Discourse analysis of archival data

Chapter 3 showed how discourse analysis beyond the speech event can be applied to ethnographic data. We presented two ethnographic case studies, each of which involved analysis of several events over time, and we showed how configurations of indexical signs across pathways of events helped establish social action. In this chapter and the next two, we apply discourse analysis beyond the speech event to two other kinds of data: archival material and new media. Ethnographic, archival, and new media data are not mutually exclusive, and many research questions will require analysis of two or all three types. In the example of students crying “racist” described in Chapter 2, in fact, Reyes (2013) supplements her ethnographic study by analyzing both new media data in which “crying racist” circulates as a comedic genre as well as archival material that shows what it has historically meant to “cry racist” in the U.S. at different points. Both examples we describe in this chapter focus on archival material, but they also draw on ethnography and new media data to some extent. We separate the three types of data, treating them as ideal types, because there are some differences in how to apply our approach to each.

In this chapter, we apply our approach to archival data, focusing on analyses of discursive artifacts. We describe work by Miyako Inoue (2006) on “Japanese women’s language” and Robert Moore (2007, 2011) on “Irish English accent.” Each of these studies presents compelling analyses, and we make no claim to reanalyze the authors’ data or draw different conclusions than they do. We use our terminology and diagrams to present their discourse analyses, borrowing their thorough accounts to illustrate how our methodological concepts, tools, and techniques can be applied to archival data.

The studies by Inoue and Moore are similar in several ways. They both investigate how ways of speaking become linked to types of people over time—kinds of Japanese women and girls in the case of Inoue, and kinds of Irish people in the caseof Moore. They both focus on reported speech as a key discursive tool for signaling social action. The studies also diverge in two important ways. First, their data come from very different sources. Much of Inoue’s data comes from print media, such as magazine advertisements, many incorporating visual images. Moore focuses on three books, one of which emerged from a collaborative blog. Second, Inoue describes more dramatic change, while Moore describes stability and change. Both describe how a set of emblems comes to mark a certain kind of speaker across a century or more—women being marked by a certain style of speech in Japan, and Irish speakers being marked by a stigmatized way of using English—and how the presupposed identity of this kind of speaker shifts in more or less dramatic ways over time. We have chosen examples that diverge in these two ways in order to demonstrate how our approach to discourse analysis can be used with a wide range of data sources and contribute to various kinds of archival projects. Any analysis of discursive data from historical materials, drawn from longer temporal and broader spatial scales, can benefit from discourse analysis beyond the speech event.