Thick Concepts in Social Research: What,Why,and How?

Abstract

Thick concepts are concepts that describe and evaluate at once. Academic discussion on thick concepts originated in meta-ethics, but thick concepts increasingly draw attention from qualitative researchers working in the social sciences, too. However, these scholars work in relative isolation from each other, and an overview of their ideas is missing. This article has two aims. The first is to provide such an overview, by bringing together these disparate voices on why thick concepts matter for the social sciences and how to work with them in qualitative social research. The second aim is to reflect on the methodological difficulties of working with thick concepts, by thinking through the example of my research on a specific thick concept—the concept of dependency. The article argues that thick concepts are invoked by social researchers for either epistemological or methodological purposes. It then goes on to claim that if we want to take thick concepts as our sensitizing concepts or as our objects of research, these two purposes really ought to be considered in unison: any methodological approach involving thick concepts must factor in the epistemological challenge thick concepts pose to social-scientific research. To show why—and to consider what this requires from qualitative researchers—I draw on insights acquired during my study on dependency. I end with practical recommendations for working with thick concepts in social research.

Thick concepts are concepts that describe and evaluate at once. Some paradigmatic examples include honest, brave, kind, and stupid. ¹ When we make use of a thick concept, we are not only doing so to refer to a particular property or characteristic of whatever it is we are talking about; we are also expressing a particular valuation or attitude toward that thing. In fact, as Thacher (2015, p. 321) points out, they do one by way of the other: thick concepts “evaluate a phenomenon by describing it, and describe it partly by evaluating it.” In this way, using a thick concept means assuming a certain value, attitude, or judgment. This makes thick concepts different from purely evaluative concepts like good, bad, right, and wrong (which are known as “thin” concepts) and also from purely descriptive concepts like apple, basket, and car.

Academic discussion on thick concepts has tended to limit itself to meta-ethics—the realm of moral philosophy concerned with the nature of moral thought and moral language (Kirchin, 2013b, 2017). But this is currently changing, as thick concepts increasingly draw attention from scholars working in the social sciences (Abend, 2008, 2011, 2019; De Vries & Gordijn, 2009; FitzGerald & Goldie, 2012; Gorski, 2013; Herzog & Zacka, 2019; Kirchin, 2017; Levering, 2002; McMillan, 2017; Parker, 2007; Sayer, 2011, 2017, 2019, Thacher, 2006, 2015). Given that social research is suffused with thick concepts, this interest seems long overdue. Thick concepts figure as our research topics and as our sensitizing concepts (Bowen, 2006). They also figure in our empirical descriptions of the social world and in the words of our participants. Inequality, violence, care, capitalism, poverty, and tabloid journalism, but also strange, hopeful, derelict, efficient, wealthy, and queer—all such concepts are thick. If, as I claim, qualitative research is saturated with thick concepts, we had better reflect on how to work with them.

In this article, I do so in two ways. First, I provide an overview of social-scientific literature on thick concepts, by bringing together some isolated voices on why thick concepts matter for the social sciences and how best to approach them in social research. Second, I reflect on the methodological difficulties of working with thick concepts, by thinking through the example of my research on a specific thick concept (Moi, 2015)—the concept of dependency. Doing so also lets me formulate practical recommendations for approaching thick concepts in qualitative research.

This article consists of three parts. In the first part, I discuss the nature of thick concepts and how they have been mobilized by social researchers. I suggest that thick concepts tend to fulfill one of two purposes for social researchers writing about them. One is epistemological; here, thick concepts are brought in to challenge the fact-value dichotomy or the idea of a value-free social science. The other is methodological; here, thick concepts are brought in as objects of social-scientific inquiry to discover more about the social world. In the second part, I argue that these two purposes really ought to be considered in unison. That is, if we want to take thick concepts as our sensitizing concepts or as our objects of research, we must factor in the challenge they pose to the fact-value dichotomy. I show this by chronicling my research on dependency. In the third part, I consider what is required from social researchers in working with thick concepts in this way, advocating for the importance of the moral imagination. In the conclusion, I list practical advice on how to approach thick concepts in social research.

Two purposes for thick concepts in the social sciences

To grasp why thick concepts found their way from meta-ethics into social research, it is helpful to understand why they came to be considered by ethicists in the first place. Talk of thick concepts in meta-ethics emerged in debates on “moral facts”: can there be, for instance, ethical knowledge as there is scientific knowledge? The common presumption amongst meta-ethicists was that it is impossible to speak of “moral facts” because moral language does not have any empirical reference—it merely expresses values, for example, in terms of emotions (Stevenson, 1972) or prescriptions (Hare, 1963). These values are not empirically contestable (like facts are) and therefore cannot count as knowledge. Foot (1958, 1959) and Murdoch (1956, 1970) brought up thick concepts (although not under that guise) to challenge this presumption. When using thick concepts such as honest or rude, they argued, we cannot distinguish between describing empirical facts about the world and expressing values we hold, as such concepts do both at once. Hence, their belief that such concepts are indicative of moral facts: they reveal, as Chappell (2017, p. 712) puts it, that our world is “shot through with values.” Williams (1985) eventually coined the term “thick concepts” to describe concepts like honest and rude. He defined thick concepts as “determined by what the world is like… and yet, at the same time, their application usually involves a certain valuation of the situation, of persons or actions” (1985, p. 129).

