The very idea of reviewing these two books together is good thinking. In terms of content, the books are alike in that both have state-of-the-art ambitions. At the same time, the two books are very different in terms of genre—in how they go about stating the art of sociolinguistics. Thus reviewing them together allows us not only to compare two ambitious presentations of sociolinguistics, but also to compare the monograph and the collection of articles as two major ways of presenting a discipline. In this review, I focus on explanations of language variation and change as my red thread, spending more time on Bayley & Lucas' newly edited collection of articles, continuing with a discussion about Chambers' time-honored monograph, and ending with some concluding remarks.

Apart from a short introduction by the editors and a lovely personal afterword by Roger W. Shuy about Walt Wolfram, to whom the book is dedicated (four pages), Sociolinguistic variation consists of seventeen chapters grouped into three parts: “Theories” (eight chapters, 157 pages), “Methods” (three chapters, 69 pages), and “Applications” (six chapters, 112 pages). The references list at the end (thirty-nine pages with close to 1,000 entries), as well as the index list (twelve pages), immediately reveal the book's very North American orientation.

To me, it must remain the editors' secret why Ch. 4 is not the first. Kirk Hazen, in “The study of variation in historical perspective” (70–89), does a good job of establishing a framework that embraces the various aspects of the study of variation represented in the book. Newcomers should start with Hazen's presentation and comments for a better understanding of the discussions and developments both within the variationist field itself and in relation to the larger field of linguistics. For my red thread, I offer two quotes: “[T]he nature of language requires variationist explanation” (89); “The extraordinary trait of variationists-cum-linguists is the inclusion of social factors as well as linguistic factors in the explanations of language variation” (71).

Returning to Ch. 1, “Variation and phonological theory” (5–23), we find that Gregory R. Guy explicitly leaves the treatment of social regularities in linguistic variation to other contributors of the volume (6). The chapter gives a clear presentation of “the reciprocal, mutually illuminating relation between phonological variation and phonological theory” (6), the central claim being that “[a]n adequate theory of phonology will offer explanations of the broadest range of sound patterns, including non-categorical, probabilistic patterns, and the study of variation will inform the construction of such a theory” (23). Thus, the concern is with linguistic regularities only, and explanations are given in terms of general, possibly universal, principles and processes of contextual conditioning. Optimality Theory models are accorded particular attention.

In “Syntactic variation” (24–44), Lisa Green is preoccupied in the same vein with new endeavors to incorporate variation into syntactic theory, as represented by Optimality Theory and the Minimalist Program, and with how “this approach may be useful in answering questions and making predictions about the possible ways dialects can vary and the limitations for options in the grammar” (27). Questions and negations in Mainstream American English (MAE) and African American English (AAE) are discussed in this perspective. Furthermore, Green argues that the advocated approach might help to distinguish developmental variation in language acquisition from that which is socially motivated. The development of the copula and auxiliary be in AAE is discussed in the perspective that there are “syntactic … and semantic constraints on the production of be that might be general developmental phenomena” (40).

In Ralph W. Fasold & Dennis R. Preston's chapter, the explanatory potentials of a theory that incorporates variation is part of a wider, and most crucial, discussion of the cognitive status of variation. The authors whip out old Occam's razor and argue for “[T]he psychological unity of inherent variability” (45–69). This is accomplished by way of constructing a modular model (visualized as boxes connected by arrows), which is modified as the authors take the reader through three levels of variationist sociolinguistics. Basically, there are four modules: an information (or contextualization) devise, a grammar, a sociocultural selection devise, and an output. Level 1 studies link sociocultural factors with linguistic forms. The probabilistic character of that linking is claimed to be a psychologically plausible model for what is going on when speakers use more or less (in percentages) of one form of a variable: “They are evidencing the influence of a set of probabilistic weights that come to bear on each occurrence, a cognitively plausible (rather simple) operation” (52). Level 2 studies “search for influencing factors among (not outside) the components of a grammar” (55). The authors state that they “believe linguistic (not sociocultural) motives for variation are strongest. … [L]inguistic influences are nearly always probabilistically heavier than sociocultural influences” (55). Level 3 studies focus on how language changes by virtue of variable use, and they seek to relate patterns of change to both sociocultural and linguistic forces. The authors discuss changes in linguistic competence, and hence in their model, in terms of postvernacular learning, the result of which is assumed to have a less deeply embedded status than the vernacular; such “grammatical weakness” (63) is introduced as a grey area within the grammar module and is “stated rather than shown [to be the areas of] psycholinguistic processes such as retrieval, accessability, memory, and the like” (68).

