Before I went to university, I didn’t think about how much of the grammar I used diverged from standard English, never mind why I might be using it. Non-linguists generally don’t spend a lot of time thinking about how or why they use language – until someone points out that their speech is ‘unusual’ or that it needs ‘correcting’, of course. But not being able to identify the relative clause in a sentence like ‘I read this book, which was fun’, doesn’t mean that you aren’t capable of using a relative clause to add information about your attitude towards reading a book. People adapt their language to subtly communicate social detail all the time. They might be able to say how they were trying to sound but, rarely, how they used language to try and sound that way.

The extent to which people are able to adapt their grammar to communicate social detail is the focus of this book. There are many, many studies which show a correlation between the use of localised grammar and low socioeconomic status, but far fewer interrogate the extent to which our use of localised grammar is constrained by our class status. Does being working class mean you can’t help but use localised grammar, or are we free to use language to develop styles and personas which transcend our place in the social hierarchy? This book seeks to answer this question.

The volume of research on social class in sociolinguistics can give the impression that we are all constantly and consciously working to present our classed selves to the world. But, just as people don’t think about the language they are using, people who aren’t social scientists or social activists don’t tend to think about social class much either. We’re all too busy just being, day-to-day. We know that there are people who are different from us and that can make us feel many things about ourselves and them, but it takes social mobility or encounters with ‘others’ to make us reflect on our precise place in the social order. I didn’t know what linguistics was until I went to university, and I didn’t know that I was working class either. I just knew I wanted to learn some more and that going to university involved stepping outside the norms of my everyday experience and usual ways of being (my habitus, Bourdieu Reference Bourdieu1990: 12–15). I had signed up for a degree in English Language and Literature, without really understanding what it involved. The English Language and Literature A level I had taken was relatively new, and the language component had taught me how to precis texts and what kinds of language to use in debates, but I don’t remember learning anything about phonetics or grammar or linguistic theory. I’d always enjoyed reading and had done well at writing. I wrote stories and poems, and people said I had a good imagination. I wanted to be a teacher. The only person I knew who had been to university was my sister, who had trained to be an occupational therapist. Both of my parents left school aged fifteen. My mum became an office clerk, but stopped working when my sister was born (she subsequently retrained as a nursery nurse when both me and my sister were at school). My dad had an apprenticeship which trained him as an electrical engineer. He’d wanted to be a teacher. They both valued education enormously and had not wanted to leave school at fifteen. Our house was full of books.

Encouraged by my parents and my school teachers, I thought I was good at literature but, on arriving at university, I quickly learned that I wasn’t. Everybody else in the classes seemed to know more than me, and it wasn’t just because they had read the texts before (although many of them had). They seemed to understand how the texts connected to history and culture, and they identified references that I missed. The other students also seemed to know how to talk about the texts and how to talk to the lecturers. I wrote copious notes and never missed a class, but the books we read didn’t make me feel anything and I only talked in class when I was picked on to answer a question. Then I found it hard to get my words out. I tried to focus on creative writing because I was confident about that, but I was told that my writing was parochial (a word I had to look up) and that the things I wanted to write about weren’t relevant. I stopped being confident about it.

The linguistics classes that formed the English Language component of my degree were different. We had done bits of work on grammar at school, but mostly in my French class. I loved the grammar classes at university – there were rules to follow and puzzles to unpick about why certain structures were grammatical and others weren’t. The lecturers were interested in – and often pleasantly surprised by – my judgements about which kind of sentences were possible in my dialect. Nobody minded that I transcribed things in my own accent when I learnt the International Phonetic Alphabet in my ‘Sounds of English’ classes. I loved Old English too because we had to learn the grammar to understand what we were reading. It was another problem to solve and one that wasn’t dependent upon what I did (or, rather, didn’t) know about the current zeitgeist (another word I’d learnt in my literature classes).

