Rewriting Social Science: The Literary Turn in Qualitative Research

Abstract

In the past few decades, there have been efforts to transform qualitative inquiry, drawing on resources from both imaginative literature and art. There have long been tensions within social science that have encouraged the use of these resources, but the recent “literary turn” is more radical. The assumptions underpinning it are examined, and it is argued that what is most important is the purpose for which modes of expression are employed and how well they serve this. The problem with the literary turn is that it frequently involves substitution of the purposes of art or politics for those of social science.

Social scientists’ research reports have always differed somewhat from those of typical natural scientists, in form not just content. This has been most obvious in the case of qualitative research. Generally speaking, since at least the early 20th century, natural scientists have reported their findings in journal articles using a relatively standard format. ¹ Quantitative social scientists have frequently employed a similar structure in journal articles (with sections such as: research questions addressed, literature review, methods used, findings, discussion, and conclusions). Some qualitative researchers have adopted something like this standard format in journal articles, though rarely very closely. And, in their monographs, the structure of accounts has frequently been shaped by what was taken to be the nature of the phenomena investigated and/or the character of the analysis carried out. For example, some accounts were structured in sections or chapters devoted to the multiple cases that had been investigated, or focused on different aspects of the phenomena studied. Yet others have tracked a chronological sequence, for example a criminological investigation following people from initial arrest, through legal proceedings, and perhaps into prison (see, for instance, Rock, 1973). Furthermore, historically, the products of qualitative research have not been restricted to written accounts: anthropologists have long employed photographs and produced ethnographic films.

This variation in formats employed in qualitative research has long been influenced by the arts and humanities, even though natural science served as its predominant model for much of the 20th century (Redfield, 1962). ² However, it also arose from conflicting ideas about both the task of social research and the nature of the social world. The main conflicts included the following:

1. A focus on the individual case versus a concern with producing general knowledge in the form of scientific laws, abstract social theories, or empirical generalizations about whole populations. Of course individual cases have often been studied on the assumption that they can provide general understanding, by being typical or symptomatic of larger processes, or by challenging previously adopted theoretical assumptions. However, there have also been studies almost entirely concerned with describing an individual person, community, organization, activity, or culture in their uniqueness. In short, in much qualitative work, there has been a concern with what Redfield (1962, p. 111) referred to as “the reconstructive effort to preserve reality within its contexts of unique positions in time, space, and quality [. . .].” This has frequently led to forms of portrayal that are closer to those used in everyday descriptions of personal experience, historical writing, biographies, and imaginative literature.

2. The idea that mechanism-like causal processes can be found in the social world, analogous to those held to operate in the physical realm, versus the belief that social processes are much more complex, contingent, and uncertain in character, or that they take the form of constructed meanings, relationships among which may be noncausal and noninferential. If the first position is adopted, a hypothetico-deductive approach can be used, but the second set of views may require a narrative strategy, documenting the particular paths that processes take in particular contexts, how meanings emerged and came to be related, and so on. Again, this can come close to historical work or even to storytelling.

3. The notion that social science should restrict itself to documenting observable behavior, and drawing inferences from this, versus the argument that what people do is shaped by their understandings of, and feelings about, features of their environment, and indeed their sense of their own identities. The implication of the latter is that the researcher must find some means of accessing and then representing these meanings. This too could lead to the use of forms of reporting that were rather different from the impersonal accounts found in natural science and quantitative research.

4. The idea that social science investigations must rely entirely on the intellect or “rational cognition,” in the sense of explicit inference, versus the claim that human social life can only be understood through the researcher’s experience as a person, including intuition, fellow feeling, empathy, or learning the culture of the people being studied. Furthermore, this latter argument can be applied to how readers of research reports are to gain the understandings offered by research. It may be argued that the sort of impersonal, factual account characteristic of natural science, and of much social science, is inadequate from this point of view; that the researcher must find ways of evoking the appropriate understandings in audiences, and that this may well require drawing on literary, dramatic, or poetic forms and strategies.

5. The assumption that research can simply discover the nature of the phenomena being investigated versus an insistence that any account is a construction that partly reflects the personal characteristics of the researcher. From this, it was sometimes concluded that rather than the researcher being rendered virtually anonymous in research reports, they should be fully included in the picture.