Since the ambition of such a unitary, modular model is to determine the output, and in that sense to explain variation in language use, its realism as a picture of “variable linguistic competence” seems problematic to me. This goes not only for the claims above about the nature and functioning of the “internal” forces behind language variation, but also for the “external” delimitation of the modeled competence from other kinds of competence: The authors begin by excluding a component of “communicative competence” (in the sense of Dell Hymes) from the model because “such abilities … interact with rules of linguistic competence rather than constitute a part of them” (45, n. 2). They conclude by stating that “the probabilistic model we have drawn here does not fully determine the output of grammars since the ‘intention’ of a speaker may interact with his or her sociocultural identity. That is, one may choose to ‘perform’ (or perform to a greater or lesser extent) an available sociocultural identity, and such choices must play a role in the activation of the selection determined by sociocultural selection. That provisio, however, makes this program more difficult, not inconceivable” (68–69). My interest here is not to discuss either the inclusion or exclusion of the communicative and performing competences in question, but rather to point out that the authors, quite independently of the demarcation issue as such, seem to operate with a distinction between usage that is determined and usage that is chosen. On the one hand, the sociocultural selection device determines how language is used by speakers in accordance with their sociocultural identities. On the other hand, speakers choose how to use language in order to “perform” an available sociocultural identity. This distinction is akin to the distinction in the grammar module between vernacular (determined) use and postvernacular (chosen) use. I certainly agree that a cognitively realistic model of variable linguistic competence will have to include a conceptualization of the relationship between determination and choice. My reservation with the proposed model concerns the general aim to “fully determine the output of grammars” and the derived distinction between a core of (probabilistically) determined use and an added margin of (freely) chosen use.

This may be the place to mention that Hazen's overview of “the study of variation in historical perspective” does not include what Penelope Eckert has termed “third wave variation studies” (see Eckert's homepage at http://www.stanford.edu/~eckert/thirdwave.html), that is, studies that analyze choice of linguistic variants as negotiations of social meanings and identities and conceive of speakers as constructive users and creators of “styles.” The reason is probably due to Hazen's (tacit) exclusion of studies in the Hymes tradition; the anthropological approach has been accorded no more than one sentence in a short appendix-like section that lists a number of “variable methodologies” (80, where Eckert's work is mentioned as bridge-building “between quantitative variationist study and the examination of social meaning,” with a reference also to work on style by Nikolas Coupland). Against this backdrop, the reader will be surprised (and the newcomer confused) by finding that Ch. 5, which follows immediately after Hazen's overview, explicitly and repeatedly places itself in the Hymes tradition. “Style in dialogue: Bakhtin and sociolinguistic theory” (90–109) is written, interestingly enough, by the only non-North-American contributor to the volume, Alan Bell. His New Zealand voice blends with the Russian voice of Bakhtin (“In truth, I doubt whether I could pick apart my words from Bakhtin's in some sections”, p. 94) in double-voiced explications of language variation, which do not so much reweigh the explanatory loads of social vs. linguistic forces so much as they take us beyond: “Variety, therefore, is endemic to language, native to it, innate. It is also multidimensional. The kind of variety that Bakhtin depicts above does not coincide with any obvious parameter such as the demographics of a speaker in a Western urban context. Rather he offers many cross-cutting social and social-psychological phenomena—groups, purposes, ideologies, registers” (100). This approach of course represents a radical critique of the idea that speakers have a basic, genuine, authentic way of speaking—the vernacular—which sociolinguists should strive to study. Bell stresses the importance of the hearer, the audience, to the sociolinguistic understanding of the nature of language (as opposed to theoretical linguistics' focus on the speaker). Language variation exists because human beings, speakers and hearers, interact. “A dialogical theory of language is the foundation of a rounded sociolinguistic theory, placing stylistic variety at the heart of our enquiry, and asserting with Bakhtin the centrality of language to our humanness” (109).