My inability to engage with the English Literature curriculum was bewildering. I thought I was clever (there are a lot of geeks in this book, and I was one of them) but I felt stupid. It felt personal. It was only in the final year of my degree, when I took a class in Sociolinguistics, that I realised that how I felt in my classes was deeply personal. As I learnt about the ways in which my language could be structured by my social class, my gender, and the location of my upbringing, I began to realise that these factors provided a framework for all of my experiences. I wasn’t engaged by the literature classes because I didn’t have the collection of social assets, including credentials, tastes, style of speech, forms of social engagement (cultural capital, Bourdieu & Passeron Reference Bourdieu and Passeron1990) to engage with them. I didn’t feel like I could talk in class because I knew my style of speaking was different. I didn’t know how to position myself as a legitimate participant in the seminar discussion.

My experience of the two components of my degree, English Language and English Literature, illustrate something about what is required for people to adapt and change (and, ultimately ‘to learn’). I didn’t enjoy my literature classes and I felt like my access to them was inhibited by some deficit in me (the lack of cultural capital). On the other hand, I enjoyed the challenge of the linguistics classes (a subject which was mostly new to everyone in my cohort) and the encouragement of my linguistics lecturers supported me to succeed, despite the fact that I felt like I didn’t belong in the university. Three factors determined my success in linguistics: access to the university environment, opportunity to engage successfully with it (facilitated in a large part by supportive lecturers) and motivation to become involved in the institution. As Eckert and Wenger (Reference Eckert and Wenger1993) have noted, all three are necessary to learn successfully (as suggested above, this definition of ‘learning’ is broader than ‘schooling’). Motivation is determined by the individual (although it can be conditioned by other factors, like confidence or sense of legitimacy), whereas access and opportunity can be externally controlled. Sometimes, individuals may want to learn, but they may not have obtained the necessary qualifications to enrol on a particular course, or they may not be able to find a course they can afford, or one which fits around other obligations like caring responsibilities. Sometimes opportunities to learn have gatekeepers who deny access to individuals who aren’t considered legitimate. People may not consciously exclude others – there is a long literature describing how implicit biases can affect the decisions people make about who is or isn’t the right fit for a job or a place on a course (see Brownstein & Saul Reference Brownstein and Saul2016 for an overview) – but if individuals are perceived to lack the qualities required of a legitimate participant, then they can be excluded. In their examination of apprenticeships, Lave and Wenger (Reference Lave and Wenger1991: 37) note that apprentices only learn successfully when they achieve a state of ‘legitimate peripheral participation’ (where the peripherality refers, not to position in relation to an abstract group centre, but to ‘an opening, a way of gaining access to sources of understanding through growing involvement’). Consequently, what matters in gaining access to a site of learning is the achievement of legitimate status.

Acquiring new language styles or adapting how we use language in our day-to-day interactions is a form of learning too. Consequently, if we want to understand why people use the language they do, we need to understand what they have been able to learn and the extent to which they are able to modify that learning. Social class is a place to start – it is central to the social order (Eckert Reference Eckert2019b: 2) and gives us access to certain linguistic variants and styles – but social class is not the place to stop. Opportunities and motivations shape what we learn too. Precisely how they do so will be explored in this book.

So this is a book about social class that is not about social class. In addition to using language to index certain social qualities, we also use language to undertake social action. Whilst we might infer something about a speaker’s social background from how they speak, a speaker’s utterance may also communicate more nuanced social detail. It might communicate subtle information about when a person is speaking, what their preferences are, what their alignment is to what they are saying, and what their feelings are about who they are saying it to. Whether or not a speaker uses a particular linguistic form may be guided by these consideration as well as, or even in spite of, their social background.

Consider Extract 1.1. It comes from a conversation between me and two girls who feature heavily in this book, Georgia and Jennifer. They are discussing Georgia’s relationships with boys.