It is also this possibility of ethical knowledge that has moved social researchers to consider and work with thick concepts. These scholars are typically occupied with studying moral life in some way by using social-scientific methodologies—an endeavor exemplified by fields such as the sociology of morality (Abend, 2010), moral anthropology (Mattingly & Throop, 2018), empirical (bio)ethics (Davies et al., 2015), and empirical ethics of care (Leget et al., 2019). In this context, thick concepts are brought up to fulfill one of two purposes.

Epistemological Purposes

A first purpose of invoking thick concepts is to make an epistemological point about the relationship between facts and values (Abend, 2008; De Vries & Gordijn, 2009; Gorski, 2013; Sayer, 2011, 2017; Thacher, 2015). For these authors, thick concepts challenge the notion of a strict dichotomy between fact and value commonly assumed in mainstream epistemologies of the social sciences (Davydova & Sharrock, 2003; Gorski, 2013). This dichotomy, typically associated with Weber’s (1949, 1958) plea for a value-free social science, is an inheritance of the empiricist claim that “‘ethics’ is not about ‘matters of fact’” (Putnam, 2002, p. 19), and hence that factual scientific description can and ought to be uncoupled from value judgments.

By combining description and evaluation, thick concepts are believed to trouble this dichotomy. ² When thick concepts enter social-scientific description, the valuation of the researcher who is doing the describing enters along with them. Consequently, the use of thick concepts precludes the idea of “neutral,” morally detached social science, as “facts” produced in social research are dependent on the researcher’s ethical orientation. Drawing on Murdoch (1970), Thacher (2015) regards thick concepts as “lenses” through which we come to see the world around us. This is because, as Murdoch (1970, p. 27) put it, thick concepts “set up” a world for us, showing us what sorts of things there are in the world and why they are important (Diamond, 2010). Take the example of the concept of human being. Human being “sets up” a world in which some beings (members of the species Homo sapiens) are distinguished from other beings and in which this distinction gains a particular ethical weight: it merits, for instance, a mode of conduct in terms of dignity. To wield the concept of human being is thus to apprehend a world in which members of the species Homo sapiens appear as beings of special importance (Diamond, 1988). In effect, argues Murdoch, the world we see depends on the thick concepts we wield: “the moral concept in question determines what the situation is, and if the concept is withdrawn we are not left with the same situation or the same facts,” she wrote (1956, p. 54). ³ Our observations are thus necessarily value-laden—as they are when we recognize a member of the species Homo sapiens as human being. Consequently, these descriptions also bear the traces of our ethical orientation to the world. As Thacher notes, “[w]hen we view a situation through the lens of such [thick] concepts, our description already contains the seeds of our ethical conclusion about it, and when it comes time to act there may be no separate decision about what moral principles to apply” (2015, p. 328). Put differently, the thick concepts we use to describe a situation also imply their subsequent evaluation. In this way, description and evaluation are intrinsically linked; the descriptions of our observations are themselves ethically charged. Thick concepts thus show us that how we see the world also affects what we see in the first place—so that, for instance, having the concept of human being implies a moral orientation that has us see members of the species Homo sapiens in a specific way, namely, as dignified beings deserving of respectful conduct.

Note that for these authors, this is no indictment of the use of thick concepts: they believe in fact that such concepts are required to get description right. Thacher, Sayer (2011) and Gorski (2013) all point out that attempts to expunge thick concepts from social-scientific description come at the cost of losing scientific accuracy. In spite of (or owing to) their ethical charge, thick concepts end up producing descriptions that appear to be more truthful. As an example, Sayer (2011, p. 45) suggests that saying “thousands died in the Nazi concentration camps” perhaps sounds more “neutral” than “thousands were systematically exterminated in the Nazi concentration camps” but also less factually accurate and truthful. For this reason, Sayer concludes that “[w]e may need to be more evaluative in order to be more objective” (2011, p. 44). Thus, in this argument, thick concepts are used to argue against the Weberian idea of a value-free social science and for the inevitability of morally engaged social research.

