Researchers who examine media discourse play in a field made wide by the polysemy of both of the constituent terms that name it. While “media” texts appear in newspapers, on television, on the web, on the radio, in film, in books, and in magazines, among others, most studies can only afford to focus on a single medium and on a single layer of media production or media language. And “discourse” presents its usual problems of site and scale: Are we examining cultures, professions, performances, texts, and/or linguistic features? Although all of these can claim to be parts of discourse, most studies can only afford to focus in detail on a single layer.
Anne O'Keeffe's Investigating media discourse confronts these challenges, in part, by focusing on the interview genre in broadcast media, presenting broadcast interviews as performative conversations. O'Keeffe assembled a corpus of 271,553 words, all originally from spoken discourse that appeared in media broadcasts on the radio and television. The broadcasts come from throughout the English-speaking world. She examines radio call-in shows, where listeners get a chance to speak on the air with a host, and television talk shows and political interview shows, where a celebrity or a political figure is interviewed by a host. O'Keeffe wants to find out how conversations change when they are broadcast on the radio and television – to understand, as she puts it, “how interactions differ when they are transmitted through a different medium” (p. 1).
The key unit of analysis for O'Keeffe is the conversation, and much of her theoretical apparatus and study design reflect her understanding of media discourse as talk that has been transmitted. She writes, “Media interactions are essentially conversations that are heard by others” (3). Following this, she draws heavily on conversation analysis and ethnomethodology, especially Goffman's “participation framework,” in her approach to the data. In chapters 2 and 3, she outlines her modes of analysis, emphasizing that researchers need to adopt an eclectic approach when analyzing media discourse. For O'Keeffe this means a cocktail of conversation analysis, genre theory, and corpus linguistics. Chapters 4, 5, and 6 present the findings from the study, with the most interesting discoveries, concerning pseudo-intimacy and identity construction, emerging in 5.
O'Keeffe shows how interviewers and interviewees exploit the discursive resources of everyday conversation in order to perform intimacy and identity for a distant and imagined audience. Specifically, she shows how interview participants deploy pronouns, vocatives, and pragmatic markers to this end. By comparing concordances of her interview corpora with four other corpora of spoken English, O'Keeffe demonstrates that her interview corpus contains a similar frequency of pronouns, especially first and second person. Then she performs a qualitative analysis of the first person plural in all of its cases. She selects random samples from her corpus, presenting results that illustrate interviewers using pronouns to include their audiences in the participation framework: “we've another caller on the line,” “can you tell us about the changes your character is,” and “our guest is Roberta Pearson” (99–100). These findings help ground her more ambitious claims about the ways that the participation framework explains the interactions at work in broadcast media interviews. She uses vocatives and pragmatic markers from her corpus as further evidence to support this argument.
O'Keeffe shows how interview participants use a variety of address terms to refer to one another – for instance, names, titles, and endearments – as they deploy the resources of everyday conversation within a media performance. These vocatives function as resources of politeness and intimacy negotiation within broadcast interviews, helping interviewers and interviewees alternately demonstrate what O'Keeffe calls “pseudo-intimacy” or execute politeness maneuvers to help their conversation partners maintain face within the interaction. For example, she shows how an Irish reporter addresses George W. Bush as “Mr. President” on a number of occasions in a particularly tense interview, where the reporter has put Bush on the defensive with questions about the Iraq war. In this case, the vocative, according to O'Keeffe, serves to balance the reporter's challenging interruptions with a polite and public reminder of the president's status. In other cases, such as radio call-in shows, interview participants address each other using proper names, suggesting an intimacy between strangers.
Another way O'Keeffe supports her argument that interview participants perform intimacy and identity is by analyzing various pragmatic markers used in interviews. She shows how these tokens of conversation management from everyday speech – including interjections, discourse markers, and stance markers – operate in broadcast performances. Providing a number of examples where interview participants use terms like right, yeah, anyway, and exactly in order to manage conversation, O'Keeffe also presents a quantitative analysis comparing the frequency of discourse markers in the three kinds of interview interactions in her study, along with a corpus of everyday conversation. From this comparison, she develops a continuum of pseudo-intimacy across her study categories. With discourse markers as an indicator of pseudo-intimacy, the results show that political interviews are the least like everyday discourse, while the interviews between unknown hosts and callers are the most similar, with celebrity interviews falling somewhere in between.
