An introduction to critical discourse analysis in education has eleven chapters, three by the editor, Rebecca Rogers. The volume attempts to apply Critical Discourse Analysis (CDA) to various formal and informal educational settings, and to situate CDA within a theory of learning. Most chapters begin with definitions of “central concepts,” and Rogers has sprinkled discussion prompts throughout.

Rogers wants to make the linguistics of CDA accessible to educators – especially graduate students – an important goal given that education is a major arena for the reproduction of power, social relations, and identity formation (indeed, a voluminous literature within sociolinguistics, anthropology, and education has asserted this). CDA, as Rogers rightly points out, has a significant role to play in educational research. Thus, it is good to see that nearly all of the authors in the volume, from diverse academic backgrounds, are educators. Unfortunately, I judge the book to fall short of achieving its goals.

Rogers's introduction covers a lot of useful territory, but it assumes too much background knowledge for an “introductory” text even for graduate students. She and her contributors want to integrate the discourse theories of the linguists James Gee and Norman Fairclough (both of whom have written excellent chapters). Rogers views Gee's and Fairclough's work as complementary, and her discussions of both these scholars take on the formidable task of summarizing and then melding the work of the two. She thoroughly reviews the theoretical orientation of CDA (particularly from its systemic functional linguistic roots) as well as Gee's distinction between “d” and “D” in discourse. Educators do read Fairclough. But they soak up Gee with special eagerness, and for good reason. Gee – whose work has focused on social linguistics and literacies – has written amply about his “d” and “D” distinction for more than 15 years. When Rogers discusses the integration of CDA into educational settings, she defers to Gee's highly influential Social linguistics and literacies and Introduction to discourse analysis. Unfortunately, her discussion fails to contribute any understanding beyond what graduate students might get from simply reading these originals. Furthermore, Rogers's preoccupation with Gee and Fairclough limits her analysis. For instance, while the linguistic anthropologist James Collins has written a foreword for the volume, no serious attention to this field, which also has a substantial “critical” dimension, is found elsewhere in the book. Fortunately, Rogers recognizes the importance of ethnography and context – a neglected area in the CDA literature – and she has been careful to ensure that all chapters address this gap. But no chapter in the book comes close to the nuanced analysis of talk and activity that can be found in recent work in linguistic anthropology or in related areas in sociolinguistics and pragmatics, as represented in the solid insight and analysis on indexicality in Wortham & Rymes's (2003) edited volume on linguistic anthropology and education.

Rogers summarizes criticisms of CDA but ultimately dismisses them as matters of implementation that are irrelevant to the foundations of the theory. For instance, in her chap. 3, “A critical discourse analysis of literate identities across contexts,” Rogers presents a case study of one woman's experiences in special education classes as a child. By providing an extremely detailed discourse analysis, Rogers hopes to demonstrate that CDA is a trustworthy methodological tool in the social sciences if it is “conducted in a standardized manner.” In terms of CDA's efficacy, Rogers also claims that “researchers should avoid starting their data analysis assuming power is embedded in the data” (68, and earlier on 15). But since Rogers does not make clear who might be accused of using “non-standardized” methods in CDA, this case comes off as a straw-man argument. And with reference to her caution about the sites of power, isn't the raison d'être of CDA the Foucauldean assumption that power circulates and is deployed everywhere? And from a CDA perspective, isn't discourse the most manifest form of ideology, the nexus of ideology and power?

A fundamental – and vital – aspect of CDA is that it takes social theory as its starting point. Curiously, no author in this book goes from text analysis to social analysis deeply enough to learn what language tells us about society. For example, in her chapter “Cultural models and discourses of masculinity,” Josephine Young analyzes interviews of the 18-year-old middle-class Latino Chavo, his mother, and his teacher to uncover how his construction of masculinity shapes his participation in school literacy practices. Young scrupulously organizes her transcripts “into lines and stanzas as defined by Gee” (153), designed to help her uncover cultural models. But the chapter fails to ask an equally important question: How does the power of school literacy practices and context shape Chavo's understanding of what it means to be a boy in a literacy classroom? To me, Young's analysis is more philosophical than “critical.” The “cultural models” that Young “uncovers” seem fairly transparent from one or two readings of the transcript, and the preparation of the transcripts for stanzas seems superfluous. Young's methodology has been to analyze the “social language” and “situated meanings” (again, using Gee's theory and method) to determine how Chavo “used language to represent himself in different social contexts” (151). For example, she writes that Chavo “used the phrase it sucks when he described the humanities class to his teammates so that he appeared to be a certain kind of guy” (151). This rather obvious conclusion is typical of Young's method and results. The central problem with this chapter is that no research at all was necessary to know that it sucks is intended to convey “a certain kind of guy.” Extracting cultural models requires more linguistic finesse than Young uses.

