that the framework can and should be improved. We have argued that it should apply beyond the confines of any given country to cover how many people’s lives are lived transnationally [2,13,16,17]. Indeed, many families today have members located in two or more nation-states and thus even the so-called “intimate” social scale of the family is transnational.

To best understand the empirical realities experienced by transnational families whose members have migrated to multiple nation-states, our research design calls for traveling to them and collecting data via different qualitative methods primarily interview techniques and participant observation (see Section 4 for details). We also share an intellectual underpinning—the Gendered Geographies of Power (GGP) framework. GGP was developed precisely to address gender in transnational migration processes [18,19]. We offer a brief overview here of GGP in order to establish its relationship to intersectionality. GGP is comprised of three analytical elements: socio-geographic scales (from the corporal to the transnational), social locations (a synonym for the feminist term “standpoints”), and power geometries (adapted from Doreen Massey [20] to underscore people’s agency given their social locations). The principal rationale behind GGP was to address how gender was not explicitly included in early efforts to examine intersectionalities involving transnational migration. That is and to underscore GGP’s role in comparison to intersectionality’s, GGP was focused on bringing gender “in” to existing studies of transnational migration and migrants’ lives. The intersectional framework, in contrast, was developed to address the need to consider gender along with additional axes of differentiation in studies that focused largely on sub-national phenomena.

This article, however, brings these two related efforts together. What we have argued recently is that intersectional analyses need to address more intentionally the multiple, related socio-geographic scales in which people live their lives. We have termed this approach “Scaling Intersectionality” (SI) [13]. In our theoretical, methodological and empirical explanations of SI we show how the same individual’s or group’s constellation of identifications varies with regard to social status and power when analyzed at the domestic scale, for instance, versus the regional, national and/or transnational scale. Hence, each person as well as each group and institution can and should be examined intersectionally across various scales simultaneously. That work also sheds light on the fact that most intersectional research is conducted within and not across nation-states. This “methodological nationalism” is a major limitation that does not capture sufficiently how people’s intersectional identifications vary according to various sociospatial and temporal contexts. In a subsequent article [2] we applied SI to a concrete case: education as a type of public policy, past and present. It is a case that is both shaped by and shapes people’s multiple social locations.

To date our efforts to enhance intersectionality have been illustrated using ethnographic studies conducted in Malaysia and Singapore (Thimm) and in India and the USA (Chaudhuri), although we specifically focus only on Thimm’s data in this article. We also draw upon previous transnational research by Mahler in Central America and the USA, also ethnographic, which is not incorporated here empirically but informs our theoretical approach. In our research we study transnational families whose members are far-flung within and outside their birthplaces. We collect qualitative data—particularly through interviews and participant observation—because we have found that qualitative data is critical to the work of intersectionality. At its core, the framework is about interpretations, about people’s perspectives on their and others’ lived experiences. It is not the kind of work that lends itself well to hypothesis testing with quantitative methods. When such qualitative research is designed carefully, it not only yields thick intersectional analyses of individual subjects’ social locations across various social scales from their perspective; it can also produce the same quality perspectives from others who are in a position to know any given subject’s positionality. Research methods are mentioned here (but developed for the case study in particular in Section 4) to identify that to date individuals have been our core unit of analysis. In this paper we begin to argue for and illustrate how intersectionality can be expanded beyond individuals to social networks.

Whereas in our most recent papers we have presented scaled intersectional analyses of our interviewees’ lives from our own scholarly perspective, in this paper we embark on a polyvocal approach. This means that we present and offer for comparison not only our perspectives; we also precede our analysis with that of informants’ own perspectives on their social locations. Additionally, we include the perspectives of people close to the interviewees—typically relatives, but also some knowledgeable friends. In sum, the subject’s/informant’s social constellation and status is given three separate times. Each perspective varies from the others in content as well as degree of intersectionality and inclusion of socio-geographic scales. We do not attempt to evaluate each perspective in terms of veracity; rather, we agree with other scholars [21] that all perspectives are partial and even multiple perspectives do not produce comprehensiveness or exhaust all analytical avenues. Rather, our purpose is more modest.

By bringing in multiple, simultaneous intersectional positions and narratives, the aim of this article is twofold: First and most importantly, we believe that such polyvocality more accurately reflects how people understand themselves and their relations to others. That is, people both identify with and are identified by their various axes of differentiation; additionally, our identifications shift based on social context. Second, we believe that polyvocality is a necessary innovation for enhancing intersectional research methodologically. It makes explicit the analytical lens(es) through which intersectional research peers—as insider (ego’s perspective), as close outsider (friend’s, relative’s, foe’s perspective), and/or as more distant outsider (researcher’s perspective). We fully acknowledge that the list of potential perspectives exceeds those we consider in this article. We have chosen a subset focused on kinship/family networks given our own data and the fact that most people know that at least some family member interprets their social relations and standing in addition to themselves. In this sense, we are interested in analyzing how an individual’s experiences influence and are influenced by intersectional identifications and how these, in turn, develop and shift according to the social location of the individual and her close counterparts. That is, to underscore the point, people are not just products of their contexts. They not only conform to laws and norms but also transgress, change, and eradicate them. What all these actions and interpretations mean for different participants varies, however. That is why we argue for polyvocal intersectionality. We are not satisfied if we understand someone’s standpoint when we only investigate her perspective or that of the researcher’s. Rather, we improve accuracy and nuance by gathering and then comparing and contrasting individuals’ views against those of their social networks. In what follows, we briefly summarize the background to polyvocality as a postmodern and feminist research and writing strategy and as an essential element of the “Writing Culture” debate. Further, we discuss various types of insider/outsider viewpoints on the way to a proposed polyvocal methodology that we anticipate will enhance and continue to advance Scaling Intersectionality. Lastly, we apply this very analytical model to one case from our empirical research: that of Thimm’s ethnographic study done in Malaysia and Singapore [22]. We will open the case with the voiced perspective of one of Thimm’s key informants, then switch to the perspective articulated by her mother and father, and end with our scholarly and ethnographic perspective.