A statement to the same effect about the nature of variation is found in Michael Montgomery's introduction to Ch. 6, “Variation and historical linguistics” (110–32): “Because it is socially situated and motivated, variation is intrinsic to natural language and is always potentially unstable” (110). Naturally enough, however, the chapter's general interest is not in theorizing (socio)linguistic variation, but in the development of better methods for the study of written texts from the past—“socio-historical methodologies” (110), which combine the methods of traditional philology, the concepts and quantitative methods of sociolinguistics, and the scholarship of social historians. Montgomery stresses the value of letters written by semiliterate individuals as entries to the nonstandard language of earlier periods, and also stresses that “audience factors come forcibly into play when analyzing an individual's letters” (119). This is claimed to be “in contrast to recorded interviews”—a claim that certainly is in contrast to the position advocated by Bell.

Chapters by the editors conclude the Part 1. In Ch. 7, “Second language acquisition: A variationist perspective” (133–44), Robert Bayley's focus is on variationist linguistics' contributions to SLA research, but some of the presented SLA research might well be considered a contribution to Fasold & Preston's psycholinguistic model of inherent variability. Their model copes with interlanguage variation by adding a new grammar (and output module) for each language learned after the first language. As the result of postvernacular learning, these grammars are assumed to differ in terms of cognitive development and status from the first language—vernacular grammar. Bayley presents empirical research designed to shed light on whether SLA proceeds by repeated restructuring, or gradually along a multidimensional continuum. “As it turns out, the literature on interlanguage variation offers examples in support of both models” (140). In Ch. 8, “Variation and modality” (145–61), Ceil Lucas reviews the history of the study of sociolinguistic variation in American Sign Language, compares variability in spoken languages and sign languages, examines internal constraints on variable units in the two modalities, and discusses modality as an explaining factor behind observed differences.

Part 2 opens with a chapter on “Sociolinguistic fieldwork” (165–89) by Natalie Schilling-Estes. This chapter should be obligatory reading on the subject in any introductory course in sociolinguistics. I would have said the same about Sali A. Tagliamonte's chapter on “Quantitative analysis” were it not for the numerous errors of false replacements and interchanges of numbers and terms e.g. the numbers in Table 10.1, figures for past and present, range in Table 10.12, “apparent time results” instead of “real time results” (208), “first person” and “others” instead of “spoken dialogue” and “internal thought” (209).

By placing Erik R. Thomas' chapter on “Sociophonetics” (215–33) at the end of Part 2, the editors have chosen to ignore his introductory claim that “it is a mistake … to view sociophonetics as essentially methodological. In fact, sociophonetics is concerned with addressing some quite basic theoretical issues in linguistics, foremost of which is how language is processed cognitively” (215f.). This is substantiated in an excellent overview of sociophonetics studies that ends with a discussion of exemplar theory as a new and different approach to phonology: “One key difference is that the lexical and indexical aspects of language are seen as interconnected, not as separate modules” (231). Thus Thomas' chapter appears to be a direct answer to Fasold & Preston, who end their article (Ch. 3) by inviting the reader to “join with variationists whose modularity is … beyond suspicion” (69).

Part 3, “Applications”, includes very worthwhile readings in social-issues-interested sociolinguistics: Carolyn Temple Adger & Donna Christian, “Sociolinguistic variation and education” (237–53); A. Fay Vaughn-Cooke, “Lessons learned from the Ebonics controversy: Implications for language assessment” (254–75); Angela E. Rickford & John R. Rickford, “Variation, versatility, and Contrastive Analysis in the classroom” (276–96); Ida J. Stockman, “Social-political influences on research practices: Examining language acquisition by African American children” (297–317); Ronald R. Butters, “Sociolinguistic variation and the law” (318–37); and John Baugh, “Attitudes toward variation and ear-witness testimony” (338–48).

In their introduction, Bayley & Lucas point to Chambers' Sociolinguistic theory as the only textbook in sociolinguistics that is primarily devoted to variation (1)—a claim that seems correct to me only if social-issues-oriented approaches (like the ones presented in Part 3 of their work) are excluded from the category “devoted to variation.” What is exceptional about Chambers' Sociolinguistic theory is its well-defined engagement with language-oriented approaches, its thoroughgoing focus on understanding variation and change as fundamental aspects of human language. The book's subtitle and first sentence signal that it is about “linguistic variation and its social significance”, and this “core area” of sociolinguistics (10) is conceptualized as “Correlations” (the title of Ch. 1): Each of the three chapters that make up the main body of the book “concentrates on one of the three major social correlates of linguistic variation—class, sex, and age—along with the web of related social factors that each of them entails” (10). I have often returned to these chapters for their high quality as overviews and discussions, and I certainly agree with Peter Trudgill when he states in his preface as series' editor that Chambers' “discussion in the final chapter of the origins and functions of linguistic variation is one of the most challenging and exciting pieces of work ever to emerge from the field of sociolinguistics” (xvii).