Methodological Purposes

A second purpose of invoking thick concepts is to propose a methodological innovation for social-scientific research (Abend, 2019; FitzGerald & Goldie, 2012; Levering, 2002; Thacher, 2006). Thick concepts are made into objects of social-scientific inquiry, usually to illuminate some aspect of the relationship between morality and the social world. Of interest to these scholars is the function of thick concepts in social life and their capacity to be expressive of value. Below, I describe three such methods for studying thick concepts.

The first comes from Levering (2002), in his proposal to regard conceptual analysis as an empirical method. Levering believes that a conceptual analysis of thick concepts can yield empirical knowledge of the functioning of the social world. This is because thick concepts are indicative of social rules. Drawing on Wittgenstein (1953), Levering sketches a picture of language use as essentially rule-bound: mastering a concept means understanding the terms of its use, whether or not it is appropriate to apply it in a given situation. If this is the case, thick concepts become especially interesting for social researchers, as the evaluations these concepts contain are telling of the moral understandings (Walker, 2008) that drive ethical life in a given society. They therefore regulate all kinds of social behavior, Levering contends. Levering’s methodological proposal is to study thick concepts by performing a conceptual analysis of a concept’s correct use; doing so allows the researcher to describe the social rules and customs the concept indicates, as well as the moral understandings that enable this word to become meaningful.

The second approach comes from Abend (2019). Similar to Levering, Abend believes that their feat of simultaneous description and evaluation means thick concepts reflect “the particular character” and the “evaluative perspectives” of the social world in which they circulate (2019, p. 214). However, while Levering focuses on the social rules implied by thick concept, Abend is interested in the institutional “enablers” that render thick concepts meaningful: in the social arrangements that anchor the “evaluative perspective” reflected by thick concepts. The task of the social scientist, he writes, is to “empirically identify, describe, and explain these enablers [of thick concepts]; find out what it is about a society, group, or field that makes particular thick concepts possible” (2019, pp. 213–214). Abend’s methodological proposal consists of two steps. First, the researcher is to develop an empirical account of the use of a specific thick concept. Second, she is to make an “armchair inference,” in which she establishes what the thick concept is “enabled or made possible by”—which, as Abend admits, is not “directly observable” and hence “requires careful analyses of the contexts, varieties, and ramifications” of the uses of the thick concept in question (2019, p. 217).

The third approach comes from Thacher (2006), in his proposal of a “normative case study,” a case study yielding both empirical and normative conclusions. One objective of such normative case studies is to analyze thick concepts in order “to investigate the proper meaning of values” (2006, p. 1632). Unlike Levering and Abend, Thacher’s purpose is decidedly normative: it is about getting a given thick concept “right.” Thacher insists that doing so must be achieved empirically (rather than philosophically), by “examining new cases from a perspective that is simultaneously descriptive and evaluative” (2006, p. 1667). What he suggests, following McDowell (1998), is to examine thick concepts as a moral educator might do, by studying unfamiliar cases:

Conveying what a circumstance means, in this loaded sense, is getting someone to see it in the special way in which a virtuous person would see it. In the attempt to do so, one exploits contrivances similar to those one exploits in other areas where the task is to back up the injunction “see it like this”: helpful juxtapositions of cases, descriptions with carefully chosen terms and carefully placed emphasis, and the like. (McDowell, 1998, p. 85, p. 85)

Thus, Tacher’s methodological proposal is to expand our ethical comprehension of thick concepts either by comparing a new instance of the application of a particular concept to an established instance (“helpful juxtaposition of cases”) or by exploring a range of thick concepts related to the concept in question (“descriptions with carefully chosen terms and carefully placed emphasis”).

Working with thick concepts in practice: The case of dependency

So far, I have shown that social researchers bring up thick concepts for both epistemological and methodological reasons. Strikingly, however, these purposes are rarely considered in unison. Social researchers discussing thick concepts to probe the fact-value dichotomy rightly attend to how thick concepts figure in social-scientific description, but rarely reflect on how to study or work with thick concepts in practice. The potential of thick concepts as objects of social inquiry remains unexplored. Conversely, those who forward thick concepts for methodological reasons as objects of social inquiry largely sidestep the methodological ramifications of the notion of “moral facts.” That is, there is little consideration of how our methodologies are affected by the idea that that we cannot separate a “hard,” empirically observable world of facts from the values we attach to. Levering (2002) and Abend (2019) both appear to approach thick concepts as social phenomena that can be readily observed and analyzed by a morally detached social scientist; in their accounts of thick concepts, the fusion of fact and value bears few consequences for our ability to study them, and their projects are emphatically social-scientific rather than normative. Thacher’s (2006) approach to thick concepts is a partial exception to this observation as his method is explicitly geared toward ethical inquiry; he proposes to study thick concepts empirically in order to improve our ethical understanding of them. Nevertheless, as I will argue, even Thacher seems to underestimate the repercussions of the merging of fact and value for our endeavors to empirically study thick concepts.