2. Questioning Scholarly Authority: Postmodern Questions and Polyvocal Replies

Beginning in the 1980s and spurred by James Clifford and George Marcus’s highly influential volume Writing Culture [23], cultural anthropologists engaged in a raging debate about ethnography. In that discussion, the heretofore presumed objectivity and scientific authority of ethnographic writing was challenged yielding a multitude of critiques of the ethnographic enterprise and of qualitative inquiry more generally. Could these ever be accepted as anything other than subjective perspectives and, if so, might it be better to embrace rather than remedy reflexivity in knowledge production and dissemination? These questions and debates absorbed anthropologists and other qualitative researchers who acknowledged the existence of implicit and/or explicit biases in researchers’ accounts and counterweighted them by emphasizing authenticity and authority in respondents’ voices (e.g., [14,15]; see below for details). They also reflected and helped to produce larger academic inquiries into authority and authoritative presentations given that most knowledge production under globalization emanates from institutions located in advantaged countries. Could scholars occupying privileged positions in universities of the Global North overcome their positionality? What (post)colonial perspectives could counter-balance or permanently imbalance the presumptions and assumptions embedded into the scientific process? The turbulent 1980s were followed by years in which these and similar questions produced nuanced debates around the “crisis of representation” particularly for ethnography. Arguably, Ruth Behar and Deborah Gordon’s [24] Women Writing Culture was one of the most influential books that specifically engaged with gender with regard to critiques of ethnography. In the book’s introduction, Behar ([25], p. 15) asked key representational queries that would prove highly influential for feminist research then and now:

Have ethnographic authority and the burden of authorship figured differently in the works of women anthropologists? What is the cultural logic by which authorship is coded as “feminine” or “masculine”, and what are the consequences of those makings? What kind of writing is possible for feminist anthropologists now, if to write unconventionally puts a woman in the category of untrained wife, while writing according to the conventions of the academy situates her as textual conservative?

These are serious questions, and responses to them are not easily forthcoming. One commonly accepted approach to the issue of “ethnographic authority and the burden of authorship” is to have the scholar/author present herself in a self-reflective way in her writing. That way she can examine herself, her standpoints and her interpretations much as she would analyze “others” in classic ethnography. By using this approach, scholars began incorporating multiple interpretations of their data—polyvocality in its most basic sense. This postmodern research and writing strategy questions the researcher’s authority in the process of representing the so-called “other”. There have been many approaches to operationalizing this critique, this early polyvocality, however (see, e.g., [26,27,28,29,30,31]). They range from merely ensuring that different perspectives are voiced with minimal intervention by the researcher, to having the researcher curate those voices to some degree, to co-producing texts with research subjects and/or activists ([32], p. 38). One influential example is found in Lila Abu-Lughod’s influential [14] book, Writing Women’s Worlds. A post-colonial and feminist standpoint ethnographer, she opted to have participants’ narratives appear without her commentary or interpretation. She does, however, begin her book with a highly theoretical introduction that can be read before, after or never depending upon the reader’s own interest and perspective. In so doing, Abu-Lughod elides requiring readers to socially locate herself as the analytical authority though she also reserves some scholarly positionality. Another important example is Margery Wolf’s A Thrice-told Tale: Feminism, Postmodernism, and Ethnographic Responsibility [15]. In this book, Wolf writes three versions of the same ethnographic event from her fieldwork in Taiwan. One version or “voice” is a purely fictional account—a story she wrote that could explain the event. The second version is a kind of rough draft voice, the version reflected in her own field notes, and the third version reflects her scholarly, academic voice. Wolf explains why she chose this unusual “writing culture” approach in her introduction not unlike Abu-Lughod. However, Wolf highlights how all ethnographic writing is subjective; the question is not if it is or not, but the degree of subjectivity.

Polyvocality also emerges from the perspectives and concerns of feminist research by and about women of color and lesbians. Many of these women write from their own experiences of marginalization within the academy. These women also remark how their communities’ voices have been among the least heard in society. As indicated above, they are the scholars who innovated intersectionality and therewith marked the dominant feminist critique as white, univocal and exclusive [9]. Wairimu Njoya [33] examines this issue in the field of political theory; Amanda Coffey and Sara Delamont [34] for educational research. Their critiques argue that it matters not only which voices are heard, but also, how they are heard. Dominated voices arguably need their own expressive forms but, these scholars argue, they have largely been forced to adapt to dominant voices. Otherwise, they might not be heard at all [35,36]. There is an alternative, however. With polyvocality each voice emanates from its own stage instead of competing from the same platform for access to time and space. With polyvocality voices can be analyzed simultaneously or separately. Thus, it is a flexible, collaborative approach to knowledge production and dissemination, one that implicitly or explicitly challenges classic publication practices. Yet, as Coffey and Delamont ([34], p. 139) argue, “texts are still authored—and selected, collected, edited, presented, written, crafted and read.” Most of this production occurs behind the scenes by academics and therefore is not typically scrutinized. In this sense, polyvocality as a creative alternative strategy of textual production does not yet resolve the major issues; there are numerous contemporary feminist researchers working toward that goal (e.g., [32,33,37]).

2.1. Multiple Voices through “Emic” and “Etic” Perspectives

At this point we take an analytical step back in time from the invention of polyvocality to an important precursor that we wish to use in this article: anthropologists’ differentiation between “emic” (insider) and “etic” (outsider) perspectives. The distinction dates back to the 1960s after a U.S. linguistic anthropologist, Kenneth Pike [38], invented them by drawing upon the linguistic concepts of phonemics (the study of sounds in a particular language) and phonetics (the study of sounds and sound systems in general) (see [39], p. 172). Pike adapted the methodological tools developed to study linguistics and applied them to the study of cultural perspectives. It subsequently has become commonplace among anthropologists to associate a cultural insider’s understanding (i.e., research subject) with the “emic” perspective. In contrast, cultural outsiders’ perspectives are characterized as “etic”. Anthropologists’ work, then, is almost invariably “etic.” As a basic distinction, this emic/etic framework can be helpful, but it is also too binary. It presumes a firm and unproblematic boundary between “us” (the researchers) and “them” (the subjects). This is untenable for those who affirm postmodern, feminist, and particularly intersectional approaches. Therefore, we do not advocate for this type of simplistic “bivocal” approach either. However, we do not wish to throw out the proverbial baby with the bathwater, either.

