2.1 Introduction
In spite of the prodigious amount of effort invested by grammarians in codifying English and innumerable pages describing and illustrating standard usage in grammar books and dictionaries, it is perhaps remarkable that a linguistically principled definition of standard English continues to evade professional scholars. Milroy (Reference Milroy, Macaulay and Fought2004:162) argues that ‘the concept of the standard is surprisingly underspecified and undertheorised’, while Coupland (Reference Coupland2000:623) laments the inability of linguists to arrive at a ‘persuasive set of principles and perspectives’ in relation to what standard English is. Though there is broader consensus on what constitutes standard written English as the result of its extensive codification, the concept of a spoken standard remains far more nebulous (Cheshire Reference Cheshire, Bex and Watts1999a:129; Cameron Reference Cameron2012:100), raising doubts as to whether a unitary standard spoken variety even exists in any objective, well-defined sense (e.g. Poplack Reference Poplack, Valdman and Hartford1982:2; see also Snell, this volume, for further discussion).
Much confusion about what constitutes standard spoken English derives from the failure of the prescriptive grammatical tradition to draw systematic distinctions between writing and speech (Crowley Reference Crowley, Bex and Watts1999:129; Milroy and Milroy Reference Milroy and Milroy2012:22). This confusion is further exacerbated by the widespread belief that written English is the most ‘legitimate’ or ‘authentic’ representation of the language. Invidious comparisons with written texts, where standardisation is most easily achieved, are at the root of public opinion that uniformity in language represents the normal state of affairs whereas variation, inherent to all spoken language, is deviant and undesirable (Cameron Reference Cameron2012:39).
Pervasive indeterminacy surrounding what counts (or does not count) as standard spoken English is further compounded by the multiple frameworks used to analyse it, not to mention the obfuscatory ideologies associated with the standard language itself. These problems have conspired in blurring the boundaries between standard and non-standard English, giving rise to numerous ‘grey areas’ and definitional ‘fuzziness’ (Milroy and Milroy Reference Milroy and Milroy2012:22).
The following sections elaborate on why the notions of standard and non-standard English pose many terminological and conceptual problems and explore alternative perspectives on standard English and the process of standardisation. Special consideration is given to everyday spoken English and its core properties of variation and change, both of which are fundamentally at odds with the pursuit of uniformity and the suppression of variability at the heart of the standardisation process (Milroy and Milroy Reference Milroy and Milroy2012:6). Because this process cannot be properly understood independently of the diachronic context in which it arose, I briefly trace the historical roots and evolution of standardisation in the English language, paying particular attention to prescriptivism and its relationship to actual usage.
Recognising that the (variable) structure of spoken English cannot be easily characterised by appealing to conventional linguistic descriptors (Cheshire Reference Cheshire, Bex and Watts1999a:129), I consider why progress in achieving a clearer understanding of the nature of spoken English has been hindered by frameworks that have downplayed or ignored the inherent variability which pervades speech. This discussion segues into a review of structural aspects of spoken English which have benefited from careful quantitative analyses, elucidating robust usage patterns that have been insufficiently acknowledged in the body of work addressing the standard language. I then turn to the concept of ‘Received Pronunciation’, assessing its associations with ‘educated’ and upper-class speech, and its status as a socially symbolic, non-localisable accent. Finally, I offer general conclusions and directions for future research.
2.2 Standard and Non-standard English: Definitional Issues
Much of the difficulty in defining standard and non-standard English resides in the terms ‘standard’ and ‘non-standard’ themselves (see Lippi-Green Reference Lippi-Green2012:62), or, more precisely, the widespread disconnect between professional linguists’ understanding of those terms and their popular, value-laden interpretations as manifested in public discourse about language.