In this section, I want to show that the two reasons for introducing thick concepts in social research—epistemological and methodological—cannot be conceived separately. That is, if we wish to take thick concepts as our objects of research or as our sensitizing concepts, we must grapple with the challenge they pose to the fact-value dichotomy; with the fact that how we see the world also affects what we see in the first place. Hence, we must factor in how our understanding and application of thick concepts—and, accordingly, the observations we make with and about them—are themselves dependent on our ethical orientation to the world. If we fail to do so, we may not only miss out on much of the work said concept does for people in the society or form of life in which it circulates; we may also limit ourselves in what we (as social researchers) can come to see of the social world through these concepts. This is why the philosophical notion of thick concepts also has distinct methodological implications; and I wish to unpack these in what follows.

To make my point, I will think through an example (Moi, 2015) of my own work on a specific thick concept: the concept of dependency. This work began as an ethnographic research project for the Dutch Ministry of Health on “experienced client dependency” among people with intellectual disabilities and their relatives, which involved shadowing (McDonald, 2005; McDonald & Simpson, 2014; Van der Weele & Bredewold, 2021) and photovoice (Wang, 1999; Wang & Burris, 1997). The study was to investigate “dependencies, big and small, formal and informal, which have a (large) impact on quality of life, self-reliance, and voice of clients.” While this brief certainly placed “dependency” in the domain of the ethical, it also made it sound like something obvious and readily observable. However, my colleagues and I soon realized it was anything but. ⁴ This is because we found the concept of dependency to be morally charged (Van der Weele, 2021a, 2021b): it did not merely describe particular persons, relations, or situations but also judged them in a particular way. In other words, we realized that dependency is a thick concept.

As the study progressed, we also began to see that dependency is not just any thick concept: it exhibits what Kirchin (2013a, 2017) calls “evaluative flexibility.” This phenomenon occurs “when a thick concept is used to imply a pro idea in one context, but a con idea in another” (Kirchin, 2013a, p. 13). The evaluative content of a thick concept can alternate; the value we ascribe to a thing (an object, a person, and a state of affairs) when using that concept is not fixed. One example of a thick concept with evaluative flexibility that Kirchin mentions is honesty. We might value some person’s honesty in some circumstances, while in others we would vastly prefer discretion (in which case someone can perhaps be said to be “oversharing” or “blunt”). Evaluative flexibility thus entails that a thick concept’s evaluative content can alternate, depending on its application and thus on the ethical orientation that underpins it. But as I will show, there is another upshot to the phenomenon of evaluative flexibility: a shift in a given concept’s evaluative content can effectuate a shift in its descriptive content, too. As a result, depending on the ethical charge a thick concept bears, it can end up “setting up” worlds that are altogether different. Put differently: if thick concepts work like lenses, a shift in one’s moral orientation calibrates the lens, so that the concept in question leads us to discern different things. This poses a methodological challenge to working with thick concepts that authors like Levering, Abend, and Thacher have so far failed to address: the observations we make with and about thick concepts are a product of our moral orientation.

To spell out this challenge (and how we might grapple with it), I offer a rough chronology of my travails with dependency. I describe three applications of dependency, underpinned by three distinctive ethical outlooks. As I will show, each of these applications functioned as a different lens on my fieldwork, “setting up” a different world and enabling me to see different things. By listing these applications of dependency, I wish to suggest that studying a thick concept—and studying the world through it—may require a modification in the ethical outlook we adopt. Throughout the analysis, I will be drawing on Levering’s, Abend’s, and Thacher’s strategies for investigating thick concepts. I will do so both to demonstrate their uses and (in the third section) reflect on their limitations.

Pejorative Applications of Dependency

As mentioned above, the Ministry had requested a study on “dependencies, big and small, formal and informal, which have a (large) impact on quality of life, self-reliance, and voice of clients.” Following Levering, we might ask what sort of moral understanding this use of “dependency” harbors. The Ministry’s definition assumed a negative relationship between “dependency” on the one hand and all kinds of autonomy-related goods such as self-reliance and client voice on the other—goods the commissioned study also implicitly endorsed. (So much for value-free social science!) Framed as the opposite of self-reliance, dependency was deemed undesirable and preferably diminished. In other words, in their use, the concept dependency had a decidedly pejorative undertone. It thus contained a moral understanding of “dependency” as an undesirable state, which was to be avoided where possible.