Despite these tensions, most qualitative research for much of the 20th century adopted some kind of rationalist epistemology and realist ontology, along with a factual linguistic mode of expression not very dissimilar from quantitative research and natural science, despite the differences in format. ³ However, this was strongly challenged from the 1980s onwards, with some qualitative researchers explicitly aiming at more “experimental” or “creative” forms of research product, for epistemological, ethical, political, and/or aesthetic reasons (Clifford, 1986; Geertz, 1988). This is what I will refer to as the literary or artistic turn (Sikes, 2012). ⁴

One of the most important landmarks of the literary turn was the publication of Writing Culture (Clifford, 1986), a collection of papers, mainly by anthropologists, focusing on issues surrounding the textual character of ethnographies. This was very much a response to a felt crisis. Anthropology’s distinctive character—a focus on “primitive” or “traditional” societies—had become discredited through influential criticisms of the discipline for its imperialist and neo-imperialist entanglements, as well as its claim to produce objective scientific accounts that are superior to indigenous knowledge. Indeed, the “natives” were now “talking back,” challenging the authority of anthropologists (see, for example, Brettell, 1993; Deloria, 1969).

The response to this crisis on the part of Clifford and Marcus and their contributors was to open up the boundary between anthropology and the arts and humanities. This was at a time when the influence of French theorists, especially Derrida and Foucault, was growing in those fields; and it was this that led to a focus on ethnographic writing, on the texts that anthropologists produce, and to the application of various forms of literary analysis to these. ⁵ These French sources were very different from earlier philosophical influences on anthropology. They amounted to a critique of the Enlightenment conception of reason as being capable of grasping the nature of reality, even in its Marxist and structuralist forms. Instead, the idea of literal, objective, or transparent representation was rejected: it was argued that all texts, including scientific ones, are constitutive in character—they construct their objects, and the meanings of these are unstable. Earlier developments within anthropology—such as the structuralism of Levi-Strauss and the “interpretive,” “symbolic,” or “semiotic” approach of Clifford Geertz and others—had laid the groundwork for this in treating culture as text. Furthermore, especially in his essay on “blurred genres,” Geertz (1980) had treated this new focus on symbols and texts as a means of dissolving the boundaries of anthropology with the humanities, along with amenable parts of sociology, such as the work of Goffman.

Another important theme of Writing Culture was the close relationship between “poetics and politics,” as represented in its subtitle. Concern about the political engagements of anthropology was not new, of course: this had already been prompted in the 1960s and 70s by continuing anticolonial struggles, the civil rights movement in the United States, as well as the rise of second-wave feminism and other “new social movements” (see, for example, Hymes, 1972). However, this concern was later deepened by Foucault’s insistence on the close relationship between knowledge and power, challenging the idea that the pursuit and reception of knowledge can be autonomous and that its effects are emancipatory. Of even more direct relevance for anthropology, in making the link between politics and textual matters, was Said’s (1978) book Orientalism; in which, drawing on poststructuralism, he claimed to show how imperialist ideology had shaped Western academic work over a long period of time. The core idea in Writing Culture’s focuses on texts was that “genre constraints govern the composition of ethnographies” and that these constraints “affirm, if only implicitly, colonial perspectives and asymmetries of power [. . .]” (Rees, 2008, p. 5). In effect, it was argued that such constraints must be abandoned, that allowing “experimental” forms of writing would serve to challenge oppressive perspectives and power asymmetries.

The call for a fundamental reorientation away from a traditional conception of social science, and towards literary or artistic experimentation with discourse and text, and with other media, was not restricted to anthropology: it extended across much of social science. Here it drew on some of the same intellectual sources, but on others as well. Examples include Lawrence-Lightfoot and Davis’s (1997) advocacy of portraiture, Clandinin and Connelly’s (2000) narrative inquiry, and Eisner’s (1991) notion of connoisseurship. And a large number of studies displaying literary innovation of various kinds appeared from the 1980s onwards: Angrosino (1998), Ashmore (1989), Crapanzano (1980), Dorst (1989), Dwyer (1982), Ghodsee (2011), Krieger (1983), Mulkay (1985), Pfohl (1992), Rose (1989), Shostak (1981), and Stoller (1999) (see also Elliott & Culhane, 2017). These works vary considerably in character, whether, for example, consisting solely of dialogue between the researcher and an informant, made up of juxtaposed quotations without the researcher supplying a narrative line; accounts presented entirely in reported speech; ones composed of fictions drawn from research data; and even of a self-undermining account of a PhD thesis and its examination. Also relevant here is the flourishing of autoethnography, of various sorts (Bochner & Ellis, 2016; Denzin, 2013; Reed-Danahay, 1997; Short et al., 2013), as well as what has come to be called “arts-based qualitative research” (Barone & Eisner, 2011; Cahnmann-Taylor & Siegesmund, 2017; Leavy, 2015, 2018). This literary or artistic turn has also extended to the use of poetry (Richardson, 2002), drama (Saldaña, 2005), collage (Gerstenblatt, 2013), and dance (Bagley & Castro-Salazar, 2020). ⁶