To understand the meaning of these utterances, we need to be able to decode the content conventionally associated with the words that are used and the way they are structured. For instance, to understand the sentence ‘he were bad, though, weren’t he?’ on l. 12, we need to know that he refers to Georgia’s ex-boyfriend, that were is a localised variant of past-tense third-person singular be, and that the verb be can depict a state or condition (importantly, we don’t need to describe grammar in this way to understand it). We also need to know that bad is a word used to describe something that is not good and though is typically used to qualify something said previously (in this case, Georgia is qualifying why her ex-boyfriend was a bad mistake). It’s also necessary to know that these words are assembled to make a declarative statement, ‘he were bad, though’, and a tag question, ‘weren’t he?’, following the conventions of English. Decoding this information gives us the semantic meaning of Georgia’s utterance: it enables us to perceive the reality and truth about what she is attempting to describe. But we don’t just use language to decode propositions or truth conditions. Our understanding of Georgia’s utterance also relies upon our ability to understand meanings that are not abstractly entailed by the words and structures she uses. Some meanings are recoverable from the fact that words and structures are used in particular ways at particular moments of interaction. For instance, we might wonder why Georgia uttered a tag question at this point in the discourse. We’ll learn much more about tag questions in Chapter 7 but, for now, it is enough to note that their structure invites an interlocutor to attend to the proposition expressed in the preceding declarative statement. Unlike regular questions, they rarely constitute requests for truth-conditional information; Georgia is not asking whether Mike is bad news – she knows full well that he is. Rather than establishing whether her statement is true or false, her tag question seems to be seeking to establish that this is an opinion shared with Jennifer. It is seeking to establish common ground.

That Georgia’s utterance includes a tag question, as opposed to being a simple declarative, could be interpreted as marked. The markedness of an utterance can help us to determine what a speaker is inferring beyond what is said in a purely semantic-referential way (Horn Reference Horn, Horn and Ward2004; Acton Reference Acton2019). This is pragmatic meaning: it requires us to consider what is implied or presupposed by an utterance, beyond its referential content. Utterances may be marked for many reasons: they may require more interpretative effort, or they may indicate something about a speaker’s alignment with the content of their talk. In Extract 1.1, the syntactic configuration of the tag question (as opposed to a less marked form) may serve to emphasise Georgia’s evaluation of Mike and conduce agreement around this evaluation. If people in our speech community use a lot of tag questions in this way, it may be that we will come to associate tag questions with especially evaluative personality types.

But utterances can be marked for reasons beyond the interpretative effort required to decode them. They may be heard less frequently than alternative utterances, or they may violate dominant social norms. In constructing the tag question on l. 12 of Extract 1.1, Georgia uses a form of verbal agreement that differs from standard English: ‘he were bad, though, weren’t he?’. I’ve previously referred to this variant as nonstandard were (Moore Reference Moore2010), but to avoid unnecessarily stigmatizing variants that differ from the standard, I refer to it here as levelled were – to reflect how the use of were is levelled across all persons in the past tense. We’ll learn a lot more about it in Chapter 4, but for now, it is sufficient to observe that it is not a speech error, but a local variant that is common in Georgia’s dialect. The were has the same semantic referential meaning as the was in ‘he was bad’ and ‘wasn’t he’, in that they are alternative ways of marking past-tense third-person be. But does Georgia’s use of were have any other kind of meaning? Acton (Reference Acton, Hall-Lew, Moore and Podesva2021) has argued that utterances can gain meaning from their sociohistorical use. That is to say, we may infer something about an utterance based upon what we associate it with and our beliefs about this association. For instance, levelled were is more frequently used by people in lower-social-class groups. If a listener is aware of this association they may decode Georgia’s use of levelled were as a symbol of working-class status or, at least, as a symbol of her alignment with working-class practice. In turn, the listener may infer that Georgia has any number of social characteristics that are associated with working-class status. These might include traits like resilience, toughness, or friendliness, dependent upon the listener’s precise beliefs about working-class people. In this way, levelled were may index social meaning associated with being working class, i.e., being resilient, tough or friendly. I define ‘social meaning’ as what can be inferred about a person’s interactional position or character on the basis of how they use language in a specific interaction. This is distinct from pragmatic meaning, which I define as what is implied or presupposed by an utterance, beyond its referential content. Acton (Reference Acton, Hall-Lew, Moore and Podesva2021) argues that social meaning is a form of pragmatic meaning (see also Hall-Lew, Moore & Podesva Reference Hall-Lew, Moore and Podesva2021), however, it is important to note that whilst all kinds of social meaning entail pragmatic meaning (as defined above), pragmatic meaning does not necessarily entail all kinds of social meaning. For instance, it is possible to presuppose something about a person’s interactional position without presupposing something about their character. The range of meanings that different grammatical variants can carry, and how these meanings are generated, will be a central concern of this book. In Chapter 3, we further explore different levels of meaning and consider how social meanings develop by exploring issues of ideology and indexicality. For now, it is important to note that the term ‘grammar’ can encapsulate a wide range of different variants. If we are to understand how social meaning attaches to grammatical variants, we need to be explicit about how we define those variants. In the next section, grammatical variation, and the range of grammatical forms that can vary, is more precisely defined.