That is, we deliberately assert here that polyvocal analyses need to differentiate—at a minimum—between insiders’ and outsiders’ perspectives. While only a starting point, this is a meaningful point of departure. In empirical research it would be extraordinary if researchers could collect and voice every possible perspective on a phenomenon. However, this is just not feasible. Therefore and as a conscious compromise, we propose in this article that there be, minimally, three voices in any polyvocal analysis: the emic view (ego or insider) and two etic views. One of these etic views emerges from what we will call for heuristic purposes the “family etic” perspective. This is the view reflected by someone very close to the ego such as a close family member or friend, someone who thus sees the ego from a cultural insider perspective but still remains distinct from the ego. There could, of course, be multiple “family etic” perspectives, but we argue that at least one is important to include as an intermediate voice between the ego and the analyst. Thus, the third voice in our polyvocal model is the researcher’s. The researcher is trained to see both within and beyond the specific cultural context in which the ego is situated. That does not mean that the researcher is immune to subjectivity. Quite to the contrary, we subscribe to the Writing Culture critique that everyone is subjective. That is, even the researcher can be a cultural insider but can provide an analyst’s etic perspective. Existing literature on qualitative research identifies different stylistic genres through which the researcher tells her subject’s story. For example, Warren and Karner [40] identify and distinguish among “analytic-descriptions”, “narrative unfolding”, and “sociobiography”. Through the weaving of individual voices (emic and etic) together to tell a collective story, they show how each one of the genres can be a powerful mode of representation. What we anticipate will make a contribution here is that this type of intentional polyvocality will yield even greater insight into the complexities of individuals’ (and groups’) intersections. We turn to that now.

2.2. Interweaving Polyvocality with Intersectionality

What we propose here is that intersectionality be conducted polyvocally. The primary rationale behind this recommendation is that the intersectional analysis of any given person will vary between her account of herself and others’ accounts of her. We wish to make this explicit in research in order to continue the legacy of feminist scholars who have argued that the scholar’s voice should not be the only voice. Thus we propose that intersectional analyses include—at a minimum—three intersectional accounts for any given subject: the subject (ego/emic), the subject’s family member, friend, close associate (family etic), and the researcher (researcher etic). Of course, to accomplish this polyvocal intersectional analysis requires careful methodological planning and execution.

To operationalize this model methodologically then, we suggest that researchers collect data not only from a particular person or subject (ego). Rather, we need to collect that person’s perspective and at least one additional person’s perspective on the ego’s data. We want to learn not only what a person says that she does and why she does it; we also construe as valuable the perspectives of others close to her who are asked to speak to what she does and why she does it. This second person(s) provides the “family etic” perspective that can and should scale intersectionally. That is, we gather a close outsider’s views of the ego’s social status and locations across different social scales—within the ego’s family, the ego’s additional local groups (e.g., religion, class, caste, race), the ego’s nation, and also, when appropriate, the ego’s status from the transnational perspective. We readily acknowledge that this is a huge undertaking! However, we underscore that people live their lives making most of these calculations on a daily basis. Our social science data collection and analysis skills, however, have not been finely tuned enough to hear how each strand of a person’s own instrument combines with another, let alone how their instrument harmonizes with others. A symphony is a good metaphor for what we recommend. To date we have been lucky to hear the notes of one instrument, or perhaps just one string being played on an instrument of numerous strings. That is the individual’s emic perspective or a scholar’s etic perspective. We now advocate for hearing how multiple strings sound together—the emic, family etic, and scholar’s etic. We hope that this approach appeals to readers, but we know that it is also abstract at this point. For that reason, in this article we not only make the theoretical case for polyvocal scaled intersectionality; we also provide an example of how it is realized in an empirical case.

3. Polyvocal Scaled Intersectionality: Applying Our Model to an Empirical Case

The current article makes an argument for the added value of polyvocal scaled intersectional analyses and then we illustrate the approach with one empirical case. In our previous publications [2,13,16,17] we have already detailed our univocal analysis of empirical cases. In those articles we focused not on subjectivity but, rather, on scaling intersectional analyses—particularly the importance of the transnational scale. In this section, however, we focus on subjectivities, on polyvocal analyses. We begin the case by presenting the viewpoint of a main subject (informant) in the broader study. This is the insider or “emic” perspective of a woman regarding her own social positioning. To her voice we then add a close relative’s viewpoint. Following classic anthropological terminology, this perspective is “etic” as an outsider, but we qualify it as “family etic” to indicate that it is someone who is very close to the original informant. Using data from interviews with the relative, we apply a scalar intersectional analysis to the degree possible given our data. Do family members or close friends view and voice the informant’s social locations similarly or differently from the subject? How and why? What axis or axes of differentiation and scale of analysis are voiced and emphasized by close outsiders versus the ego? Lastly, we offer our (the analysts’) perspective, which is the classic outsider’s or “researcher etic” voice. By deferring our analysis to the end, we anticipate privileging the views of those closest to the social constellation we examine, a strategy that we feel would be thwarted if we began by detailing our own analysis and then offering others’ perspectives. Nonetheless, we acknowledge that making this decision continues to reflect, not eliminate, our privilege.

3.1. Ethnographic Case from Malaysia and Singapore: Doreen Hemmy1

The empirical case we provide is that of Doreen Hemmy, a woman interviewed by Thimm as part of her research into gender and migration between the Malaysian capital, Kuala Lumpur (“KL”), and Singapore [22]. Before proceeding with Doreen’s own perspective, we provide here only the briefest of introductions to the case so that the emic and family etic voices that follow can be understood. In 2003, Doreen migrated as a 15-year-old girl from KL to Singapore, where she completed her secondary school education with an ASEAN-scholarship from the Singaporean government. As explained below, Doreen was viewed as a “foreign talent” whose immigration and education were underwritten by the Singaporean government. After high school, she received a scholarship to an elite college there and, after that, another to the National University of Singapore (NUS). At NUS, Doreen studied English Literature from 2007 to 2011 and graduated with honors. During her undergraduate years she also studied abroad in Europe. As a condition of her scholarship, Doreen was obligated to stay and work in Singapore after graduation. She did so for several years. After more than 10 years abroad, Doreen moved back to KL in 2015, where she rejoined her family, none of whom had emigrated, and enrolled in law school.