What, in Abend’s terms, are the “enablers” of this application of dependency? The project’s application of dependency certainly was in line with the political moment of the time. The Dutch welfare state, as many others in Western Europe, is characterized by political discourse that valorizes self-reliance and autonomy (Newman & Tonkens, 2011). In this political climate, “dependency” tends to become a stigmatized condition. Fraser and Gordon observe that “dependency” has grown to become an “ideological term” with “strong emotive and visual associations and a powerful pejorative charge” (Fraser & Gordon, 1994, p. 311). The stigma attached to dependency is thought to reveal itself in the trope of the “welfare mother” (who is believed to be reprehensibly dependent on the state) but also in the lesser social position of people with special care needs who are particularly dependent, such as people with intellectual disabilities (Kittay, 2019). These tropes can only circulate in a society that is hostile to dependency. In such a political discourse, “dependency” acquires a pejorative meaning, akin to an accusation—as it did in the commissioned project.

In its pejorative application, then, dependency describes skewed relations in which the dependent subject relies on another subject or object for fulfillment of an important need; this reliance is subsequently judged as harmful and even depraved. (Note that by describing dependency in this way, I put Thacher’s method at work: I am compiling a list of secondary thick concepts (skewed, reliance, need, harmful, and depraved) to better understand the meaning of the first.) It was this pejorative application of dependency I initially took with me to the field. Its negative evaluative charge primed me to perceive care relations in the group home in a specific way: as lopsided, unequal, and probably problematic. It implored me, in other words, to be suspicious of care; and so it functioned as a lens that “set up” a world of practices and protocols that were potentially disempowering. Who was making plans? Who was making choices? Who was carrying out tasks? As a pejorative concept, dependency effectively had me look for ways in which residents were made dependent, to the detriment of autonomy-related goods. And I found plenty of such examples, in the smallest details of everyday life. I took notes of support workers ladling out potatoes and vegetables, even if residents could perfectly do so themselves, and of support workers pouring out tea and coffee, even if residents had no problem pouring out their own. Such moments of care became suspect to me, signs of micromanagement, of unnecessary coddling, and of overprotection. (Yet more thick concepts!) And not just to me: whenever “dependency” came up in conversation with support workers, they also spoke of it as something that had to be diminished. In their conception of what their work was about, “dependency” was the name of a problem that needed solving. The pejorative application of dependency really did pervade the field.

Appreciative Applications of Dependency

Had we delivered a report on these observations to the Ministry, they undoubtedly would have been pleased: the negative relationship between “dependency” and autonomy-related goods had been confirmed. However, if we think of thick concepts as lenses shaped by our moral orientations, it is not surprising that working with a pejoratively charged concept brought us to perceive and criticize disempowering practices: the normative conclusion had been part of the concept we used to approach the field in the first place. It was as Thacher (2015, p. 328) puts it: “our description already contains the seeds of our ethical conclusion about it.”

However, as the project progressed, we stumbled upon a counterpoint to this pejorative application of dependency in literature on the ethics of care (Held, 1993; Ruddick, 1989; Tronto, 1993), which challenged the Ministry’s use of the concept. In a repudiation of the political hostility toward “dependency,” some feminist scholars have sought to reconsider “dependency” by casting it as central aspect of the human condition. Miller summarizes the position well:

Human finitude necessarily gives rise to myriad circumstances of dependency; illness and injury serve as just two cases in point. In addition, dependencies often function as bookends bracing either side of a life: infants are born into radical dependency, while the elderly often encounter it in their waning days. As such, the human experience of dependency is unavoidable… a fact with noteworthy ethical ramifications. (Miller, 2005, p. 140)

We might follow Levering again in asking about the moral understanding that underpins this use of “dependency.” In the care ethical application, “dependency” still refers to the state of relying on another subject or object for fulfillment of an important need, but this state of reliance is no longer deemed harmful or depraved. Rather, it is considered ordinary and inevitable; a precondition of existence itself. Consequently, “dependency” is no longer contrasted with the various autonomy-related goods I mentioned above but instead becomes the condition under which something like autonomy can become realized in the first place (Mackenzie & Stoljar, 2000). It can therefore be seen as an appreciative application of dependency. To be sure, this appreciative application of dependency depicts it as an essentially neutral condition; however, its ordinariness also paves the way for its subsequent appreciation. Kittay, who is a prominent voice in this field, insists that our condition as what MacIntyre (1999) calls “dependent human animals” begs for a revalorization of dependency, arguing that “our value derives from the chain of dependent relations that make all our lives possible” (Kittay, 2019, p. 163).