As we have seen, these developments not only built on long-standing conflicts within social research but were also shaped by ideas associated with “new social movements,” notably feminism (Gordon, 1988), as well as by various philosophical approaches that started to have a major impact on some parts of social science in the second half of the 20th century. As well as poststructuralism, these included phenomenology, structuralism, postmodernism, and (most recently) “new materialisms”; this last being represented in what Somerville (2020, p. 443) calls “the ‘new new’ of the post-qualitative.” Furthermore, in many of their forms, there was a political dimension to these developments: realist ontology and rationalist epistemology were often charged not just with being based on false ontological and epistemological assumptions, but also with reproducing and supporting the socio-political status quo: in particular, they were denounced as reflecting a dominant white, male, Western, heterosexual, and/or ableist culture that suppresses approaches to understanding the world that are more characteristic of marginalized groups. One aspect of this was the argument that, in conventional forms of qualitative research, informants’ voices are under the control of the researcher, being reported simply when it serves the latter’s purposes. In effect what was proposed instead was a kind of textual “democracy,” in which participants speak for themselves within research texts; this was often related to wider notions of participatory inquiry. Also influential here was the promotion of “indigenous,” African, and other non-Western philosophies. Very often these developments led to rejection of the idea that one account of the world can be superior to another, in the sense of representing it more accurately, in favor of judging accounts according to what are taken to be their political implications, as well as their ethical and/or aesthetic value (for examples, see Denzin & Lincoln, 2017).

In more specific terms, the ideas underpinning these developments included the following:

1. Rejection of the assumption that there can be universally valid knowledge, in favor of the notion of multiple epistemologies, these reflecting the particularities of time, place, and local culture.

2. The argument that all claims to knowledge constitute the phenomena that they purport to represent, rather than simply documenting features of objects that have an independent existence. An implication of this is that research reports not only employ many of the same textual strategies and rhetorical forms as imaginative literature but necessarily are fictions themselves. Those adopting this position often insist that the constructed character of research accounts must be made explicit, some even advocating forms of writing designed to subvert the idea that reality is being described—using strategies similar to those employed in modernist novels (for an example, see Ashmore, 1989)—or ones that display obvious gaps to make clear their imperfection (Price, 2002).

3. An insistence that what counts as knowledge in the West has been shaped by the exercise of power; that it has been designed to serve powerful groups or interests, capitalist, patriarchal, and/or imperialist. Indeed, sometimes the argument has been that any claim to knowledge amounts to an attempt to exercise power over others, and is therefore oppressive. There are two aspects to this. First, it is insisted that claims to knowledge tend to enforce one perspective rather than another as “the truth.” Second, the argument is that by their very nature, these claims position the researcher as knowledgeable or wise while treating others as ignorant or stupid, and thereby set up or reinforce hierarchical relationships.

4. It is argued that the rationalistic, cognitive styles of knowing that are characteristic of Western science do not exhaust the forms of understanding available to human beings and perhaps exclude the most important ones. Here, appeal has been made, for instance, to indigenous cultures, and/or to the perspectives of various marginalized groups within Western societies. One result has been that the task of research has sometimes been re-framed as to give voice to these groups, or to facilitate their pursuit of research themselves.

5. The related idea that what is of value is often ineffable, for example because discourse can never offer fully clear and stable representations (appeal being made here to the work of Derrida); at the very least, it is always perspectival and perhaps even mythical in character. This is part of a broader opposition to the rationalistic character of much Western thought, with its celebration of natural science and technology.

6. The notions of discontinuity and paralogy are championed against any assumption of continuous, or even dialectical, growth of knowledge. In addition, the possibility of dialogue aimed at reaching agreement may be rejected as incapable of bridging the deep divides between ideologies or research paradigms, since these are taken to be incommensurable.