3.1.1. Doreen’s (Emic) Perspective

During her first interview in October 2008, Doreen discussed her social background and identified how she socially located herself at that time:

I am Eurasian. My mother is Chinese. My father had a Dutch father and a Sri Lankan mother; she was a Tamil. In Malaysia, the society is segregated into Chinese, Malays, Indians and Others. I am officially classified as Others. So…in Malaysia, I don’t have any race! I’m just Others! (…) But, culturally speaking, I’m practicing mostly Chinese cultural elements, so there is no need to adjust in Singapore [where Chinese are a demographic cultural majority]. In my family, the most important elements are the Chinese and Christian. All my close relatives are Christians. Christmas and Chinese New Year are the two biggest festivities in my family.

Six weeks later, Doreen further elaborated on these issues after she was asked about the different cultural backgrounds of her ancestors. She said,

I have no idea about the Netherlands or Sri Lanka, I don’t know what kind of Dutch or Sri Lankan cultural elements could have been part of my life! I don’t have any connections to such elements. I mean, I really don’t have any race! I have the feeling that I do not belong to anything. The Chinese part is most important for me since I have a very strong relation to my mother’s family. But I don’t even look Chinese. Actually I think that this is a reason why I like postcolonial literature. Because postcolonial theories deal with people living between boundaries and at the margins of society; it’s about how they cross spaces and how hybrid identities are created. This crossing of boundaries and identities, the hybridity… I find this within myself. Postcolonialism is very much about myself and how I live in spaces ‘in-between.’

In Doreen’s commentary from 2008 it is clear that she understands how she is classified ethnically by the societies in which she has lived. In Malaysia she does not enjoy any recognized ethnic group status. She is an “other.” Meanwhile, in Singapore she is deemed Chinese despite her mixed background and her own sense of identification. She expresses and embraces her multiethnic background, her hybrid identity. Nonetheless, at the social scale of the family she feels most strongly attached to her Chinese heritage largely due to her strong relationship with her mother. This identification appears to have been strengthened in Singapore. When asked about her social positioning there, she responded,

It is okay being in Singapore. Actually, this is not an important question for me. I neither feel disadvantaged nor preferred. I myself, here in Singapore, I feel most comfortable with people who speak English [her mother tongue]. Being or not being Chinese is not that important. My life in Singapore made me be so used to the English language. I even increasingly approach people in Malaysia using English and not Malay.

When Doreen was interviewed again in 2010, two years after the initial interview, she spoke about her relationship to her church.

I want to change my group in church. I don’t supervise my girls’ group anymore because they should lead their group now by themselves. I want to participate in a Church group where Malay is spoken. I can practice my Malay then and that [Malay-speaking] group needs some support.

From this interview, then, Doreen expresses something of a return to a stronger KL identification than before.

To summarize Doreen Hemmy’s emic perspective, she recognizes that how she is socially positioned and perceived varies from Malaysia to Singapore. She is able to mobilize different identifications from her background in each society. Ethnically, she originally appeared to embrace her Chinese heritage over her Dutch and Sri Lankan background, and two years later she was moving further toward Malay, at least in her language practice. It is likely that her education has made her more conscious of these phenomena, helping her to find a comfortable mixed identification with “hybridity.” Most consistent is her religion, which, unlike other aspects of her stated identifications, remains firmly Christian. Had Thimm been researching these issues in depth during the time of the interviews, she would have pursued these topics more. However, the need for understanding Doreen’s own perspective (as well as those of close others—family etic) to accomplish our purpose in this article means that the body of empirical data we have at our disposal from Thimm’s original research is less than ideal. We address that issue as a limitation later in the article. Regardless, we hope to have provided Doreen’s voice in this section, albeit not as fully fleshed out as is desirable, and now move into a discussion of her parents’ perspectives as an example of how a family etic analysis can contribute to a more holistic intersectional analysis overall. The parents’ perspective also suffers here from the same empirical limitations as Doreen’s emic perspective, but we provide it as an example—not a full-fledged model—of how those close to a person tend to see that person’s social positioning neither exactly as the person herself does nor as the outside researcher would. To underscore a fundamental point made earlier in the article, our feminist position is that there is no one truth but, rather, all perspectives are partial. Moreover, whereas the inclusion of family etic perspectives may multiply subjectivities, this does not mean that these combine into an “objective” perspective, however holistic. Rather, our view is that data collection and write-up are always informed by the researcher’s positionality, in addition to that of her subjects. Thus, when we collect and present multiple perspectives on any given phenomenon, we may get closer to but never arrive at “the truth.”

3.1.2. Doreen’s Parents’ Perspectives (Family Etic)

In June 2009 Doreen’s mother, Magdalena, was interviewed by Thimm but, again, not with the direct intention to gain her perspective on Doreen’s identifications. That has to be imputed from the interview transcript. What follows are excerpts that provide the most valuable, though partial, information. “Well, she didn’t want to worry us,” Magdalena related, referring to when Doreen was very sick while in Singapore. “I know she was so sick and she stayed in her hostel room. She had very good friends there from the church. So she always had friends who came to see her when she was sick. So that will be sort of a compliment she is not really alone. Because she had a church family, she has very close friends in school.” Later on during the same interview, Doreen’s mother returned to this theme. “She is very independent [in Singapore]! I think she never gives us cause to worry. I always tell a friend, that, you know, she is very strong in her Christian beliefs. She knows her values… she is always kept under the values, the Christian values. So we don’t worry, there is no reason.”