In its appreciative application, then, dependency describes inevitable relations of care in which the subject relies on another subject or object for fulfillment of an important need; in some cases, this reliance is even embraced as it enables this subject to survive and flourish. (Note, again, the associated thick concepts: care, reliance, survive, and flourish.) This appreciative application of dependency is not suspicious but welcoming toward care. It therefore enabled me to discern rather different things when conducting my fieldwork. The residents with intellectual disability I shadowed had many needs they could not realistically meet on their own. Those with few support needs might occasionally need assistance in managing their job or organizing their social life; those with many support needs needed help with getting dressed, eating, bathing, and every other aspect of everyday life. The support workers I shadowed helped fulfill these needs, so that the people for whom they cared could live (some version of) an ordinary life. Without support workers, many of the residents would have been helpless. To be sure, these dependency relationships were skewed in some ways, and generally far from reciprocal. But they were also a precondition for flourishing—and as such, even empowering. The appreciative application of dependency formed a lens that “set up” a world of situations in which such connections were foregrounded.

As I noted above, the pejorative application of dependency circulated widely in the field. By contrast, I rarely encountered the appreciative application of dependency explicitly. However, there was a logic to the practice of support that had to assume the appreciative use of dependency at least tacitly. The very idea of the group home itself implies that people with intellectual disabilities need assistance to flourish in everyday life. In this sense, using Abend’s vocabulary, the appreciative application of dependency has institutional enablers as well, albeit hidden in plain view. Therefore, in its appreciative application, dependency also offered a critical angle on the seemingly ubiquitous hostility toward “dependency” I encountered. In light of the inevitability of dependency, the ambition of promoting self-reliance and autonomy by minimizing dependency which I found to be central amongst support staff suddenly became paradoxical: supporting residents to become more independent only reaffirmed the dependency that necessitated care in the first place (Van der Weele et al., 2020, 2021).

Critical Applications of Dependency

We came across one third application of dependency in our fieldwork, this time in conversations with people with intellectual disabilities who participated in our photovoice project, as they spoke to me about life in the group home. To illustrate how they wielded the concept of dependency, I will recount some parts of my conversation with Annemiek (see also Van der Weele, 2021a, 2021b). (Once more, I take my cues from Levering by looking at a concept’s use.) Annemiek is a young woman with an intellectual disability who uses a wheelchair and receives around-the-clock physical care and support in daily life. Reflecting on her dependency, she remarked:

These days, I see my dependency because it’s been that way for years, no longer as dependency, perhaps. So I just take it for granted, while to an outsider it might look like dependency.

Annemiek said that she no longer sees her dependency as “dependency” because over the years, she has come to take it “for granted.” In effect, she is creating a distinction between dependency as it is perceived by outsiders (whom she assumes to apply the concept pejoratively) and as it is experienced by her. Later in the conversation, she explained:

I might be dependent in multiple things, but it doesn’t feel that way… Taking off my coat, putting on my coat, let’s see, that the door is opened for me. When the door is closed, I can’t do anything. So yes, actually, a lot, actually quite a lot of things… But I don’t see that as such… a big problem.

Annemiek admitted being dependent in many ways, as her disabilities resulted in various support needs. Still, these needs did not make her feel dependent because she did not regard such instances of dependency as a “problem.” Only when dependency emerged as a problem did she begin to feel dependent. In other words, dependency only became a useful concept to capture her experience when her dependencies started to matter to her because they began to affect her negatively.

Later in the interview, Annemiek showed me a picture of herself lying on her bed.

Here I’m in bed and I have to wait. When I get up… I have to wait to be taken out [of bed]. So here I am also very dependent. … When I’m in my wheelchair, it’s not as bad because I can do stuff. But when I’m in bed, I’m just staring at the ceiling for hours.

Here we have an example of dependency becoming a “problem” for Annemiek. Annemiek’s need for help getting up in the morning may look like dependency to an outsider (in its pejorative sense), but it does not feel that way to her. Annemiek only begins to call herself “dependent” when her support needs lead her to experience discomfort—because the need is not met, or not met well. In Annemiek’s application, dependency is a circumstantial condition, tied to moments when her care needs are not properly met: it marks dissatisfaction about normal, taken-for-granted dependencies becoming unbearable because the care that is supposed to relieve them falls short. This is why I call Annemiek’s concept of dependency critical: rather than judging a subject’s reliance on care, it judges the practices that respond (or fail to respond) to this reliance. (Note once more the related thick concepts: unbearable, dissatisfaction, and failure.) The moral outlook underpinning this critical application of dependency is marked by an ambivalence toward care and the need for others to survive and thrive. (In this, it differs from pejorative applications of dependency, which express a total rejection).

This critical application of dependency, which was deployed by many participants, also provided me with a third lens with which to approach the field. As a critical concept, dependency drew my attention to moments of tension in care relationships: it “set up” a world of situations in which the needs of residents were met poorly, were ignored, or went unseen altogether. “Dependency” came to describe an experience rather than a condition. The Ministry’s brief of the research project had been to investigate “negatively experienced dependency”: as a critical concept, however, dependency is only used to describe negative experiences in the first place. Only if it is a negative experience will it register as dependency.

