Analysing sociolinguistic variation fulfills the expectations of a modern textbook. It includes exercises (one at the end of each chapter) and breakout boxes containing tips, summaries, tables, and data analysis examples. However, I can envision myself using it more in my own research than in the classroom. This is not because it lacks much as a textbook, but because its detailed explanations and breadth of coverage concerning variationist statistical analysis should direct scholars to more thorough and responsible methodology.
The book has a preface, 12 chapters, a concise glossary of terms, references, and an index. Chapters 5 through 9 provide the bulk of the book. Chapter lengths range between 9 and 33 pages, making the book well suited for a classroom. The book clearly foregrounds data: data collection (chap. 2) from sociolinguistic interviews (chap. 3), which necessitates data management (chap. 4). From this point in the variationist analysis, the researcher must assess the linguistic variables (chap. 5) by first creating and then testing sociolinguistic hypotheses (chap. 6). To carry out this hypothesis formulation (and reformulation and retesting), Tagliamonte strongly recommends the variable rule program (chap. 7), detailing the logistics of a quantitative analysis (chap. 8), including how to do distributional analysis (chap. 9) and multivariate analysis (chap. 10). As the last steps, Tagliamonte advises the researcher about interpreting the results of the statistical analysis (chap. 11) and about connecting these results to a complete social and linguistic analysis (chap. 12).
In the early chapters, Tagliamonte expounds on how to gather data, plan a quantitative study, and conduct a sociolinguistic interview. This methodology has not been detailed often enough in the sociolinguistic literature. As Tagliamonte points out in the preface (p. ix), knowledge of variationist methodology has most often been passed on by word of mouth, from teachers to students, like a system of masters and apprentices. In chapters 1–4, Tagliamonte provides the topics all researchers must wrestle with: how to find subjects, what questions to ask in an interview, and what sets of information a researcher should end up with. Prominently, she keeps all these topics related to their effects on the sociolinguistic research questions. Some of these topic discussions are short; but this may be best, given that this is a textbook. A companion website, not listed by URL in the book itself but available on the press's site (multimediafluids.uam.elogim.com/catalogue), does have appendices with further information, such as Tagliamonte's current transcription protocol. The practicality of this kind of aid makes this book valuable on a regular basis. Accordingly, I have two copies of the book, one for my office and one for the West Virginia Dialect Project lab.
In chapters 5–7, Tagliamonte delves into the rational substance of variationist sociolinguistics. She examines concepts of the linguistic variable and works through how it should be examined. She strongly suggests that “circumscribing the variable context” of the linguistic variable be both a starting point and ending point in variationist analysis. In a research project, this step can be an important contribution to the study of language. The explicit focus on formulating a hypothesis is important for students because it emphasizes that the methodology is not the end goal in itself: Sociolinguistic work is labor-intensive to the point that researchers can develop amnesia about the larger goals of the research. These chapters also provide helpful small bits of advice about conducting variationist analysis. For example, in extracting tokens of the variable and its surrounding environment in a typescript, set the token in capital letters, because with several hundreds (or thousands) of tokens taken over weeks (or months), visually marking the tokens saves time. With this kind of practical advice, Tagliamonte fulfils her description in the preface (x) that this work is “one user's tried and true manual of best practice.”
As a researcher, I am drawn to chapters 7 through 10, which explain how to work Varbrul and why to work it in certain ways. The reader should have data on hand and a Varbrul program loaded on the computer. Reading these chapters requires active participation, which makes the theories and methodologies of variationist analysis come to life. Tagliamonte's methodological directives are integrated with sociolinguistic goals. She is acutely aware that the sociolinguistic questions being pursued shape the minutiae of the quantitative analysis: “The coding schema is where your analysis resides” (111). In the first six pages of chapter 7, she provides a theoretical discussion and defense of the Varbrul program, reviewing important components such as the variable rule, the null hypothesis, and the concept of likelihood. The remaining 23 pages of chapter 7 describe the terminology of Varbrul analysis, including step-up/step-down analysis, statistical significance, and the (dis)favoring of factor weights. Other general Varbrul topics are also detailed here. For example, the first thing I looked for when I received the book was how to best compare Varbrul runs; the answer is found in the practice part of the chapter (149).
Chapter 8 details practical step-by-step aspects of Varbrul analysis, such as how to construct the token file, condition file, cell file, and results file. It also introduces the idea of cross-tabulation, the preliminary procedure for the “number crunching” discussed in the next chapter. Chapter 9, “Distributional analysis,” describes how Varbrul provides frequency statistics for the dependent variable through cross-tabulations, and the importance of such statistics as a means to check the quality of the data and to begin to intimately understand its patterns. Chapter 10 centers on multivariate analysis, assisting the reader in assessing the output files and playing with the program to gain a better understanding of its statistical parameters.
Chapters 11 and 12 delve into how to transform the multitude of results created through the previous chapters into good sociolinguistic argumentation. Tagliamonte holds up three results as the best lines of evidence for determining which factor groups are most important to understanding variable patterns of the feature under analysis: the statistical significance of the factor group (whether it has any effect or not); which factor group is most important (having the largest range of factor weights); and the order of the factors within a factor group (constraint ranking). The import of constraint ranking is that “the hierarchy of constraints constituting each factor is taken to represent the variable grammar” (240). These lines of evidence allow the researcher to test hypotheses about expected patterns for linguistic and social factor groups and allow for comparisons between speech communities. In describing what to provide in an article or a presentation, Tagliamonte enumerates a set of information (such as input Ns) that every researcher should make known. One of my hopes for the impact of this book is more uniform reporting in quantitative analyses so that different studies can be compared more easily.
Having heard Tagliamonte give numerous presentations and having read her journal articles over the past decade, I can say that the voice of the author is very much Tagliamonte's own. She is direct, confident, precise, and open. There are no hidden agendas in this book. Some readers may disagree with aspects of her approach, but Tagliamonte does not dwell on disagreements. For example, perhaps the most important sociopolitical split in the field of modern sociolinguistics is the divide between those practitioners focusing on the linguistic system and those focusing on the social system. Tagliamonte concisely addresses this divide in chapter 1, arguing: “When attempting to synthesize both internal and external aspects of language, the challenge will always be to fully explore both” (4). There is more to analyze in any given situation than one researcher's focus can encompass. According to Tagliamonte, “Variationist sociolinguistics is most aptly described as the branch of linguistics which studies the foremost characteristics of language in balance with each other – linguistic structure and social structure; grammatical meaning and social meaning – those properties of language which require reference to both external (social) and internal (systematic) factors in their explanation” (5). For the remainder of the book, there is no wavering in her variationist resolve: Language should be modeled with quantitative patterns, and this analysis is made possible through the best known methodologies, the lifeblood of variationist sociolinguistics.
Possible suggestions for a second edition include exemplification from languages other than English, more exercises, and exercises that are component-driven and can be done in class. The current exercises should work well for out-of-class directed research, but they do not provide small steps and data for in-class work. As to the variation patterns examined, (t/d) deletion and (ING) are the main exemplars throughout all the chapters in this book. Although she notes that (ING) has as its strongest constraint the grammatical category in which the dependent variable is found, Tagliamonte considers it a phonological variable. Other researchers (Labov 2001:86–88; Hazen 2006) consider it a grammatical variable. Last, the glossary is oddly formatted. It has two columns (with head words on the left), but the right column with the definitions has neither punctuation nor capitalization, yielding some interesting semantic blends.
In all, the book is an important step toward the further development of sociolinguistics. It helps to demystify the process of quantitative sociolinguistic research and provides a clear path for quality work for years to come.