CDA relies on interpretive links between everyday texts and institutional and social configurations. This is best exemplified in the current volume by Haley Woodside-Jiron's “Making sense of public policy,” a two-year study of California's public policy related to literacy education. Woodside-Jiron extends critical analysis to policy “to include explanations of how political power constructs and is constructed by larger social practices” (200). She focuses on understanding how radical policy changes in reading education policy occurred in California between 1995 and 1997. What, she asks, made such upheaval possible? Woodside-Jiron examines the process of “naturalization” by combining a Halliday/Fairclough framework with that of Basil Bernstein, and by examining the structural analysis of various policy documents to examine how ideologies are embedded.

Good CDA is rigorously grounded in the text. While not stating it explicitly, Woodside-Jiron adheres to Fairclough's three-dimensional framework for conceiving of and analyzing discourse. But Woodside-Jiron, like so many scholars in this book, leaves herself open to a common criticism of CDA: that it is inattentive to reception. She does not show that the texts she analyzes might not be read in other ways, nor does she demonstrate that her reading is like that of what we might call “average consumers” of these texts.

In “Discourse in activity and activity in discourse,” Shawn Rowe connects CDA with sociocultural approaches to learning. He also wants to demonstrate “a way of transcribing and analyzing talk and activity simultaneously” (80). He provides two transcriptions, demonstrating how his “microgenetic analysis” allows for the “analysis of privileging, appropriating, and rejecting particular members' resources and mediational means as part of activity” (93–94). Rowe does this effectively enough. However, he neglects a substantial existing literature in transcription by linguistic anthropologists such as Elinor Ochs, Charles Goodwin, and John Haviland, the latter two especially important for their work on the nonlinguistic semiotic systems that are important to Rowe.

One question raised even by the stronger chapters in the volume is “Why CDA in contrast to another system of analysis?” For example, Cynthia Lewis & Jean Ketter's “Learning as social interaction: Interdiscursivity in a teacher and researcher study group” will be useful for those involved in teacher education. In this four-year study, the teachers read multicultural young adult literature with Lewis and Ketter “in ways that would help them make decisions about whether and how to teach these works in their community” (118), and to see how texts are shaped by ideological power and how they position readers. The teachers were encouraged to explore their collective assumptions about issues of race and identity and how these assumptions shaped decisions about text selection and pedagogy. The authors analyze several phases of the conversations over time, sometimes pointing out the obvious and at other times delving more deeply. What remains unclear is why CDA – specifically – was the necessary tool for this analysis. In a better article about the same material (Lewis, Ketter & Fabos 2001), Lewis and Ketter do not mention CDA once, but their analysis is the same as in the new work. Why is CDA more useful now? No case is made. Is this analysis critical or just smart discourse analysis?

What I found most troubling about this volume is the extent to which Rogers and her contributors seem to be decentering power from their analyses and replacing it with social identities. That's a perversion of what CDA was meant to do. (Doesn't “critical” mean “about power”?) There is no empirical basis for asserting that “identities” are the universally relevant organizing feature of linguistic interaction, or that “identities” are more salient than “power: in that regard. If anything, “identities” is the narrower, more specialized concept and should therefore generate more suspicion than “power” does as an analytic framework. Yet Rogers seems to find “power” suspect and takes “identities” for granted. Why? What model of power is she proposing? It can't be Foucault's; it can't be Marx's or Althusser's.

An introduction to CDA in education was a good idea, and I looked forward to reading it. Unfortunately, the many problems with the volume keep it from being the introduction that will place CDA at the center of our thinking about education.