From a sociolinguistic perspective, a standard language is typically used in a wide range of social functions (maximal elaboration of function) and is characterised by structural uniformity (minimal variation in form), achieved via the suppression of optional variability in language (Milroy and Milroy Reference Milroy and Milroy2012:6). But, as Cheshire and Milroy (Reference Cheshire, Milroy, Milroy and Milroy1993:15) point out, the term ‘standard’ can also denote ‘something to be aspired to’, whereas its converse, ‘non-standard’, can have socially undesirable connotations corresponding to inferiority and functional inadequacy. Indeed, negative ideologies associated with non-standard English usage abound and continue to fuel a long-standing ‘complaint tradition’ (Milroy and Milroy Reference Milroy and Milroy2012), couched in tenacious beliefs that (perceived) deviations from standard English are leading to linguistic corruption and degeneration. So entrenched are popular beliefs in the infallibility and general superiority of standard English that deviations from its precepts have even been linked on occasion with anti-social behaviour (Cameron Reference Cameron2012:95). Such tendentious associations serve as a reminder that popular discussions about standard usage are frequently embedded in a ‘double discourse’ in which concerns about putative linguistic decline are often more broadly symptomatic of deeper social, political or moral anxieties (Cameron Reference Cameron, Beal, Lukač and Straaijer2023:24).
In the court of public opinion, standard English is often construed in aesthetic and evaluative terms as the ‘best’, ‘most correct’ English, which is largely considered synonymous with the norms of standard written English. For many speakers, public figures, policymakers and pedagogues, standard English, however idealised and ill-defined, remains the embodiment of the English language, whereas for linguists it is one variety among many (Trudgill Reference Trudgill, Bex and Watts1999:118), albeit one invested with privileged social functions and considerable ideological freight.
Though standard English has a strong psychological hold on individuals, reinforced by prescriptive doctrines propagated by the education system, it is nonetheless evident that it can have multiple indexical associations for speakers (see Britain Reference Britain2017:290). This poses problems for reductive characterisations of standard English that define it in terms of its purported prestige, or speaker-based attributes such as level of educational attainment. There is evidence to suggest that speaker-based judgements of standard English are multidimensional, context-sensitive and by no means invariably positive (see for example, Garrett, Coupland and Williams Reference Garrett, Coupland and Williams1999). Similarly, definitions which posit standard English as the language used by ‘educated’ speakers (Quirk et al. Reference Quirk, Greenbaum, Leech and Svartvik1985:18) typically fail to acknowledge that the term ‘educated’ encompasses a range of interpretations (Coupland Reference Coupland2000:628; Lippi-Green Reference Lippi-Green2012:57).
Attempts to delineate standard and non-standard English by appealing to linguistic criteria also run into difficulty because they have tended to privilege certain conspicuous phenomena at the expense of other less noticeable ones. Inspection of the literature reveals that a core repertoire of morphosyntactic forms – a rogue’s gallery of ‘errors and solecisms’ (Cameron Reference Cameron2012:103) – is regularly invoked to demarcate standard from non-standard grammar. Among commonly cited non-standard grammatical forms are the following:
(1)
Multiple negation I don’t wanna catch nothing (LIC/Kieran/284)Footnote 1
(2)
Ain’t he was telling us that he ain’t gone back (LIC/Kate/868)
(3)
Non-standard past-tense forms then after that in the fifties come all the West Indians (LIC/John/082)
(4)
Subject–verb agreement they was all single women (LIC/Joan/637)
(5)
Non-standard relative markers cos in our school now we’ve got this camera what goes out (LIC/Kate/699)
(6)
Pronominal usage me and her never ever liked each other (LIC/Kelly/502)
Each of these examples illustrates non-standard phenomena that represent easy targets for prescriptivist critiques because of their relative social salience. Several exhibit regional and social patterns of variation and have lengthy histories extending back to Early Modern English (and beyond), undermining popular beliefs that these features are the products of deterioration from an earlier (and mythical) ‘Golden Age’ when the state of the language is assumed to have been ‘more grammatical’ (Labov Reference Labov2001:514). For many speakers, these non-standard variants represent everyday community norms, and they may even be conscripted as markers of group identity (see for example, Cheshire Reference Cheshire and Romaine1982). Speaker allegiance to such community-based norms is known to play an important role in bolstering their maintenance in the face of countervailing pressures to conform to their prescribed counterparts (Milroy Reference Milroy, Macaulay and Fought2004: 170).