In her interview to this point, Magdalena made no mention of language or ethnicity as Doreen did in hers. Rather, for Magdalena the axis of differentiation she raises several times is religion—that her religious peers constitute Doreen‘s family while she is physically away from her natal family. It is important to note here that Magdalena’s perspective is not equivalent to Doreen’s, though there is some overlap. The emphasis is different. It is not that Magdalena ignores ethnicity and language completely; they are just not as significant to her. That is, later in the interview Magdalena spoke about how Doreen had to repeat some high school in Singapore because “our education system is in Malay. So her education [here in Kuala Lumpur] was always in Malay. And then in Singapore [it was] in English.” She is aware of the ethno-structural challenges Doreen faced but she does not dwell on them. Rather, it is Doreen’s father, Nicholas, who focused on the relationship between ethnicity and schooling when he was interviewed in 2009:

I think basically all this [education-related issues] is a question of opportunities. Because, if you are a non-Malay [like Doreen], you don’t have so many opportunities. So a lot of the business community, they have more opportunities in Singapore. So who are the brains in Singapore? I think the education system in Singapore is open, where people have opportunities. I remember when Doreen was in her first year, she found the education exciting. Because here [in Malaysia], the standard has to be lower because of the Malays. So she always felt that she could obtain more there because of the education system. [Here in Kuala Lumpur] they lower the standard because of the Malays.

Nicholas expresses an understanding of ethnicity and opportunity structures that neither Doreen nor her mother raised with the same sense of importance. Given the purpose of our article, it would have been better had Thimm specifically asked Nicholas to socially locate his daughter and thus we could have directly compared and contrasted his (family etic) perspective with that of his daughter and wife. However, the research design was slightly different. We nonetheless see that each family member emphasizes different axes of differentiation in Doreen’s intersectional constellation. In so doing, each adds more detail to the picture that develops. The sum is greater than its parts, but that same sum continues to be partial. It is now time to add another perspective—that of the researcher/analyst’s etic portion.

3.1.3. Researcher’s Etic Perspective

At the outset, it is important that we note the following: the researcher’s voice we add now echoes a multitude of data gained and interpreted during long-term ethnographic research. Ethnographic research generates extensive and in-depth knowledge through informed exposure and analysis over time. Therefore, we anticipate introducing more detail and nuance to this article’s polyvocal scaled intersectional inquiry into Doreen Hemmy than has been presented to this point. Given the greater abundance of empirical data we enjoy, we will attend more systematically to the scalar aspects of our model. Our discussion begins at the family or intimate scale of analysis before turning to the national and transnational scales. The first axis of differentiation we focus on is gender since this involves one of the case’s more profound social shifts and because Thimm’s data were collected with a focus on gender.

To understand Doreen’s social locations, whether emically or etically, it is critical to know about not only her mother, Magdalena, but also about her aunt Laura and their mother—Doreen’s grandmother. Both Magdalena and Laura enjoy professional jobs. Magdalena is a secretary who has worked in the Finnish embassy in Kuala Lumpur since 1983. She graduated from school with a General Certificate of Education. She joined secretarial school after that but did not finish a degree program. Instead, she began her secretarial career. Later on, she received an offer to study journalism at a university in the Philippines but did not go owing to personal circumstances. Her youngest sister, Laura, was thus the first female in her extended family to earn a university education. She works as an editor at a major Malaysian newspaper. Laura was able to pursue journalism because she earned a scholarship to attend university in Hawaii and migrated there. She exceeded her siblings’ educational achievements given that they, including Magdalena, had to attend to obligations within the nuclear family (see [41], p. 33). She was also advantaged by the fact that her mother, Doreen’s grandmother, made the decision to send her children to English-speaking and Christian schools. A non-Christian, Hokkien-speaking Chinese, Doreen’s grandmother chose to educate her children in these types of schools after British colonial rule ended because she regarded them as most likely to prepare her children for future social mobility. During her interview with Thimm, Magdalena explained that it had been her mother, not her father, who pushed her children to achieve higher and better education. To pay for her children’s school fees, Doreen’s grandmother worked extremely hard in the rice paddy fields.

What made Doreen’s grandmother particularly unusual and especially interesting in terms of gender was that she strove to educate her daughters as well as her sons ([42], p. 132). This was not typical for Malaysia during the 1950s and 1960s; at that time, educating sons was normative. Moreover, this same time period witnessed women’s struggles as part of anti-colonial movements. In the course of these struggles and debates, women, particularly urban middle-class Chinese Malaysians, sought to enhance their social status in the society through formal education ([43], p. 34). Upon Malaysia’s independence from Britain in 1957, institutionalized education took on an increasingly important role for social mobility, and the period also marked greater opportunities for women. These national-level changes affected the opportunity structure faced by Magdalena and Laura, with Laura most able to capitalize on them. In Thimm’s interview with Magdalena, she reflected upon how her mother overcame her own educational disadvantage in order to devote herself to breaking new ground for her children, including her girls:

Yes, education is like the most important thing. Even my mother, she... because her parents were from China. [They were] very old-fashioned, and they didn’t believe in sending the daughters to school. So my mother is about 82 years-old now and she never went to school. But she knew the value of education. So she always would tell us, you know, ‘I can give you money but there is no use [in that]. The best thing I can give you is education.’ So this is the same thing that we... this kind of value that we also have and pass to our children.

We now shift our scalar focus from Doreen’s family and extend it to the national and transnational scales since Doreen’s educational path took her from Kuala Lumpur to Singapore. When asked if Magdalena planned to educate Doreen abroad or if that move had been merely a stroke of luck, Magdalena responded, “Well, we had sort of heard about this scholarship, the ASEAN-scholarship […]. It was always in our mind that, you know, when she gets to a certain age, she can try [to apply]. So...I don’t think there was any other option. Either this or nothing else.” Thus, Doreen’s parents strategized from early on to send her abroad. Why would they take this approach? To answer this question requires a discussion of the educational opportunity structures in Malaysia, to which we turn now.

In Malaysia, being non-Malay, especially ethnic Chinese, is currently a disadvantage given Malays’ demographic and political power there. The converse is true in Singapore, where Chinese dominate ([44], p. 9; [45], p. viii). The ethnic axis of differentiation, however, has shifted over time in accordance with historical processes that merit discussion. Under British colonial authority (1786–1957), Malayan society—the landmass that is now separated into Malaysia and Singapore—was segregated by what the British termed “race” and “culture,” categories still largely operative today ([46], p. 4). A principal vehicle for achieving and maintaining this distinction was schools that segregated students along ethnic lines, but also by gender, class, and religion [47]. There were two types of British, English-speaking schools: National “Free-Schools” and private boys’ and girls’ schools founded by Christian missionaries. In the girls’ schools, the students were taught English language and housekeeping. To educate them in reading, writing, mathematics, and geography like in the boys’ schools came second ([48], pp. 14–16). Social mobility on the basis of British, i.e., English-speaking and mostly Christian, education was therefore most profitable for the urban male Chinese population. The same continues to be true now. At present, approximately 90% of the non-Malay population, particularly those of Chinese heritage, attends independent schools with English, Mandarin, and/or Tamil as the language(s) of instruction. These schools are, in part, still Christian-oriented ([49], p. 69). Another form of current independent schooling stems from the English-speaking, Christian Mission schools, which have existed since British colonialism. Though somewhat changed from that era, they apply a similar model. One example is the Methodist Girls’ School in Singapore, which Doreen attended.