1.1 What Is Grammatical Variation?

Grammar is the way in which we structure our utterances by (i) combining meaningful units of language (morphemes) into words and (ii) putting strings of words into interpretable units (clauses). In sociolinguistics, grammatical variation is often referred to as ‘morphosyntactic’ or just ‘syntactic’ variation, but these labels can depict a broad and diverse range of linguistic units. Table 1.1 provides a simplified representation of some of the types of grammatical variation discussed by sociolinguists (Romaine Reference Romaine1984: 419; Winford Reference Winford1984: 272; Cheshire Reference Cheshire1987: 261–262).

It is not equally easy to identify sociolinguistic variation across these different grammatical types. To evaluate why one linguistic variant is used over another, sociolinguists have endeavoured to decipher the linguistic choices available for communicating a given state of affairs (Labov Reference Labov1978: 5). This is a relatively straightforward process if we can easily determine the alternative forms and compare their social value. For linguistic units like phonemes, identifying alternatives requires us to know which phonemes can denote which sounds. For instance, in British English, there are – broadly speaking – four different ways to pronounce the ‘th’ sound in a word like thing:

It is easy to see the four forms in Examples 1.1–1.4 as alternatives because, whichever of them occurs, the same word is still articulated. They are, quite literally, alternative ways of referencing the same thing (Winford Reference Winford1984: 269). Importantly, the forms in Examples 1.1–1.4 have no inherent value in themselves. They reference a sign vehicle (‘th’), but not because they have any underlying properties that intrinsically mean ‘th’; they are arbitrary pairings of sound and sign vehicle (Romaine Reference Romaine1984: 410). This arbitrariness frees phonemes up to be carriers of social meaning. Although not all alternates necessarily carry social meaning, if forms are used variably in discourse and across communities of speakers, they are potential carriers of social meaning. In the absence of any intrinsic meaning, linguistic variants like those in Examples 1.1–1.4 can take on social meanings via associations with who uses them and when.

However, unlike phonetic variants, grammatical constructions are composed of contentful morphemes and words. And the way in which these are ordered determines how constructions function. For instance, in the expression Georgia’s boyfriend, the word ‘Georgia’s’ is comprised of the proper name ‘Georgia’ and the morpheme ‘s’, which denote a person and her possession of ‘a boyfriend’, respectively. Similarly, in a sentence like He were bad, the verb were denotes ‘being’, but its form also references tense (past) and person (third person in Georgia’s usage). Furthermore, there are grammatical rules about where were can appear in the string of words that contain it.