The belief that the forms in (1)–(6) are resolutely non-standard derives from the conviction that they have canonical standard counterparts from which speakers deviate as the result of perceived carelessness, ignorance or incompetence. The received wisdom typically pays scant attention to the fact that linguistic behaviour is extremely sensitive to contextual factors, such that differences in context may elicit qualitatively different linguistic forms depending on the level of formality. The official consensus, as reflected in public policy documents (see for example, Cheshire Reference Cheshire, Bex and Watts1999a:147), at least until recently, is that the forms in (1)–(6) are not really appropriate to any social context. Indeed, Snell and Cushing (Reference Snell and Cushing2022:204) report that British school teachers habitually interpret features such as non-standard was, illustrated in (4) and documented since at least the Middle English period, as being ‘not even a dialect thing … just completely incorrect’, and ‘grammatically wrong’.
From a linguistic perspective, the categorisation of usage norms in terms of non-standard–standard grammatical oppositions, while superficially useful for variables such as we was/we were or I seen/I saw, cannot do justice to the full range of structural variation found in spoken English (see Cheshire Reference Cheshire, Bex and Watts1999a, Reference Cheshire, Cornips and Corrigan2005 and Section 2.4). Because such binary distinctions are strongly influenced by the codified forms associated with the standard written language, they ignore features of the spoken language that have eluded the codification process, notably, those that are intimately associated with the interactional and context-bound nature of speech. The indeterminate status of such features as either non-standard or standard, together with the recognition that conventional accounts of written English are inadequate descriptive surrogates of speech, has led to the rejection of standard spoken English as a coherent, reified entity. Scholars now stress the importance of envisioning standardisation as a ‘process that is permanently in progress’ (Cheshire and Milroy Reference Cheshire, Milroy, Milroy and Milroy1993:3) and theorise the standard spoken language in terms of a ‘set of abstract norms to which usage may conform to a greater or lesser extent’ (Milroy and Milroy Reference Milroy and Milroy2012:19).
Because the historical process of standardisation has played, and continues to play, a pivotal role in shaping lay perceptions of these abstract norms, it will be instructive to examine briefly how standardisation unfolded diachronically in order to arrive at a clearer understanding of its ideological ramifications.
2.3 The Ideology of the Standard Language
2.3.1 The Process of Standardisation and the Roots of Prescriptivism
The process of standardisation is one that does not arise naturally in the course of linguistic evolution but represents a case of deliberate human intervention (Milroy Reference Milroy2001a:535). As far as English is concerned, the roots of this process lie first and foremost in attempts to develop and regulate a uniform written language. The selection of a written variety, the first major step in the standardisation process according to Haugen (Reference Haugen1966), was an Early Modern English development. The rise of a written standard was catalysed by the advent of the printing press brought by William Caxton to England from the low countries towards the end of the fifteenth century. Its capacity to enable widespread dissemination of written texts fuelled the need for a written variety of the language which would be widely intelligible. Early attempts to develop such a variety were indebted to multiple sources (Samuels Reference Samuels1972), including Chancery English, the variety of written English used in state documents produced at the Exchequer in Westminster, as well as input from London-based and Central Midlands dialectal features (see Brinton and Arnovick Reference Brinton and Arnovick2017:316; see also Auer, this volume).
Subsequent phases in the evolution of a standard written variety acquired major impetus in the Modern English period, notably during the eighteenth century, widely regarded as the ‘heyday’ of prescriptivism (Beal Reference Beal2004). This was a period characterised by fervent efforts to codify the language and impose uniformity on rampant variability by prescribing some usages as ‘correct’ while condemning others. Some of the more strident injunctions of the English prescriptive tradition have filtered down virtually intact into the modern era. Thus, Lowth’s (Reference Lowth1776:139) formulation that ‘Two Negatives in English destroy one another, or are equivalent to an Affirmative’ continues to inform the vilification of multiple negation (e.g. I don’t have none), based on the ill-founded belief that such constructions violate mathematical and logical principles. In the quest to impose uniformity on English usage and repress variability, appeals to logic and to revered classical languages featured among the favoured strategies used by eighteenth-century grammarians to promote one variant as superior to others (Poplack, Van Herk and Harvie Reference Poplack, Van Herk, Harvie, Watts and Trudgill2002:94). Yet another strategy that paved the way to the stigmatisation of non-standard variants, while conferring legitimacy on counterparts sanctioned as standard, involved the attribution of specific non-standard forms to ‘undesirable’ social groups such as working-class speakers. This cemented associations in the public imagination between non-standard usage and social class which became increasingly prominent during the nineteenth century (Mugglestone Reference Mugglestone1995). The legacy of these associations persists in modern-day injunctions to avoid pronunciation shibboleths such as /h/-dropping, linked with ‘vulgar’, ‘ignorant’ and ‘lower class’ speech (Mugglestone Reference Mugglestone1995:107), despite the variable absence of /h/ being a widespread feature of British English dialects and attested in the language for centuries.