The first edition came out in 1995. Subsequent changes have been small—in particular from the second 2003 edition to the revised (and for some unexplained reason, not third) 2009 edition—and hardly significant in terms of how the core area of sociolinguistics is conceptualized. Issues concerning the external demarcation and internal constitution of the human language faculty are central to the first chapter of the book. Like Fasold & Preston, Chambers argues for a modular model. He adopts (explicitly elaborated only from the second edition) Noam Chomsky's tripartite model with a grammar, a conceptual system, and a pragmatic competence, putting particular focus on the latter by way of introducing Hymes' notion of communicative competence supplemented with research evidence for “the independence (or modularity) of communicative competence” (31) as one of the “organs of the language faculty” (30).

There is not necessarily a contradiction, I guess, between Chambers' model of language faculty and Fasold & Preston's model of linguistic competence, which, respectively, includes and excludes communicative competence. However, having excluded, as noted above, “such abilities as those Hymes speaks of” from their model of linguistic competence, Fasold & Preston include a sociocultural selection device that connects linguistic facts with social facts, stressing at the same time what seems to be their main message: “If one chooses to call that connection a psycholinguistic one, that is taking a broad view of psycholinguistics, and we are not opposed to it, but we want to be clear, however, about the separateness (or modularity) of the devices that are at stake here” (47f.). Even though I find that “social evaluation” is strangely marginal (“strangely” not least because Preston is half responsible) to their whole discussion of “the psycholinguistic unity of inherent variability,” I take it that the psychological processes that connect linguistic and social facts do include social evaluation.

In contrast, it is less clear to me whether and how Chambers' model incorporates social evaluation as an integral, and thus essential, element of correlating. The model seems to lack a psycholinguistic device for the role of social evaluation in linguistic variation—unless the evaluation aspect of correlating linguistic and social facts should be seen as a part of the communicative competence module. None of the discussion points in that direction, however, as all of the examples given in favor of the independence of communicative competence concern conversation skills. As far as I can see, it is not only that Chambers' model avoids addressing the place of social evaluation in variation, but more generally avoids addressing the psychological reality of correlations. Correlating is highlighted only as “the primary empirical task of sociolinguistics” (18), not as an aspect of the language faculty. In theorizing language variation, Chambers appears far less interested in the role of correlating as a psychological process involving social comparisons and evaluations than in how correlations of linguistic and social facts emerge from “deeper,” possibly universal aspects of the language faculty. His focus in terms of where the future developments in our understanding of the nature of language variation can be expected to come from is clearly with studies of a language-typology kind (search for “primitive” or “natural” processes of the language faculty; see §5.7, p. 258ff.; see also Chambers Reference Chambers2005) rather than language-specific studies of evaluative reactions to variants and varieties.

My point here is not to signal dissatisfaction with Chambers' focus—on the contrary, I find it exiting—nor is it to suggest that Chambers ignores the role of social values and evaluations—again on the contrary, I have more than once borrowed authoritative support for my own claims by quoting the first sentence of his last section: “The underlying cause of sociolinguistic differences, largely beneath consciousness, is the human instinct to establish and maintain social identity” (266). My point is rather that the unfolding of this instinct, its working to establish and maintain social identity, is taken as an unproblematic “given” instead of being addressed theoretically and empirically as an object of study. This does not prevent Chambers from making strong claims about the decisive role of social-identity marking and its nature in terms of consciousness, as in the quote above, and in this one: “social stratification is often crystal clear because of systematic subconscious symbolization” (53). I think that a number of such previous “givens” could have disserved a “revised” treatment.

For instance, in view of the general lack of explication and discussion of what to understand by notions such as “subconscious” and “beneath consciousness,” I actually find it more reasonable to claim that consciousness, in some sense, has to be involved for social meaning-making and identity construction to take place. In the same vein, I am not sure how to interpret “completely unmonitored” in a statement like “Of course the style most desired for sociolinguistic purposes is completely unmonitored” (19), but I am quite sure that fewer sociolinguists today than in 1995 would subscribe to the statement as obviously right, as presupposed by the initial “of course.” The issue of determinism vs. free choice in the midst of social values and pressures is touched upon in a comparison of social class and networks as norm-enforcement mechanisms, including a short mentioning of Eckert's work—but the 2009 text is not expanded here to embrace the “community of practice” and the “ethnographic” approach to studying social meaning-making advocated in studies that take this social construct as their base (§2.6.3, p. 76ff.). Let me also mention the treatment of the notion of “covert prestige,” which in all three editions is proclaimed “now established as a functional principle in linguistic variability” (142). However, as William Labov for quite some time now has treated the empirical evidence for “covert” social values as highly questionable, with radical theoretical consequences in that “mechanistic” linguistic processes have replaced social pressures as the main driving force behind linguistic variation and change (Kristiansen 2010), I would have expected to see some kind of repercussion of this development in Chambers' treatment.