7. The concept of transformative action, in which the new is produced through a creative act or an inexplicable event. This stands opposed not just to conservatism but also to any commitment to piecemeal reform or gradual progress. Here, again, there is an emphasis on discontinuity. ⁷

These lines of argument have a variety of origins, can take different forms, and are not entirely compatible. They have been drawn on selectively by social scientists, and interpreted in diverse ways.

Many of these ideas involve challenges to rationalist epistemology and realist ontology, by relying on skeptical and relativist arguments or on standpoint epistemology (Hammersley, 2023). Also drawn on are notions of research as critique, and as properly activist in character—these deriving from Marxism and Critical Theory but also from the idea that art and literature should subvert conventional understandings of beauty, morality, or truth (the “decadence” of Baudelaire, the radical new beginning offered by Dada and Surrealism, and so on). ⁸ As a result of these influences, those advocating the use of “experimental” literary or arts-based forms have very often seen the purpose of social science not in terms of producing value-relevant knowledge, this being dismissed as impossible or insufficient, but rather as bringing about personal and/or social change. While these forms of representation have sometimes been adopted so as to engage more effectively with key audiences (Richardson, 1990), or to broaden the character that research presentations can take, often what is involved is closer to consciousness-raising—even if this is disguised as the celebration of “subaltern forms of knowledge”—or to delivering “the shock of the new” to undermine the old.

It is important to stress the tensions among these ideas. There is flat contradiction between relativism or skepticism, on one hand, and standpoint epistemology on the other hand (Hammersley, 2023, ch. 5). Furthermore, one important strand underpinning the literary turn has been an emphasis on the importance and value of individual subjective experience—that of researchers and/or of those they study; as exemplified in the fields of life history and autoethnography, especially where those involved are categorized as belonging to marginalized groups. ⁹ However, this is potentially at odds with the constructionist assumptions of structuralists and poststructuralists, or those of Bakhtin, which treat all experience as socio-culturally, and politically, constituted. Furthermore, both of these strands are in conflict not only with the “critical” orientation that is to be found in much work drawing on literary and artistic models but also with the “new materialisms” that have emerged more recently. Moreover, as MacLure (2015) recognizes, the latter two strands are themselves incompatible with one another: if new materialist ideas are applied consistently, it is difficult to see how there could be anything that would count as “critique.”

Evaluation

The fundamental issue that the literary or artistic turn raises is whether qualitative research is intended to be social science; and, if not, what its purpose is. Of course, the term “science” is not unequivocal: there are other conceptions than those modeled on natural science, such as the senses of the term used by Goethe and Hegel, or those offered as distinctive to the social and historical sciences in the writings of Dilthey and Rickert (see Hammersley, 2023, ch. 2). But it would seem that many advocates of the literary or artistic turn wish to abandon science altogether. Even though the task is still frequently taken to be, in part, to produce factual knowledge about the social world, as we have seen other goals have come to be prioritized, especially those concerned with bringing about personal and/or political change, whether this reflects an ethical interest in facilitating authentic personal experience, an aesthetic concern with modes of expression, or a view of research as militant activism that employs literary and artistic devices, along the lines of agitprop. ¹⁰

The waning of commitment to social science on the part of many qualitative researchers has been based, to a large extent, on the idea that the differences between science, literature, and art can be ignored or “erased”; and that doing so leads to a re-vision of science. For example, it is argued that, even in writing conventional research reports, the aim is necessarily to persuade readers of the truth of the findings, so that rhetoric is unavoidable (see Hammersley, 1993, 1995, ch. 5, 2008, ch. 8). However, while there is some truth in this, it is fundamentally misleading. As in the case of a lawyer presenting a case in court, such persuasion must operate within rules that are designed to ensure that the overall goal is served. In the case of the law the goal is justice, whereas in that of academic research it is to discover true answers to research questions. In each case, some rhetorical strategies are legitimate while others must be suppressed because they interfere with pursuit of the underlying goal. ¹¹ So, the central point is that while there are certainly some shared features among science, literature, and art, there are also significant differences, not least in what they aim to produce and the social functions this is intended to serve. These differences must be recognized if each activity is to be done well, in particular their implications for the modes of expression used in research reports should be taken into account.

Thus, the key issue is not the use of literary or artistic textual forms within social research but rather the functions that these are designed to serve. For instance, some kinds of fictions have long been employed in science as devices to facilitate analysis or to enhance the presentation of scientific findings to wider audiences (see Suarez, 2009). Nevertheless, there are fundamental differences between the function of scientific research reports and the aims of artistic representations and imaginative literature. This is not just a matter of how we choose to interpret or “construct” these activities and their functions: particular constructions of them are institutionalized in current practices and cannot be changed at will, nor without this having significant consequences (see Hammersley, 2022).