Summarising, over the course of the eighteenth and nineteenth centuries, the ideological foundations of standardisation were laid, leading to an increasing tendency in public discourse to interpret usage in terms of polarised contrasts: good/bad; right/wrong; prestigious/vulgar; educated/uneducated, and so on. Among the most enduring legacies bequeathed by the English prescriptive enterprise, strengthened by advances in literacy and mass education, is the popular conviction that the codified norms of written English, enshrined in grammar books and dictionaries, are the ultimate arbiters of usage, as opposed to the communicative competence of millions of speakers who use the language on a daily basis (Milroy and Milroy Reference Milroy and Milroy2012:22).
2.3.2 Prescription Versus Usage
In view of the ideological pervasiveness of the standard language, the dissemination of prescribed norms via normative vectors such as the school (see Snell, this volume), as well as public preoccupations with the ‘quality’ of the language, it would seem reasonable to enquire whether prescriptive discourse has had any appreciable impact on actual usage, as measured by its role in retarding linguistic change or curtailing linguistic variation. These outcomes are, of course, avowed goals of the prescriptive tradition, although until quite recently, they had received very little empirical attention (Auer and González-Díaz Reference Auer and González-Díaz2005:318; Anderwald Reference Anderwald2012:267; Hinrichs, Szmrecsanyi and Bohmann Reference Hinrichs, Szmrecsanyi and Bohmann2015:807). The current consensus is that prescriptive norms have had little effect on speech, although their impact on the standard written language is reported to be more profound. For example, Hinrichs et al. (Reference Hinrichs, Szmrecsanyi and Bohmann2015:819) note an increasing tendency in standard written English for restrictive relative clauses to be marked by that (e.g. that’s the house that Jane built) and a concomitant reduction in the use of relativiser which in these constructions. They ascribe this development, at least in part, to the effects of grammatical prescriptivism, whose precepts mandate the use of relativiser that in restrictive relative clauses, but stipulate that which should be confined to non-restrictive relative clauses (e.g. that’s Jane’s new house, which is very beautiful), in accordance with the normative predilection for establishing an isomorphic relationship between form and function.
Though speech is regarded as much less permeable to the effects of prescriptivism, there is some limited evidence that the visual properties of the written language have influenced the spoken language in the realm of ‘spelling pronunciations’, such as [wɛɪsˈkaʊt] and [fɒəˈhɛd] (‘waistcoat’, ‘forehead’) for earlier [ˈwɛskɪt] and [ˈfɒrɪd], as well as the use of [h] in hospital, herb, humble and humour, where it had previously not been pronounced (Mugglestone Reference Mugglestone and Mugglestone2006:359).
If prescriptive injunctions affect the course of linguistic change, then variants that are stigmatised, or deemed otherwise undesirable, should be expected to diminish over time, whereas their normatively sanctioned counterparts should increase (Poplack and Dion Reference Poplack and Dion2009:561). Studies of linguistic change offer very little evidence in support of this scenario, however. Even multiple negation, sometimes believed to have receded from standard written English as the result of eighteenth-century proscriptions, was in decline in most kinds of writing by the end of the seventeenth century (Nevalainen and Raumolin-Brunberg Reference Nevalainen and Raumolin-Brunberg2017:71) before doctrines of correctness held sway. This suggests that prescriptive ideologies may have reinforced a change in progress in the written language rather than having actuated it.
In yet other instances, prescriptive dictates have clearly had no influence on the trajectory of linguistic change. Witness the case of the nineteenth-century stylistic aversion to the progressive passive (e.g. the house is being built), well established in the language by the 1830s (Mugglestone Reference Mugglestone and Mugglestone2006:350), which has since replaced in both speech and writing its earlier prescribed alternative the house is building.