Malaysia’s other major ethnicities, the Malay and Indian populations, usually could not afford the English-speaking and Christian schools during colonialism because they did not get the same support from the colonial government as the Chinese did. The British colonialists not only favored the Chinese, but male Chinese were particularly advantaged since the British employed them as intermediaries in the trading sector ([50], p. 17). This male, English-speaking, Chinese elite was challenged and shattered in the 1970s when the postcolonial Malay-dominated government reversed their privilege and, instead, favored the education of Malays and girls. This shift intersected with Magdalena and Laura’s life trajectories, opening up avenues to their education—as girls but not as Chinese—from which they had previously been excluded. By educating women, the Malaysian government could market itself as “modern” vis-à-vis other nations on the global stage, and, in so doing, it could entice foreign investment into industries that would employ these women, advancing the country economically ([22], pp. 61–63).

This national-level tale of gender, class, and ethnicity takes a different intersectional turn in Singapore. The key to understanding that city-state lies in knowledge of the education policy instituted by Lee Kuan Yew, who was the first prime minister of Singapore from 1959 to 1990. Lee built an educational strategy upon a false racial premise. He believed that intelligence is inherited ([51], p. 197). Consequently, he promoted higher fertility among educated women and birth control for women with less education ([52], p. 51). His party, the People’s Action Party, thus used its education policies to discriminate based on gender, ethnicity, and class. As with Malaysia, a purposive, capitalist, modernization strategy entwined these axes of identification. As a result, the NUS, the country’s elite university, became predominantly female and Chinese. The government also recruited talented female students of Chinese heritage from other countries such as Malaysia by offering them scholarships. Doreen benefitted directly from this “foreign talents” policy and her parents groomed her to take advantage of it. Not surprisingly, then, the outcome is gendered, ethnicized, and classed.

To understand Doreen’s social locations (scaled intersectionality), then, there is a need to understand how policies made and enforced at the national scale impacted decisions made at the domestic scale and how those decisions, in turn, sent Doreen abroad where her social locations would become transnationalized. This impacted both Doreen‘s insider/emic understanding of herself as well as her family’s etic understanding. At the time of her interviews in 2008/2009, Doreen had already been living in Singapore for more than six years. During that time, she lived in a student residence hall on the NUS campus and managed her daily routines independently. She was responsible for her everyday needs and the organization of her activities, even outside the campus. In Chinese or rather Chinese–Malaysian contexts, this kind of autonomy would not be normal for a young woman. On the contrary, daughters’ but not sons’ activities are highly patrolled and controlled by their parents ([42], p. 76). That was also the case in the Hemmy family, despite their more liberal stance on female education. Every time Doreen returned home for a visit, she returned to this gender-based vigilance, while her brother was free to do as he pleased. This gender double standard means that although Doreen used public transport freely in Singapore, she was forbidden to do so after dark in Kuala Lumpur, even if she was accompanied by friends. Her parents justify this asymmetry by arguing that Singapore is much safer for women and thus they do not have to worry about Doreen while she is there. We note here how the domestic scale attempts to extend itself transnationally with regard to gender and its intersections with generation and class (the latter is a function of education). The key may not be how effective that extension is in reality—in whether Doreen’s parents’ wishes are honored or not in Singapore—but its appearance of being maintained in KL. That is, Doreen’s parents need to be perceived by their close family and friends (family etic perspective) as ensuring their daughter’s safety (and, in so doing, safekeeping her sexuality) since this reflects on their family’s (domestic/intimate scale) social locations.

At this stage, we anticipate that readers will have identified the added value of performing polyvocal analysis rather than a scaled intersectional analysis. Doreen has expressed her emic/ego understanding of herself as a cultural hybrid (not as an “other,” as she is officially classified) based on some insights into her own gendered and ethnic mixture. She does not, however, express a profound understanding of how these reflect and refract differently at each scale of analysis. Nor does she express much recognition of how her family has taken advantage of these to enhance her education and, in so doing, her and their social standing. We are quite sure that she does recognize these and would have articulated them had she been asked to do so during an interview. However, most people do not “naturally” articulate such matters in general conversation; they have to be elicited. Doreen, however, identifies strongly as Christian and credits her religion with her values. Whereas her sense of belonging based on ethnicity and language shift throughout temporal and geographical scales, her Christian identity remains stable. Through this stable religious identity, we could get an impression of how she negotiates and shifts her references to ethnicity and language. However, she does not appear to recognize how these stable Christian values play a role in her gendered vigilance in KL. We say this not to criticize Doreen but, rather, to underscore how perspectives are partial. Similarly, Doreen’s parents fill in more gaps than Doreen does in mapping gender and education. Their input led Thimm to pursue a broader and deeper analysis of policies at the colonial and postcolonial national levels. This, in turn, filled in more gaps, but there still are many gaps in the overall story. For instance, while the data we present address changes impacting women in these societies, we leave to one side their effects on men. Furthermore, we did not cover the effects of education and economy-driven policies on the Indian Malaysian population, nor on the Malay Malaysians, especially Malay women, although the latter gain the most benefits. Another lacuna stems from notions of modernity, which developed in the course of educating women and are strongly related to Christianity and English language use within the Chinese Malaysian population.