The main assumption embedded in the institutionalized structures of social science is that research aims to provide true answers to factual questions about the phenomena investigated. To a large extent, this still underpins the practices surrounding the examination of doctoral dissertations, the reviewing of articles for journals, how research proposals are evaluated by funding bodies, and even how book proposals are assessed by publishers. That there have been some moves in all these areas away from that traditional conception of research and its function is true enough, but they all still rely on that conception for their public rationale. Fundamental change in this underlying logic will have negative consequences for the funding of social research and for public attitudes toward it—eroding already attenuated trust.

Central to the traditional scientific conception of research is that it is a collective endeavor to produce specialist factual knowledge that is likely to be more reliable than that available from other sources. Thus, individual studies should be aimed at facilitating the process of collective deliberation in which the validity of competing knowledge claims is rationally determined (see, for example, Kaufmann, 1944; Popper, 1966, pp. 217–220). This requires writing that has a single voice, presenting a clear and coherent argument and providing sufficient evidence and information to allow appraisal of findings by fellow researchers (see Hammersley, 2011, ch. 7). By contrast, art and literature have long been aimed at prompting responses of various kinds on the part of audiences, whether this is boosting religious devotion, increasing sympathy for others, providing erotic excitement, or shocking people out of conventional attitudes. My point is not that social science is superior to literature and art, simply that it is different in its goal and character, and that respecting this difference is desirable for both sides of the boundary. ¹²

What is also often involved in the literary and artistic turn is a reaction against the specialized focus of academic disciplines that seek to answer specific factual questions, albeit value-relevant ones. Instead, it is insisted that what are required are accounts capturing the whole human experience of some person or group, or ones that reconstitute our understanding of social life in aesthetically or politically appealing ways, or provide a basis for, or lead directly into, practical action. Yet academic disciplines arose precisely from the fact that only a selective approach to understanding the world allows the pursuit of deeper understanding. Erosion of this commitment has resulted from increasingly weak boundaries around academic social science. Various factors probably lie behind this: external demands that research must have “impact”; the understandable desire on the part of many researchers that their work should “make a difference”; and perhaps also the rise of mass higher education, which is increasingly concerned not so much with initiating students into disciplines, either to prepare them for becoming academics themselves or as a form of liberal education, but with attracting and retaining them for financial reasons and/or preparing them for future occupations outside of universities. In my view, these are pressures that must be resisted.

Conclusion

In this article, I have reviewed some of the arguments behind the artistic or literary turn that has been advocated and practiced by a significant proportion of qualitative researchers in the past few decades. I suggested that this reflects tensions that have long been present within social science, particularly in anthropology and sociology. However, the recent turn has gone much further than earlier efforts to deal with those tensions, often seeking to blur or erase any boundary with imaginative literature and art. I noted how this was prompted by a variety of political concerns and shaped by philosophical ideas, from phenomenology to new materialisms. I emphasized that these ideas are often in conflict with one another, so that there is no single coherent perspective underlying the literary or artistic turn but rather a motley of assumptions. Yet the discrepancies and oppositions among these ideas have not been given the attention they deserve.

These assumptions challenge the traditional conception of scientific research, on the basis of various kinds of epistemological relativism and skepticism, some form of standpoint epistemology, “critical” orientation, or “new materialism.” But, for reasons I have explained elsewhere (Hammersley, 1995, ch. 1, 2000, ch. 1, 2008, 2023), none of these alternatives is convincing, and they undercut not just social science but any conventional interpretation of the nature of research or inquiry.

In conclusion, I argued for the importance of recognizing the difference in function between science, on one hand, and literature and art, on the other hand. Fictional devices, of various kinds, have long been used within science, even natural science, but in the past, this was usually designed to serve its distinctive aim of producing a cumulative body of factual knowledge, and/or to convey this knowledge to lay audiences. By contrast, much recent qualitative work drawing on literature and art as a model has taken over the characteristic purposes of those domains, being aimed at having an impact of one kind or another on audiences. In my view, this places it outside of social science. This does not, of course, mean that it is without value, but its value ought to be judged against the work of novelists, short-story writers, poets, biographers, and artists: it should not shelter in the domain of social science while simultaneously attacking and undermining it.