Among compelling recent demonstrations that prescription has little effect on the course of change is the vertiginous rise of quotative be like (e.g. I was like, ‘What courses have you taken?’), used to report speech and thought. This is widely touted as one of the most dramatic linguistic changes in spoken English during the past three decades (Labov pers. comm. in Cukor-Avila Reference Crystal, Watts and Trudgill2002:21–2). This construction, together with the use of like as a discourse marker (e.g. there was like a stranger coming towards me), has attracted much negative attention in the media, where it is stereotyped as an emblematic feature of the ‘inarticulateness’ of contemporary youth (see Buchstaller Reference Buchstaller2014:234–7). In spite of these unfavourable associations, recent research suggests that quotative be like has staked out its niche as the unrivalled exponent of the quotative system of contemporary youth, accounting for as much as 68 per cent of quotatives used by younger British speakers (Durham et al. Reference Durham, Haddican, Zweig, Johnson, Baker, Cockeram, Danks and Tyler2012). This is a change which comes at the expense of more traditional and prescriptively endorsed variants such as say (see Pichler and Cheshire, this volume).
2.3.3 Preposition Placement: A Cautionary Tale of the Relationship between Prescription and Praxis
Though the results of numerous investigations indicate that the prescriptive tradition wields little authority over spontaneous spoken language, these findings have not prevented scholars from reaching opposing, if highly questionable, conclusions. An excellent case in point concerns preposition placement in English, a feature of English grammar that constitutes ‘one of the favourite targets of … assiduous prescriptivists’ (Bergh and Seppänen Reference Bergh and Seppänen2000:312).
Illustrating with preposition placement in relative clauses, it can be seen that the position of the preposition depends on choice of relativiser. In oblique relative clauses (7–10 below), the preposition can either precede a WH-relativiser (who(m), which, whose), as in (7), a phenomenon known as ‘pied-piping’, or it can appear without an NP complement in front of the relativised position, as in (8)–(10), known as ‘preposition-stranding’ (Hoffmann Reference Hoffmann2005:1–2). In Modern English, variable preposition placement is licensed only with WH-relativisers. With the relativisers that, as in (9), and Ø, as in (10), preposition-stranding is obligatory.
(7)
I think the speed in which we speak as well is – it’s a lot faster (DECTE/ Y07i006a/0665)
(8)
I know people who have moved away but not who I’m really close with (DECTE/ Y10i024a/0057)
(9)
that’s the lady that we’re doing this for (LIC/Mandy/278)
(10)
yeah that’s the only place Ø I know Hackney for (LIC/Lou/741)
Pied-piping is generally considered the more stylistically ‘appropriate’ option, especially in careful public writing (Biber et al. Reference Biber, Johnson, Leech, Conrad and Finegan1999:107). This preference is believed to have its roots in eighteenth-century prescription, as illustrated in Bullen’s (Reference Bullen and Gedge1797:132) condemnation of preposition-stranding as ‘a most inelegant construction, to which the idiom of our language seems strongly addicted’ (see also Auer, this volume).
Beyond formal prose, the nature and extent of pied-piping and preposition-stranding remain less clear. Some ostensibly descriptive grammars of Modern English refer to pied-piping as the ‘unmarked’ option, while preposition-stranding is said to be the ‘marked’ alternative (Quirk et al. Reference Quirk, Greenbaum, Leech and Svartvik1985: §9.6). Perhaps the strongest and most controversial statement to this effect is made by Van den Eynden (Reference Van den Eynden, Klemola, Kytö and Risannen1996:444): ‘[d]iachronic, dialectal and standard English all show that stranding is not really an option with WH-[relative pronouns]; not now and not in the past’. Yet examination of written and spoken sources of British English depicted in Figure 2.1 calls this claim into question. Though pied-piping is preferred over stranding with WH-relativisers in all sources shown in Figure 2.1, in no dataset is pied-piping categorical, particularly in the spoken language data, where the rate of preposition-stranding is proportionally greater.
Figure 2.1 Rates of pied-piping and preposition-stranding with WH-relativisers in written and spoken Modern English. LOB = Lancaster-Oslo-Bergen Corpus; BNC = British National Corpus.