We recognize that these gaps still affect the appreciation of this particular case but continue to argue that the analytical model we put forth in this paper is additive, not exhaustive. A nuanced analysis of people’s intersectional standpoints is enhanced by examining those intersectionalities simultaneously across various social scales and from multiple perspectives. Our project is to advance intersectionality by adding emic (ego’s) and “family etic” (ego’s close relatives/family) perspectives to the standard or, in our terminology, the “researcher etic” perspective. The added value of an analysis that results from documenting and engaging with the subject’s emic perspective and one or more family etic perspectives is two-fold: it counterweights the power and authority of the scholar in the research and writing process, and it produces a research, writing, and analysis strategy that analytically broadens the current state-of-the-art in intersectionality research. Our intention and goal in this section has been to draw upon our empirical data to illustrate our model and its added value, while being mindful that what we offer is not perfect.

4. Materials and Methods

The data for this article were collected independently and prior to the authors’ collaboration. Data collection was conducted as a transnational and multi-sited case study, however. There are additional methodological commonalities that assist in making the case work for our current argument despite not being designed this way. First of all, Thimm utilizes multi-sited ethnography [53]. This approach calls for doing fieldwork in multiple, diverse, connected places and multi-sited research is recommended for studies in which, like ours, the phenomenon under study is transnational. As indicated above and detailed below, Thimm’s 2008–2009 research occurred in Singapore and then in Malaysia. She conducted follow-up research twice afterward.

Following the recommendations by Marcus ([53], pp. 106–10), Thimm first identified her key informants in Singapore and later worked in their homeland of Malaysia. During her first six months of fieldwork in Singapore, she closely accompanied 13 Chinese Malaysian female educational migrants (“foreign talents”). In the second half of her fieldwork, she traced her interviewees’ social ties back to their families living in each migrant’s hometown. This took her to Malaysia’s capital, Kuala Lumpur, and to other cities like Johor Bahru and Georgetown/Penang. By conducting research in their homelands, she was able to better comprehend the social environments and opportunity structures from which they originated. To complement her work on educational migrants, Thimm also studied eight female Malay Malaysian non-migrant students. She located them at universities in Malaysia and examined why their educational paths did not lead them to Singapore. At the end of her long-term ethnographic work, she focused her work intensively on five Chinese Malaysian families, including Doreen Hemmy’s. All told and over the course of this project, Thimm conducted 55 interviews. Typically, these were more open and narrative than structured. To this data she added participant observations at the students’ different university campuses, in their Christian churches, and in bars and restaurants. She also attended some student parties. In the specific case of Doreen Hemmy, featured in this article and in many of her previous publications, Thimm met with her alone most of the time. Occasionally other family members were present. Thimm complemented the data collected through interviews and participant observation with library research to uncover and understand the broader educational policy history and structures that affected her informants.

The principle of research design followed here was Grounded Theory (GT) [54]. In GT, data collection and analysis typically occur interchangeably and inform each other throughout the course of the research. More specifically, the principle of “theoretical sampling” ([54], p. 73) dictates that the researcher pursues the collection of new data in order to maximize the findings’ conceptual variation and/or to pursue any unanticipated lines of inquiry that emerges sui generis during the research itself. With GT, data collection stops when further sampling is not expected to widen the range in variation already encountered in the data or when the researcher reaches saturation—when the addition of new data does not uncover significant new findings. The possibility of sampling theoretically and altering or expanding the data pool remains open throughout the course of data collection and analysis.

Subjects were recruited for the study using the principles of convenience sampling and snowball sampling. As Handwerker and Wozniac [55] validated experimentally, convenience samples should not be dismissed due to presumed problems with representativeness. Conversely, these authors’ work shows that convenience samples can yield data reflecting the same heterogeneity of the underlying population’s demographic traits, life experiences, and social contexts as that found in random samples. Gravlee [56] has further shown that convenience sampling is desirable when specific populations need to be studied. Representativeness is not expressly sought or even desired in much GT research, or qualitative research more broadly speaking. Rather, validity is more important. That is, this type of research values telling stories accurately over telling stories that are widely shared. This brings us to data analysis and the selection of individuals featured in this article.

In accordance with the principles of GT, the data analysis was conducted inductively. The interview transcripts, verbatim records, field notes, photographs, and statistical data were analyzed using “open coding” in Atlas.ti, a software designed for qualitative data analysis. Notes taken during and after an interview began the process of identifying salient themes. Once a theme (pre-code) was identified, the data for each instance of the theme had a code applied to them. The instances of the theme were subsequently re-read to ensure consistent use of the code and also to identify how themes were logically connected to the other themes. These additional themes were then open-coded in the software. When a passage of data is coded for more than one theme, it creates a digital connection that aids with identifying underlying thematic connections. With GT analysis, themes and their corresponding codes can and often do shift as data analysis progresses. That is, this process can be iterative. This is expected given the inductive, not deductive, analytical approach. As groups of logically related codes emerge, they are named and become higher-order codes, or macro codes. In so doing, GT analyses progress from the “messiness” of the original data to a small number of important themes/topics that really explain the data through meaningful coding and become the basis for much of the analytical write-up. It may take several iterations of theme identification and coding for this cohesive picture to emerge and that is what transpired in this case.

From Thimm’s 13 Chinese Malaysian respondents, she selected Doreen’s case to use here because it represents the main themes emerging from her overall analysis and because her interviews with Doreen and her family contained sufficient (but not optimal) data for the emic/etic analysis we propose in this article. This was not true for all the other informants since Thimm was not able to interview their relatives in every case. To reiterate what was stated previously, Doreen was not interviewed with the express intention of using her narration for an intersectional analysis of her own social position. Instead, she was interviewed about her gendered education path from Malaysia to Singapore. Although in interviewing Doreen Thimm did not emphasize this woman’s self-identifications, Doreen herself discussed them to a degree. She addressed her ethnic “hybrid” identity and how it changed as a consequence of her migration. That is, she understood herself in a transnational social context, which also made her a good person to feature here. In sum, the decision to select Doreen’s case for the current purpose was pragmatic and also driven by theoretical and methodological reasons.

Finally, in an article that makes an argument for polyvocality we wish to return methodologically to address how, when “telling the story,” researchers decide who tells the story. Whose voice(s) and experiences narrate a story given that any story has at least two perspectives—that of the storyteller and that of the listener? Scholars such as Laurel Richardson ([57], p. 39) caution qualitative researchers in this respect. She warns that

Deciding how to present voices and lives is a continuous problem for qualitative writers. Because we use the voices and experiences of the people we study, both for their own sake and as evidence of our credibility, we are constantly making writerly decisions about who gets to say what and how often in the text, and who the narrator talks about, how and how often.

