The Use(fullness) of Theory
Stefan Bargheer

Theory, as Robert Merton pointed out more than half a century ago in “Sociological Theory”, is not a unified concept. There were at least six different notions of theory in use at the time he was writing, and it seems that since then the list has grown longer rather than shorter. A text identified as good theory from the point of view of one of these notions may not qualify as such based on the criteria of another. I would like to carry the argument one step further: not only does the value assigned to a text as theory change depending on which category of theory is used, but the classification of a text as “theory” in the first place is likewise contingent. Scholarly texts have careers and biographies. Some of the most recognized sociological texts started as one thing and ended up as another – that is, they were labeled as “empirical research” when they were first published and became, over time, re-classified as “sociological theory.”

Erving Goffman’s dissertation research, later published as The Presentation of Self in Everyday Life in 1959, is an example. To say that its status changed over time is not to say that it was initially marginalized. To the contrary, the book was an instant success. It won the ASA MacIver Award (today called the Distinguished Scholarly Book Award) in 1961, and it received favorable (if not enthusiastic) reviews in some of the most notable social science journals, including the American Sociological Review, the American Journal of Sociology, and the American Anthropologist. What made these reviews glowing was the perception that the book was presenting original empirical observations and thus breaking new ground.

There was only one criticism shared among these reviews – the book lacked theory. One reviewer explicitly pointed out that The Presentation of Self in Everyday Life could have benefitted from a discussion of the work of Max Weber. From today’s perspective, such criticism and advice is bewildering. At the present, the text is part of the textbook canon of contemporary sociological theory and taught in introductory courses on the topic. The text itself is a theory, and the reason that it is still read is the very fact that the author did not bother to discuss Weber but instead presented his own point of view.

How could the status assigned to the text change so dramatically—that is, from a text that lacks theory to being itself theory? This transformation unveils something about the way the concept of theory is used in sociology. I want to argue based on this example that theory is not an intrinsic quality of a text – it is a product of the way a text is used. Goffman’s text was used because not only the reviewers considered the book to be original and to offer a new perspective. In the years since it was first published, The Presentation of Self in Everyday Life was applied to new cases: it was put to use in other scholars’ work and thereby transferred from its original research context to other research contexts. My argument is that the book acquired its status as theory through this process of de-contextualization and re-contextualization. Put differently, theory is not an intrinsic quality of a text that makes it particularly well suited for application to a vast number of empirical cases. It is the other way around: texts that are applied to a large number of cases will, over time, come to be referred to as theories.

Texts that have such careers or biographies are texts that I will refer to as theories by use. If it is indeed the use of a text that turns it into a theory, then every text can potentially become a theory. Theories by use are simply the empirical research that sticks. In this sense, today’s theories are yesterday’s empirical research, and, by extension, today’s empirical research will be tomorrow’s theory. The argument is a reversal of the relationship between cause and effect usually employed to make sense of the use of theory. It is at the same time a reversal, at least partially, of the commonly assumed relationship between the producer and the consumer of theory. Theories by use are not produced by authors, but by the audiences who apply their texts to other cases. Theory by use is thus a collaborative enterprise.

These observations do not only apply to Goffman and his famous book. The same can be said for a large number of well known sociological texts, including texts written by American as well as European classical and contemporary authors. When I presented a paper at the Junior Theorists Symposium a few years ago, I analyzed the reception of Max Weber’s The Protestant Ethic and the Spirit of Capitalism. This classic European text experienced the same transformation from empirical research into sociological theory as the more contemporary American text written by Goffman that I use as an example here. The only necessary requirement for empirical research to be relabeled as sociological theory is the passing of time. By definition, a text cannot already have been transferred to other cases at the very moment it is first published. It needs to be used repeatedly, and such a process takes time.

Although there is a long list of notable examples for such theories by use, there are a number of equally notable texts that explicitly aim to be theories. These texts I call theories by design. Anthony Giddens’ The Constitution of Society, published in 1984, is an example. Theories by design, like Giddens’ work, realize their aim to be recognized as theories by relying almost exclusively on texts already classified as theory, including, for instance, the work of Max Weber. A further analysis of theories by design and by use might compare their success rates over time. Are theories by design as prominent in the long run as theories by use, or are most texts that endure the test of time in fact theories by use?

Regardless of how one assesses the resonance of either Goffman’s or Giddens’s work, it is important to note that failing to distinguish between the two substantially different ways that a text can become theory has consequences. The conflation of the two provides an underpinning for the view that theories, both classical and contemporary, are foundational in the sense that every research project should use them as a point of departure, and that the relative value of each project can subsequently be assessed based on how well it relates to these texts. This, in a sense, is the way the reviewer who admonished Goffman for not discussing Weber seems to have thought about theory. From that perspective, theories are what can be called foundational knowledge. The foundationalist view of theory simply fails to recognize that many of our most cherished sociological theories were once identified as empirical research.

Taking this blind spot of the foundationalist perspective into account, I want to make an argument for what I call junk knowledge. Junk knowledge is a relational category. No text is intrinsically junk. Junk knowledge is merely that knowledge which has no direct payoff or use for the empirical case a researcher is working on at the moment. A history of brain surgery in India would qualify as junk knowledge for an ethnographic researcher working on public housing in New York, and vice versa. Although such knowledge has no direct value when it is acquired, it nevertheless has a potential value in the long run. A large range of junk knowledge at a scholar’s disposal improves the quality of his or her empirical research because it assists the formulation of original ideas in the encounter with empirical evidence. Research lives thanks to discovery, the new, the unexpected. Having a wide range of interpretative options available and the ability to transfer knowledge from one context to another facilitates the process of making sense of new data, of translating this data into equally new arguments. Many scholars have addressed the value of such a multiplicity of perspectives, calling it theoretical pluralism, but I would like to make the case for pluralism pure and simple. There is no need for texts to be classified as theories first to make pluralism a worthwhile endeavor.

The notion of theoretical pluralism – if put forward with a strong emphasis on the word theory – is based on the misplaced assumption that texts classified as theories are more foundational than those that are not. Building on the distinction between theories by use and theories by design just introduced, I argue to the contrary: singling out certain texts as foundational impoverishes the creative potential that pluralism offers. It limits the range of knowledge a scholar has at his or her disposal and the willingness to take this knowledge seriously. The assumption that theory is foundational comes at the expense of imagination and innovation. The latter qualities rely on the ability of the researcher to de-contextualize and re-contextualize texts, that is, to transfer texts from one case to another. We should, as a consequence, train sociologists in the ability to make such transfers, not make them memorize the arguments of texts that have once been classified as theories and expect them to relate every research finding to Max Weber or comparable authors.

This does not mean that sociologists should stop reading other scholars’ works and exclusively engage in research of their own. To the contrary, I think that abandoning the notion of foundational knowledge and endorsing the notion of junk knowledge instead would lead us to read more, not less, and to take other scholars’ work more seriously, not less. It would make our discipline even broader than it already is, and ensure that original research does not get overlooked because it aims to be just that—original.

Back / Open Book
Back / home