Equally damaging to Van den Eynden’s (Reference Van den Eynden, Klemola, Kytö and Risannen1996) claim are findings based on recent investigations of spontaneous speech, where WH-relativisers in oblique relative clauses are explicitly contextualised in relation to the other options with which they compete, namely that and Ø. Table 2.1, based on Levey and Pichler’s (Reference Levey, Pichler and Asahi2020) analysis of relativisation in four modern British varieties, shows rates of preposition-stranding and pied-piping in Berwick-upon-Tweed located near the Scottish border (Berwick), Tyneside in the north-east (DECTE), Salford in the north-west (RoSE) and the London boroughs of Havering and Hackney (Linguistic Innovators) in the south-east.Footnote 2
Table 2.1 Distribution of preposition-stranding and pied-piping in oblique relative clauses according to relativiser choice in four British speech corpora [Total N = 306]
| Relativiser | All data | ||||||
|---|---|---|---|---|---|---|---|
| WH- (including what) | That / Ø | ||||||
| Corpus | Variant | N | % | N | % | N | % |
| Berwick | Stranding | 2 | 100 | 68 | 100 | 70 | 100 |
| Pied-piping | 0 | 0 | 0 | 0 | 0 | 0 | |
| RoSE | Stranding | 3 | 100 | 31 | 100 | 34 | 100 |
| Pied-piping | 0 | 0 | 0 | 0 | 0 | 0 | |
| DECTE | Stranding | 8 | 80 | 63 | 100 | 71 | 97 |
| Pied-piping | 2 | 20 | 0 | 0 | 2 | 3 | |
| LIC | Stranding | 12 | 92 | 116 | 100 | 128 | 99 |
| Pied-piping | 1 | 8 | 0 | 0 | 1 | 1 | |
| Total N | 28 | 278 | 306 | ||||
A first important finding to emerge from these results is that over 90 per cent of oblique relative clauses in the aggregated dataset (N = 278/306) are marked by that or Ø, which only license preposition-stranding. In the varieties depicted in Table 2.1, WH-relativisers play only a very minor role in oblique relative clause constructions.
Moreover, even with WH-relativisers, pied-piping occurs no more than 11 per cent of the time (N = 3/28), contrary to Van den Eynden’s (Reference Van den Eynden, Klemola, Kytö and Risannen1996) claim. And, unusually, one instance of pied-piping occurs with non-standard what, as in (11), sometimes analysed as a complementiser (reported to inhibit pied-piping) rather than a relative pronoun.
(11)
well they knew how to teach but they didn’t know the subject on what they was teaching (LIC/Jennifer/203)
In brief, the results in Table 2.1 suggest that in everyday spontaneous spoken English, preposition-stranding in oblique relative clauses is the unmarked option and that pied-piping is a very minor alternative. Even where speakers are confronted with the possibility of pied-piping, as in the case of WH-relative clauses, they opt for preposition-stranding most of the time (see Biber et al. Reference Biber, Johnson, Leech, Conrad and Finegan1999:105). This finding dovetails with McDaniel, McKee and Bernstein’s (Reference McDaniel, McKee and Bernstein1998:309) claim that ‘preposition pied-piping is not a natural option in English, but rather a prescriptive artifact probably picked up during schooling’.
An important caveat to emerge from this cautionary tale is that the kinds of data and analytical frameworks that researchers appeal to in order to characterise the grammatical structure of spoken English severely constrain the conclusions they reach. A preoccupation with written forms of English, against which speech is often misleadingly assessed, coupled with a propensity to privilege particular theoretical accounts of grammatical use over the variability found in actual speech, can lead analysts down the garden path of misidentifying or ignoring important structural regularities in spoken English.
The following sections confront these issues in more detail and argue that by abstracting away from the variability inherent in spoken English, analysts run the risk of obscuring the way the language is actually used.
2.4 The Grammar of Spoken English
2.4.1 Approaches to the Study of Spoken English
Within generative and cognitively oriented frameworks, there is a long tradition of basing structural accounts of English on isolated, decontextualised examples, often accessed via introspective methodologies, rather than drawing on naturally occurring spoken data (Carter Reference Carter, Bex and Watts1999:153). As Milroy (Reference Milroy2001a:545) argues, it is difficult for analysts using these frameworks to remain immune to a set of assumptions that are conditioned by the ideology of the standard language. Chief among these is the belief that spoken English is relatively uniform and stable, that each difference in form is associated with a specific meaning or function (form–function symmetry), and that the forms and structures used in speech are isomorphic with those used in writing.