For over three decades, qualitative scholars have debated issues of voice, authority, and knowledge production. We do not anticipate settling these deliberations with this article and its approach. We do hope, however, that research including polyvocality will solve some of the problems identified to this point. Nonetheless, we also recognize that such research involves important trade-offs—trade-offs that affect our work here as well. That is, we did not conduct our research with polyvocality built into the research design. We arrived at the necessity of bringing in multiple perspectives as part of our ever-evolving collaborative work to improve upon an already exceptional framework—intersectionality. Thus, we readily admit that our article shortchanges the “emic” (ego) and “family etic” perspectives because we did not elicit these during the original research for this article. To be sure, Thimm did interview family members of her principal informants. What she did not do is interview these family members expressly to collect the views of the key informants. She collected data that ex post facto has been marshalled for this purpose. Obviously, that is not ideal and is a definite limitation. We also recognize that conducting full-fledged polyvocal research augments typical qualitative work and therefore likely further diminishes the (already small) sample size of the research. However, if validity is the primary goal this trade-off is can be substantiated.

5. Conclusions

Our purpose in this article has been to continue advancing intersectionality as an analytical approach by documenting that there are multiple, varied perceptions of people’s intersectional constellations. To date, scholars’ views have largely dominated the literature but there is reason to include the views of the very people being analyzed as well as others close to these subjects. This commonsense addition of polyvocality to any given discussion or analysis recalls and reinvigorates postmodern critiques of “writing culture” [14,15]. If all knowledge production and, in our case specifically, ethnographies, reflects partial, limited perspectives, then one remedy is to ensure that multiple voices, multiple perspectives are collected, presented and analyzed. In this article we do not just make a theoretical case for polyvocality adding to our own scaled intersectional analyses. We also illustrate how this might be accomplished for a particular case, including not only our own interpretation of our data but also that of sociocultural insiders—ego and ego’s relatives. In this sense we advocate for intersectional researchers to move beyond the typical unit of analysis—individuals—to a small network unit of analysis. We also illustrate how intersectional constellations shift from voiced interpretation to voiced interpretation and, in so doing, deepen, expand, and problematize these same analyses. We readily admit that including three perspectives (emic, family etic, and researcher etic) is not exhaustive. However, in providing these three we hope that we have sufficiently illustrated the added value of polyvocality to intersectional scholarship. This methodological expansion has theoretical ramifications given that, in its very essence, intersectionality is about identifying and being identified. How any given individual views herself will typically vary from how she is perceived by others and these interpretations also shift, we argue, according to the social context and scale in which they are lived and analyzed. Thus, from our vantage point, scaling intersectionality with polyvocality contributes not just to understanding mutually constitutive intersectional entanglements, but also leads to a richer treatment of polyvocality. In this effort, gender continues to be treated as operating not in a vacuum but along with and across other axes of differentiation. The polyvocal approach adds analytical depth and breadth, particularly in conjunction with scaling intersectionality. That is, any given person’s constellations of identifications vary not only by context (the social scale included here), but also by who is viewing and voicing those identifications. Few people in a subject’s social circle are knowledgeable enough to provide perceptions of that subject’s social locations across multiple scales. A person familiar with the subject’s family might not be aware of the subject’s life overseas as a close family member might be. As illustrated for Doreen Hemmy, the very same person’s social statuses can and often do vary depending upon who is voicing them and about what context. And the subject’s intersectional statuses affect those voicing their opinions, including researchers. We advance our careers, and thus our lives, much as our subjects do, by negotiating these identifications carefully across contexts.

By proposing a modification to the research design of some intersectional research and demonstrating how that design can be realized, albeit partially, we hope to initiate (but not to conclude) a methodological conversation about intersectional research. We acknowledge here that the illustrations of the emic and different etic perspectives provided in this article have limitations. First, they are quite uneven. There are various reasons for this; arguably the most important is the fact that the data we draw upon were not deliberately collected to realize a polyvocal intersectional analysis. Rather, we had to dig deeply into the data corpus we had at hand to find material that reflected—but were not elicited specifically to reflect—the emic and family etic perspectives. We researchers, therefore, still maintain our privileged position vis-à-vis our data; we are the ones to choose what to present here and we alone interpret this material without verifying it with our original interviewees. That is another limitation. We, the analysts, still define which respondent to feature in the article and to what extent to analyze her corresponding data. We determine at which point we shift from emic to family etic and then to the analysts’ voices. We have selected which details to narrate and which to exclude. We remain privileged even in the act of ensuring that our voice is not the only voice heard.

As mentioned, another limitation is that we did not specifically design our research around the model we propose in this article because it was only later that we collectively identified the potential of polyvocality. Ideally we would have designed research that would have engaged our key informants and their close family and friends in these identification discussions during the stages of data collection. We also would have checked later to ensure that the excerpts to be published were discussed and verified with their articulators (see [58,59]). This is an advantage of GT which encourages iterative work—often involving going back and forth with subjects about the degree to which their voices accurately and sufficiently reflect their perspectives and experiences. Iterative research typically uncovers new details and even new themes that can be integrated into narratives. To return to the musical metaphor from before, single-stranded melodies can become harmonies and even full-fledged symphonies of voiced meanings.

Although our work certainly has its limitations, we nonetheless highly recommend that even those researchers who do not design their intersectional research to include polyvocal data collection, keep polyvocality in mind during their research. This merely means considering questions such as: Who is speaking and to whom about what? Who is not speaking? Who is speaking for whom? What are people saying about themselves and why? What are they saying about others and why? How does context affect their voiced perspectives? These questions are not new, of course; they are classic qualitative research questions. We hope that drawing a line of consistency between past and potential methodological practice will encourage more scholars to bring polyvocality into their intersectionality studies. Indeed, intersectionality keeps proving its versatility and adaptability such that there may always be room for innovations like we propose here. It is a hallmark that can enhance this incredibly versatile framework.