Thick concepts and the moral imagination

I have enumerated these different applications of dependency to show that the observations we make with and about thick concepts are a product of our moral orientation. Thick concepts can be fickle: their evaluative content is prone to shifting, and as it does, so does their descriptive content. In this way, a single thick concept can effectively “set up” different worlds for us to discern. Studying a thick concept (and the world through it) may require a modification in the ethical outlook we adopt. As the example of dependency shows, if we fail to do so, we risk more than just overlooking the work said concept does for people in the society or form of life in which it circulates; we risk limiting ourselves in what we (as researchers) can come to see of the social world through these concepts. If, for instance, I had stuck with the Ministry’s initial application of dependency, I would have missed much of the work this concept does in everyday life and also much of what this daily life looks like.

The methodological approaches offered by Levering, Abend, and Thacher enabled me to gain many insights about the concept of dependency. Levering’s method was useful for grasping the moral attitudes and conventions enclosed in the concept; Abend’s method for discerning the institutions that sustain and reproduce these moral conventions; and Thacher’s method for getting to the essence of a particular application of dependency, by specifying the range of its potential meanings. Depending on the research question at hand, each of these approaches have a lot to offer in the social-scientific inquiry of thick concepts. Since Levering and Abend regard thick concepts as indicative of the moral outlook of a given society or social group, their strategies are most useful for questions dealing with the social embeddedness of moral beliefs and practices. Thacher’s approach, meanwhile, is more interested in thick concepts as “moral facts”: his take on thick concepts is suitable for research looking to specify the meaning of ethical notions through empirical examples.

However, my discussion of dependency in terms of its evaluative flexibility has also exposed a methodological challenge to working with thick concepts that Levering, Abend, and Thacher have so far failed to address. While combining their approaches can result in quite holistic analyses of one specific use of a thick concept, none helps to sensitize us to how our moral orientation steers our perception—to that how we see the world also affect what we come to see. If we are to work with thick concepts in social research, we need a method that helps us gather the rich details that let us appreciate different applications of the same concept, so that it can bring into view the different worlds thick concepts “set up” for us as we carry out our fieldwork; one that encourages us to make the modification in the ethical outlook we adopt that is required to discern these different worlds.

So how can we make this modification? It is worthwhile here to return once more to the moral philosophers who pondered thick concepts. Murdoch (1970) and Williams (1985) both argued that grasping a thick concept requires identification with the moral outlook in which said concept might come to be useful, sensible, necessary, or true. To achieve competence in a concept like chaste, for instance, one must participate in the moral life in which discerning chastity is ethically meaningful. Only if one shares the value underpinning the concept does one grasp the concept at hand and apprehend the world it “sets up.” This position offers a bleak outlook for empirical researchers: studying a thick concept would require total immersion in different value systems. (Indeed, the example often brought up in such discussions is that of the anthropologist studying a completely foreign concept.)

Murdoch and Williams are surely right that to see truth in a thick concept is impossible for someone not wholly committed to the moral outlook that underpins it. Achieving such commitment does not seem an attainable goal for social researchers. There is probably only a next-best thing: an attempt to identify in imagination with the moral outlook of a given thick concept. Putnam writes that “to use [thick concepts] with any discrimination one has to be able to identify imaginatively with an evaluative point of view” (2002, p. 39). For Putnam, in other words, grasping the world a given thick concept “sets up” does not necessarily require us to adopt the value espoused by said concepts; rather, it can involve an imaginative identification with the moral outlook in question. I want to suggest, therefore, that the study of thick concepts requires from researchers the exercise of their moral imagination—their capacity to familiarize themselves with an ethical outlook different from their own, to imagine the life lived from such an ethical outlook, and to imagine what might be important in such a life.

Take my research on dependency. In order to discern what Annemiek discerns, for example, when she thinks of herself as dependent, we have to envision the sort of life she lives and what might be important in this life for her concept of dependency to begin to make sense. We must, in the words of Diamond (1988, p. 266), think about what it would mean to participate in “life-with-the-concept”: about how the concept “enters the ordering and articulating of our experience, how it contributes both to the ways we make sense of what we do and what happens to us, and to how we see the shape of our lives” (1988, p. 268). This is because these concepts are not just neutral descriptors of our everyday realities: as moral concepts, they are indicative of how we see ourselves and the world and what in it matters to us. In fact, as I have repeatedly stressed, there is a sense in which they “set up” these worlds in the first place. This is not to say that studying thick concepts in this way will provide some existential insight in another’s authentic, lived experience. But that is beside the point. If we neglect to engage with the moral outlook underpinning thick concepts, we also close our eyes to the worlds these concepts enable and the lives illuminated by them.