With the advent of variationist sociolinguistics (e.g. Labov Reference Labov1972), these beliefs have been severely challenged. A key revelation of the application of the variationist framework to spoken English is that variation is an inherent property of speech. This variation is not random or chaotic, but is rule-governed and structured (Weinreich, Labov and Herzog Reference Weinreich, Labov, Herzog, Lehmann and Makiel1968). At the heart of the notion of structured or orderly variation in speech lies the core theoretical construct of the linguistic variable. A variable can be defined as alternative ways of expressing the same referential meaning or grammatical function; for example, I have to leave/I’ve got to leave/I must leave are all alternative ways of expressing obligation, or deontic modality, in English. Crucially, the same grammatical function may be distributed across several different grammatical forms, suggesting that form–function asymmetries are ubiquitous in natural speech, contrary to the widely espoused assumption that every grammatical form has a dedicated function or meaning. Quantitative modelling of the distribution and conditioning of competing variants within a variable context offers a principled means of characterising the nature of the structured system hosting the variability in question.
Cumulative advances in sociolinguistics, based on fine-grained analyses of large quantities of natural speech data, have shown that the codified rules detailed in prescriptive grammars frequently have only a very tenuous connection with the patterns that underlie actual spoken English. Furthermore, few people have a scientifically informed idea of the structured nature of everyday speech, not least because ‘intuitions cannot begin to capture the robust and systematic … community norms that dictate usage’ (Poplack and Torres Cacoullos Reference Poplack, Torres Cacoullos, MacWhinney and O’Grady2015:275). Though these norms elude casual observation, they can be detected using an accountable quantitative methodology.
The following section addresses aspects of spoken English that have benefited from careful quantitative analysis, often generating findings that are at odds with prescriptive as well as theoretical accounts of the same phenomena.
2.4.2 Aspects of Spoken English
Subject–verb concord has been a long-standing preoccupation of prescriptive grammars. A much-discussed case concerns variable concord in existential there-constructions. According to prescriptive rules, the presence of a referentially plural post-verbal NP in such constructions should trigger verbal agreement or concord (e.g. there are books on the table). Yet in spontaneous spoken data, non-concord is often by far the preferred option with a post-verbal plural NP, as in (12) from a London pre-adolescent speaker:
(12)
cos there’s some girls I don’t like (ELPAC/20/6866)
Moreover, belying any expectation that such constructions are the province of the less ‘educated’, there is evidence indicating that in certain speech communities, it is the more ‘educated’ speakers who spearhead the use of non-concord in existential constructions, as shown in Figure 2.2, based on data from York English (Tagliamonte Reference Tagliamonte1998). With the exception of males educated beyond sixteen years, non-concord is the default option in York English for all other speaker cohorts, and it is most frequent in the speech of females educated beyond sixteen.
Figure 2.2 Rates of non-concord in plural existential constructions (past tense only) in more and less educated male and female speakers in York English.
What can account for non-concord with a plural NP in existential there-constructions? A stock response to such usage facts is that they are the reflex of change, but inspection of the historical record shows that non-concord has considerable time-depth in the language, with attestations documented as far back as the Old English period (Visser Reference Visser1963:62). Although theoretical treatments have often resorted to highly elegant and complex syntactic theories to account for agreement phenomena in existential there-constructions (see Meechan and Foley Reference Meechan and Foley1994), a recurrent observation is that in contemporary spoken English, such constructions tend towards invariance. In fact, in the pre-adolescent spoken language corpus from which (12) was extracted, there are no instances of there are + plural NP. This suggests that there’s in (12) behaves like its French and German invariant counterparts il y a and es gibt, neither of which exhibits concord effects with a following plural NP (Cheshire Reference Cheshire, Bex and Watts1999a). In other words, the weight of the available evidence suggests that there’s is in the process of becoming a ‘fixed’ or lexicalised construction (Cheshire Reference Cheshire, Bex and Watts1999a:138). Further support for this interpretation can be adduced from the multiple discourse-pragmatic functions of the existential there-construction, including topic management and the introduction of discourse-new referents (Cheshire Reference Cheshire1999b:71). For example, in the case of discourse-new referents, it is well known that subject position in English correlates strongly with given rather than new information. Existential there-constructions can be deployed as a preferred syntactic strategy to avoid placing a discourse-new referent in subject position (e.g. there’s a woman who wants to see you). Thus, informational-structural preferences, together with on-line, context-dependent factors, may conspire in creating a preference for there’s to be stored and accessed as a prefabricated or unanalysed sequence (Cheshire Reference Cheshire1999b:71). As Pawley and Syder (Reference Pawley and Syder1983:552) observe, such usages may be outgrowths of the specific ‘ecology’ of spoken grammar, in which constructions evolve to suit the conditions and purposes of face-to-face interaction.