Nonetheless, some grammatical variants involve alterations in what is produced in a clearly circumscribed linguistic ‘slot’. For instance, in the case of Definite Article Reduction (DAR), we are dealing with different ways in which the word the is articulated. Similarly, with morpholexical variants, like the use of na in Scots, there may be a simple process of deletion at work, rather than an alternation or substitution of linguistic form(s) (Smith Reference Smith2001). To some extent this makes these types of grammatical variants similar to phonetic variants in that it is a discrete and isolatable unit that varies (i.e., internal word structure or lexeme) rather than any kind of complex syntactic structure. However, the lower we get down Table 1.1, the harder it is to determine what ‘slot’ the variation falls into, and what the linguistic alternatives to a particular form might be. It also becomes more difficult to talk about the semantic equivalence of anything we might consider to be an alternative. If we think about morphosyntactic variation, this, by definition, involves some kind of structural alternation to form. For instance, with negative concord, there is repetition of negation via the use of multiple negative particles. Although, early on, Labov (Reference Labov1978: 5) argued that negative concord is ‘by definition multiple negation with the same truth value as single negation’, elsewhere, he has argued that the repetition of negative particles is intensifying (Labov Reference Labov and Schriffrin1984; Eckert & Labov Reference Andrews, Robinson and Hutchinson2017: 469) – something that could be argued to affect the state of affairs that is communicated. If negative concord communicates something different to standard negation by virtue of its grammatical structure, it becomes more difficult to think about standard negation and negative concord as functioning as simple linguistic alternatives.

As we move to the bottom of Table 1.1, the effects of grammatical configuration make it even more difficult to discern what counts as viable alternatives. The syntactic example in Table 1.1 references a study by Cheshire (Reference Cheshire2005a), who observed variation in how speakers introduce something new into their discourse. Imagine an interaction, where a garden centre is mentioned for the first time. There are many ways of introducing this discourse new entity. These could include the following:

Examples 1.5 and 1.6 are both marked ways of highlighting that the garden centre is a discourse new entity. Example 1.5 uses an existential construction (it was a garden centre that my friend went to), and Example 1.6 has a left-dislocated component (a garden centre, that was where my friend went to). Cheshire found that the type of marked strategies illustrated in Examples 1.5 and 1.6 were used in similar ways (to highlight new information) by all speakers irrespective of their social background. However, most commonly, speakers didn’t explicitly mark discourse new items – instead simply presenting them as bare noun phrases as in Example 1.7: my friend went to a garden centre. Unlike the marked discourse new strategies, the use of bare noun phrases did pattern sociolinguistically – with girls and working-class speakers more likely to use examples like those in Example 1.7 than boys and middle-class speakers. Cheshire argues that this is because boys and middle-class speakers tend to use marked discourse new strategies to highlight the discourse moments when key, factual, information is revealed, whereas girls and working-class speakers tend to be more focused on the affective content of their discourse, rather than its information structure.

Cheshire’s study shows that decoding the social meaning of syntactic variants requires us to focus on the function of expressions rather than their form. In Cheshire’s study, there is no clear ‘linguistic slot’ that bare noun phrases or marked discourse new entities fill. What makes these two strategies alternates is that they are both ways to present discourse new information. Furthermore, when Cheshire compares these forms, she finds that the sociolinguistic variation is rooted in pragmatics – it occurs where speakers are communicating different messages about their orientation to the discourse and their interlocutors. Unlike phonological variants, syntax can come with in-built dispositions to certain pragmatic functions. For instance, many of the constructions that Cheshire identifies as marking discourse new entities avoid placing the grammatical subject of the main clause in an initial position in the utterance. This violates the general principle of given-before-new information (Cheshire Reference Cheshire2005a: 486). Consequently, the syntactic configuration itself (placing the new information at the beginning of the utterance) facilitates the articulation of certain pragmatic inferences (in this case, Cheshire’s analysis suggests that the speaker is focused on communicating new, key, factual, referential information rather than on building interactional rapport).