Apart from sharing a predominant North American focus, the two books under review have surprisingly little in common and present very different pictures of variationist sociolinguistics. The core area of correlations, which is the strict focus in Chambers' book, cannot be said to be a central concern in any of the contributions to the Bayley & Lucas collection. The reason for this is not, however, that the collection is intended as a coherent alternative based on a different conception of variation in language and variationism in (socio)linguistics. As mentioned, Bayley & Lucas point to Chambers' Sociolinguistic theory as the only existing variation-devoted textbook and seem to offer their own collection of “overviews of the major areas of concern for students of linguistic variation” (1) as a kind of supplement to Chambers–a supplement that takes his conception of the “core” as a “given” to be built on rather than discussed. Consequently, the two books are very similar in their limited and rather haphazard treatment of theoretical positions that challenge the traditional conception of the core in terms of (the psycholinguistic reality of) modules, correlations, and vernacular speech.

Noticably absent from both books is any discussion of social evaluation as an independent object of empirical study in variationist sociolinguistics, which I believe is symptomatic of the general situation in the field. Chambers, having established that “[c]orrelating … is the primary empirical task of sociolinguistics,” goes on to say that “[b]eyond this primary empirical task lies the problem of interpretation. … What is the social significance of linguistic variation?” (18). These formulations do not exclude that the problem of interpretation is to be treated as an empirical task, albeit of a secondary kind. However, Chambers' discussion of empirical research into social evaluation, as an integrated part of variationist studies, consists of little more than a three-page reference to findings by Labov in New York and Peter Trudgill in Norwich in the first halves of the 1960s and 1970s, respectively (see §5.4.1, p. 235–38). Likewise, Bayley & Lucas do not treat social evaluation as a language-variation-and-change issue to be discussed in their section on “Theories”; although adequately included in Schilling-Estes' program for sociolinguistic fieldwork, the role of ideologies and attitudes is treated only as a language-and-social-problems issue—and is, therefore, very much in accordance with common textbook practice, relegated to their “Applications” section.

In spite of Labov's original and persistent insistence to the contrary, both in theory and practice (Labov Reference Labov1966, Reference Labov1972, Reference Labov, Baugh and Sherzer1984), variationist researchers have often abstained from engaging empirically with “the evaluation problem,” passifying themselves with interpreting social significance on the basis of pattern of use alone. At the same time, most variationists have always been interested in the internal mechanisms of language variation and change. It seems to me that the latter interest has prevailed in recent years as traditional variationism has moved away from explanations based on social values. This is clearly the case in Labov's own work (Kristiansen Reference Kristiansen, Wodak, Johnstone and Kerswill2010), and I think the two books under review mirror the general situation in the field quite well: the study of social evaluation as a relevant explanatory force is left aside in favor of preoccupation with deeper, language-faculty dependent, principles. In view of this tendency, we have seen it as a challenge to maintain and develop empirical research in social evaluation as a crucial part of our efforts to describe and explain the development of the Danish language and speech community (Kristiansen Reference Kristiansen2009).

Being both very similar (sharing basic conceptions of what variationist sociolinguistics is about) and very different (picturing the core and major areas, respectively), we can indeed argue that the two books make perfect complementary readings: the new edited collection makes a nice supplement to the time-honored monograph. Nevertheless, it is hardly risky to foretell that the Bayley & Lucas volume will never come close to the must-have status that Chambers' monograph has obtained. Whether it is well-deserved in this case or not—to some extent I think it is—the edited collection will suffer from the genre's reputation of lacking internal coherence. A particularly bad reputation attaches to what at my end of the world is called a festschrift, which perhaps explains why the festschrift aspect of the Bayley & Lucas volume, from my continental European viewpoint, appears strangely downplayed. Actually I remain ignorant of why Wolfram was celebrated in 2007.