These findings are some of O'Keeffe's most interesting and revealing because they show how her corpus results can provide specific linguistic data to ground generalizations from previous research about intimacy among media interview performers. In addition, these findings, especially the quantitative ones, demonstrate the added value of corpora in this kind of research. It is one thing to claim that broadcast interview participants interact in ways similar to those in everyday discourse based on certain shared features, while it is another to be able to present the frequencies of these features in comparable corpora. These quantitative analyses help to fill a research gap that often remains open in studies of discourse. Without quantitative data, researchers must identify features and then assume or speculate about their representativeness.
However, O'Keeffe is less persuasive when extrapolating in order to make the case for her eclectic approach to media discourse in general. In particular, there is a lack of fit between one major purpose of the text, presenting a framework for analyzing media discourse, and the specific research question that O'Keeffe poses and the data that she examines. Her framework, depending heavily as it does on conversation analysis, would seem to be a good fit for her study with its specific focus on broadcast interviews. If broadcast interviews are overheard conversations, as O'Keeffe claims, then her eclectic combination of conversation analysis, genre theory, and corpus linguistics may be the appropriate framework for analyzing them. However, broadcast interviews are only one particular genre presented in two specific media. Whether they are typical or exceptional in the wash of other genres presented across a variety of mass media – news stories, profiles, blogs, novels, features, music programs, films, advertisements, weather and traffic reports, RSS feeds, and sports play-by-play, among many others – is an empirical question that is not answered here.
If O'Keeffe's book were entitled Broadcast interviews as performative conversations, then this problem of a limited data set standing in for a broad and complex field of discourse would disappear. However, some problems with the foundational analogy of the research would still need to be addressed. Are broadcast interviews in fact overheard conversations, as O'Keeffe suggests? In my revised title, I propose “performative conversations” because this emphasizes the efforts by participants to perform their identifications and intimacies through their language, something that is well supported by O'Keeffe's data. O'Keeffe's “overheard conversations” emphasizes a foundational analogy between everyday conversations and broadcast interviews that introduces some unnecessary difficulties. On the one hand, this analogy justifies the use of conversation analysis as a method for describing the interactions between interview participants. On the other hand, it leads O'Keeffe to position her object of study as an inauthentic variety of conversation (cf. “pseudo-intimacy”), leveraging everyday talk as a norm. While we may not want to accept Alan Bell's assertion that “people in Western countries probably hear more language from the media than they do directly from the lips of their fellow humans in conversation” (Bell 1991:1), the saturation of homes, businesses, and public spaces with media texts and broadcasts should at least raise questions about O'Keeffe's working assumption of a clear divide between “everyday” and “media” conversations and her assertion that “our primary experience of talk is through everyday conversational interactions” (1).
The series in which O'Keeffe's book appears may help explain the slippage between the title of her text and the research that it reports. The previous titles in the series are Investigating classroom discourse and Investigating workplace discourse. Both the classroom and the workplace are material, institutionally bounded places that a researcher could visit in the manner of an ethnographer. By contrast, “media” is not a place. The newsroom is a place. The television studio is a place. The broadcast booth is a place. O'Keeffe has produced a study that examines the talk that occurs in these kinds of places, talk that also happens to be broadcast over the television and radio airwaves.
Investigating media discourse is most valuable as an application of existing methods of analysis to a novel data set rather than as a new approach to analyzing media discourse. In drawing conclusions at the ends of each of her chapters, O'Keeffe favors verbs like focus on, apply, survey, and look at rather than verbs like discover, demonstrate, and show. In the chapters that discuss her framework, this language suits her. In chapters where she presents her empirical discoveries, this language shortchanges her.