The ability for syntactic constructions to encode pragmatic meaning by virtue of their grammatical configuration makes them quite different from phonological variants. It is a difference that has long been recognised (Lavandera Reference Lavandera1978; Dines Reference Dines1980; Romaine Reference Romaine1984; Cheshire Reference Cheshire1987, Reference Cheshire1999, Reference Cheshire2005a; Cameron & Schwenter Reference Cameron, Schwenter, Bayley, Cameron and Lucas2013), but the focus on phonetics and phonology in sociolinguistics has hindered our understanding of the relationship between grammatical variation and social meaning. In the next section, I outline the way in which this book attempts to explore the relationship between grammatical variation and social meaning.

1.2 How Will This Book Examine the Relationship between Grammar and Social Meaning?

So far, the vast majority of work on grammatical variation in sociolinguistics has focused on morpholexical and morphosyntactic variation. This work has been important in demonstrating how these types of grammatical variable correlate with macrosocial categories such as social class, gender, age and ethnicity (see Tagliamonte Reference Tagliamonte2012  for an overview). However, the extent to which grammatical variation can encode these and other types of social meaning remains unclear. For this reason, the social meanings associated with different types of grammatical variable will be explored in Chapters 4–7, which form the analysis chapters in this volume. In order to increase our understanding of a wider range of grammatical variables, Chapters 4 and 5 will apply new methods to the study of traditional morpholexical and morphosyntactic variables, whereas Chapters 6 and 7 will focus on less frequently studied and more ‘purely’ syntactic phenomenon.

The research that has been undertaken on morpholexical and morphosyntactic variation has suggested that these variables are less subject to social evaluation than phonological variables (Labov Reference Labov1993; Labov Reference Labov2001: 28; Levon & Buchstaller Reference Levon and Buchstaller2015) and that the types of social meaning they index are more restricted than those typically found for phonological alternatives (Eckert Reference Eckert2019a: 758–759). In particular, it has been suggested that, because these variables are the focus of educational attention in a way that phonological variation is not, they develop social meanings directly associated with social class and institutional orientation (Eckert Reference Eckert2019a: 758–762). However, few studies of syntax have given serious consideration to the local contexts in which language operates. This means that our understanding of the social meaning of syntax is dominated by correlations between linguistic forms and demographic categories (to the extent that variants are depicted as being ‘working-class’, for instance). More significantly, the social meanings of categories have been inferred in line with dominant ideologies about class. That is to say, research has highlighted outsider perspectives on the meaning of working-class practice. This is despite the fact that the notion of covert prestige (coined by Labov as far back as 1972) implies that speakers orient to localised linguistic variants on the basis of social meanings which differ from overt prestige norms. Whilst it may well be the case that the social meanings of grammatical items are more likely to be limited to those associated with social class and institutional orientation, we simply do not have sufficient data to evaluate the extent to which this is true.

Part of the focus on category-linked social meanings stems from research suggesting that whilst we can continue to acquire lexical and morphologically conditioned changes across our lifespan, syntactic change is more difficult to acquire as we age (Kerswill Reference Kerswill1996). If there are psycholinguistic conditions on the acquisition of syntax, then what we learn in childhood may constitute a default style of speech: ‘the form first learned, most perfectly acquired, which we use automatically and unthinkingly’ (Labov Reference Labov2013: 3). This style will reflect the circumstances (place/time) of our acquisition. Although there is evidence for the ‘cognitive primacy’ of what we acquire in childhood (Sharma Reference Sharma2018: 26), there also remains insufficient data about the extent to which we can adapt or change our use of grammar, irrespective of the social meanings it encodes.

By focusing on the range of social meanings that can be articulated by a linguistic variant of the type most frequently studied in sociolinguistics (what I have earlier referred to as ‘levelled were’), Chapter 4 will consider the extent to which speakers use morpholexical/morphosyntactic variation in stylistically sophisticated ways. More specifically it will suggest that speakers can adapt their use of these variables providing that (i) they have access to them (i.e., they are available in their input) and (ii) that they are motivated to use them by virtue of their utility as a social symbol.