So given that thick concepts like dependency determine not only how we see things, but also what we see, they ask from researchers an exercise of the moral imagination—an attempt to imagine the sort of life that would accompany the moral outlook in which a particular thick concept might come to be useful, sensible, necessary, or true. Social researchers are surely already used to this, even if we are unaware of it. This is the everyday moral work of empathy, of humility, and of attention—work in which we are exercising our moral capacities as well as our analytic, scientific ones, and which is practically already part of any qualitative research project.

Conclusion: Working with Thick Concepts in Social Research

This appeal to the moral imagination may seem somewhat abstract. But the example of my research on dependency has also brought out more hands-on strategies for navigating social research on thick concepts, which help to give practical shape to the idea of exercising the moral imagination.

1. Recognizing thickness. The first step in working with thick concepts well is to recognize their thickness—and hence, that they describe and evaluate at once. This may seem obvious, but in practice, it often is not. For instance, many concepts driving practice-oriented qualitative research—health, inclusion, autonomy, recognition, well-being, equality, dignity, and so on—are thick, but their evaluative dimension is seldom made explicit. If we notice that such concepts are also moral concepts, we can begin to adapt our methodologies accordingly.

2. Questioning your concepts. Thick concepts teach us that how we see the world also affects what we see in the first place. For this reason, it is important to remain vigilant of what these concepts—by way of our moral orientation—may obscure from our vision. It pays off to be wary of dominant discursive uses of the thick concept under study—especially if these are provided by an external source, such as a commissioning party. As conceptual “lenses,” these uses may limit your field of vision and inhibit your observations. This also means being open to questioning your own use and understanding of the thick concept under study. I found three principal ways to challenge these, all already part of our research practice: fieldwork experiences, new literature, and collaboration with other researchers.

3. Gathering perspectives. Thick concepts are used by different people to express different things, so that they come to reflect different moral orientations. For this reason, research on thick concepts benefits from an expansive and triangulated data collection that brings into view a breadth of perspectives. As my research on dependency suggests, ethnographic methods and forms of participant observation work well because such methods allow for interactions with a wide variety of participants and for witnessing a wide scope of events and encounters. Additionally, methods that focus on acquiring the perspective of individuals—such as shadowing, photovoice, and narrative analysis—also seem particularly suitable.

4. Analyzing language use. Depending on their use, the meaning of thick concepts can shift, both in their descriptive and evaluative content. Analyzing how your research participants use specific thick concepts may provide new insights in a concept’s meaning, the rules that guide a concept’s use, or a counterpoint to established discursive uses—like Annemiek’s use of “dependency.” It may subsequently also indicate alternative moral orientations. To gather relevant information, it helps to ask participants for their definition of thick concepts and to pay close attention to moments in which they introduce new thick concepts in conversation. This strategy employs the conceptual analysis approach offered by Levering, but builds on it by insisting that abstract conceptual analysis does not suffice: language use in the field can differ from your expectations and must be taken equally seriously. To distinguish between uses, it is useful to follow Thacher and define different uses in terms of other thick concepts—as I did with the different applications of dependency.

5. Crossing disciplines. Interdisciplinarity is key in working with thick concepts. If thick concepts are both descriptive and evaluative, social research involving thick concepts is a project of social science as well as ethics. As Thacher argues, this means moral philosophy on thick concepts greatly benefits from empirical research. But to add to this, it also means that social research on thick concepts can greatly benefit from the input of moral philosophers. Hence, it may be worthwhile discussing or even analyzing your data teamed up with ethicists.

6. Reflecting on values. As I have argued, thick concepts almost inevitably enter our research—and along with it, so do our values, judgments, and attitudes. This does not mean our findings cannot be trustworthy. As Sayer (2011) argues, our descriptions can become all the more truthful for being value-laden. It does mean, however, we should remain mindful of what our moral orientation enables us to see, as well as what it veils from us. Working with thick concepts requires reflexivity (Finlay, 2002) to appreciate both how our values steer what we come to see and what we might fail to notice.

These insights might seem familiar to some readers. After all, many of us already work with thick concepts, even if not typically under that guise—if not by way of sensitizing concepts, then at least in how we configure our descriptions of the world and views of our participants. Nonetheless, even if we do not outright use thick concepts in our research designs, the discourse on thick concepts is worth our while because it provides a clear vocabulary to think about some challenging aspects of qualitative research all of us have to grapple with: for instance, about managing the balance between empirical and normative claims (Leget et al., 2009; Molewijk et al., 2004) and about the politics of our descriptive practices (Vitellone et al., 2020). Since so much qualitative research already implicitly engages with thick concepts, I hope this article will motivate other social researchers to draw on their vocabulary and join their social-scientific discussion.