The variant studied in Chapter 4, levelled were, is situated on the boundary between morpholexis and morphosyntax. What other factors do we need to consider when we examine a variable that is more firmly morphosyntactic in character? Recall that the social meanings of grammatical variants which are structurally embedded are likely to be tied to their semantics. This makes it vital that any exploration of social meaning includes consideration of the pragmatic meanings that a form’s semantics and structure permits it to portray. Earlier, we noted that the repetition of negation via the use of additional negative particles in negative concord can be intensifying. But we also know that negative concord is ‘arguably the most common stigmatized variable in the English language’ (Eckert Reference Eckert2000: 216). Does the pragmatics of the repetition of negative particles have any role to play in the social meaning of negative concord, or has its social stigma bleached out this pragmatic function? We will explore the balance between social association and pragmatics in Chapter 5 via a study of negative concord.

Both of the variables studied in Chapters 4 and 5 match the profile of the highly stigmatized grammatical variant. However, Chapters 6 and 7 will move beyond this prototypical variant type to explore more purely syntactic variation that does not easily conform to notions of stigma and nonstandardness. The pragmatic utility of syntactic forms – like those used to introduce discourse new entities – may be common to all speakers, even if they are used at varying frequencies. For these linguistic forms, variation is not necessarily driven by the variety one acquires in childhood and how this variety is placed along the axis of prestige and stigma, but by pragmatic utility and communicative need. That is to say, differences in social meaning that are rooted in the pragmatics of a construction may engender differences in how frequently certain groups use a linguistic form. Put more simply, whether or not an individual uses a particular syntactic item might depend upon ‘what speakers choose to talk about’ (Cheshire Reference Cheshire, Cornips and Corrigan2005b: 99) and, perhaps more importantly, how they choose to talk about it. The linguistic form studied in Chapter 6 is right dislocation. This structure consists of a tag at the right periphery of a clause, such as She’s fun, my mum. The tag (‘my mum’) is co-referential with the preceding subject or object pronoun (the subject pronoun ‘she’ in the example). Right dislocation is very common in informal spoken language – even people who do not consider themselves to have a local dialect will use it from time to time. However, its precise grammatical composition can vary. Consequently, Chapter 6 will consider how the precise grammatical environment in which these forms occur may intensify or attenuate their pragmatic effect, such that – despite the universality of their occurrence – different social groups interact with this variable in socially meaningful and socially nuanced ways.

Chapters 4–6 focus on occurrences of discrete grammatical items. However, syntactic variants are contentful: they don’t just differ by their syntactic structure, they also differ by their lexical content and, in speech, by their phonetic content. Do these different levels of linguistic architecture work individually or synergistically to create social meaning? The focus on phonetics and phonology in sociolinguistics has meant that we are a long way from understanding how different elements of language might work in related ways to affect social meaning. As Eckert and Labov (Reference Andrews, Robinson and Hutchinson2017: 485) note, ‘the realization of a phonological variable is a short (and frequent) event in a syntactic series of events’. Consequently, it is likely that the nature of ‘a syntactic event’ will constrain the precise social meaning articulated by any of its constituent parts. For this reason, we will think about the relationship between syntax and other types of linguistic structure in Chapter 7. By examining tag question constructions (like the example, He were bad, though, weren’t he?, in Extract 1.1), Chapter 7 will show how grammatical environment can work synergistically with other levels of linguistic architecture – including phonetics – to create social meaning. This chapter will attempt to propose recommendations for how to better integrate the study of syntactic variation into a wider understanding of the social meaning of language more generally. In this chapter, we will also think about whether the universality of syntactic variables like tag questions (i.e., variables that everyone uses to some extent to express interactional positioning), means that they do not acquire the types of social meanings found for other linguistic variables.

The chapters described so far, (Chapters 4–7), form the data analysis chapters in this book. But, of course, sound empirical analysis relies upon good quality data. Chapter 2 will present the ethnography on which this book is based. The ethnography took place in a school, Midlan High, and this institution will be contextualised socially and geographically. The social groups within the school will also be described. This chapter will reflect upon the fieldwork process and the intricacies of doing fieldwork within an educational context. It will also discuss the types of linguistic data collected during the ethnography, and reflect upon their relationship to other types of data typically collected in variationist research, and ways of categorising speakers in studies of linguistic variation.

Sound empirical analysis also draws upon (and refines) theories about a particular set of concepts. Understanding the social meaning of linguistic variation requires that we study language as it relates to social practice and forms of social engagement. Chapter 3 will more deeply interrogate how we study social meaning and the processes involved in meaning making. What concepts do we need to know to understand how social meaning develops and what techniques are required to understand how these concepts operate? The ways in which variationist sociolinguists have examined the social meaning of linguistic variation, with a specific focus on how these have been applied to the study of grammatical variation, will be considered in Chapter 3. Given that social meaning may interact with pragmatics, this chapter will also highlight the need to combine research on the pragmatics of spoken language with variationist work on the social embedding and social distribution of linguistic variables.

The research described in Chapters 4–7 will demonstrate how grammatical variation contributes to the social meanings and styles that the young people of Midlan High articulate in everyday interaction. Drawing upon the range of grammatical variables studied in the preceding chapters, Chapter 8 will argue that grammatical variables are subject to social evaluation just as phonological variables are, but that the type of grammatical variable may affect the types of social meaning associated with the specific variable studied. Consequently, the final chapter of this book will reflect upon what it means to view grammar as fluid, flexible and as a social resource. Educational linguists have pointed out the benefits of teaching grammar explicitly to school pupils, but have noted that this teaching is most effective when grammar is presented as a semiotic resource for meaning making (Myhill et al. Reference Myhill, Jones, Lines and Watson2012; Myhill Reference Myhill2018). However, it has been argued that current educational policy not only shapes how teachers conceptualise grammar (as rigid and inflexible), it also encourages negative and potentially damaging responses to any deviations from the standard variety of a language (Cushing Reference Cushing2019a, Reference Cushing2019b, Reference Cushing2020; Hudson Reference Hudson, Giovanelli and Clayton2016). This is despite there being no significant evidence that speaking in a localised dialect affects literacy (Snell & Andrews Reference Snell and Andrews2017), or that children’s literacy improves by decontextualised grammar teaching (Elley Reference Elley and Asher1994; Hudson Reference Hudson2001; Wyse Reference Wyse2001; Andrews et al. Reference Andrews, Beverton, Locke, Low, Robinson, Togerson and Zhu2004). This chapter will reflect upon the wider implications of foregrounding social meaning in our understanding of grammar. What are the potential benefits of viewing style shifting as a form of linguistic skill that children already have at their disposal? In this way, the final chapter of this volume will seek to apply the theoretical discussion of previous chapters in order to evaluate (and generate) the impact of the research. I began this chapter by talking about social class, and this final chapter will demonstrate how misconceptions about grammar and its social meaning help to perpetuate class-based social inequalities.

The following pages contain discussions of data I collected, but my analysis is rooted in the words of the lively, fascinating, vibrant young people of Midlan High. As I hope to make apparent in what follows, their discourse doesn’t just facilitate our understanding of the social meaning of language; the young people of Midlan High use language to actively animate their social world. In order to redefine models of grammatical variation on the one hand, and better inform debates about how children use language to make social meaning on the other, we need to properly appreciate the social and pragmatic functions of the grammatical features embedded in ordinary people’s everyday use of language. With this in mind, the next chapter introduces the young people of Midlan High.