There are other instances where the descriptive categories and models developed in relation to the written language either impose misleading patterns on speech or fail to account for regular grammatical tendencies that characterise everyday spoken English (Cheshire and Stein Reference Cheshire, Stein, Cheshire and Stein1997:11). One such case involves the use of this in spoken English to introduce an indefinite NP, as in (13), where prescriptive norms would likely require a(n).
(13)
then he… he met this young lady and he’s been with her(i) ever since she’s(ii) er what they call it Thailand somewhere over that way […] she’s (iii) a nice woman(iv) very nice Margee(v) very nice get on I get on well with her(vi) (LIC/Doug/562–4)
Like there’s + plural NP, the use of indefinite this exhibits no known patterns of regional affiliation. Opinions remain divided as to whether it is a non-standard grammatical feature (Prince Reference Prince, Joshi, Wenner and Sag1981) or a general feature of spoken English (Trudgill Reference Trudgill, Bex and Watts1999), pointing once again to the complexities of establishing firm boundaries between colloquial and non-standard varieties (Cheshire and Stein Reference Cheshire, Stein, Cheshire and Stein1997:2).
In conventional grammatical descriptions tailored, as we have already seen, to the written language, this is characterised as a proximal demonstrative in structural opposition with the distal demonstrative that (e.g. this book here versus that book there). Yet this structural opposition fails to provide an adequate characterisation of the use of this in (13). Absent a satisfactory conventional account, the use of this to mark a specific discourse-new referent (i.e. this young lady = a certain young lady) tends to be handled as an exceptional case, or as an extension of its ‘primary’ use as a spatial deictic.
Closer scrutiny of the distribution of indefinite this in relation to a(n) suggests that in spoken English, indefinite this introduces discourse-new referents that are ‘more pragmatically important’ (Cheshire Reference Cheshire1989:51), ‘prone to topic shift’ (Chiriacescu Reference Chiriacescu2011:48), and thematically central to the unfolding discourse (Rühlemann and O’Donnell Reference Rühlemann, O’Donnell, Aijmer and Rühlemann2014:355). This is illustrated in (13), where this young lady is referentially persistent after its initial mention, as evidenced by pronouns and nouns (indicated by subscripted numerals) in the ensuing discourse that are anaphorically co-indexed with the this-marked referent. Systematic quantitative analysis has revealed that indefinite this, as opposed to a(n), has a dedicated discourse-structuring role in spoken English, where it is used as a discriminatory marker of pragmatically prominent indefinites, notably, those which encode human referents and which are probabilistically prone to surface as topics in the subsequent discourse (see Levey, Klein and Abou Taha Reference Levey, Pichler and Asahi2020).
Had the analysis of indefinite this been confined to the sentence level and appealed to isolated occurrences, as per conventional grammatical analysis, the ‘hidden’ patterns outlined above would have remained all but invisible. Such findings sound a cautionary note that claims about the way English is used, based on traditional accounts of the standard written language, should not be uncritically equated with the data of actual spoken usage (see Poplack Reference Poplack, Shin and Erker2018:8).
2.5 Received Pronunciation
No discussion of standard English would be complete without mentioning Received Pronunciation (hereafter, RP). At the outset, it is important to acknowledge that RP is not a widespread accent, although it has served as a reference model in the teaching of English as a foreign language, and it has been claimed to be an ‘implicitly accepted social standard of pronunciation’ (Cruttenden Reference Cruttenden2001:78). It is of course perfectly feasible to approximate the grammatical norms of standard English using any accent. In fact, this is a very real possibility in view of the fact that rough estimates of the proportion of RP speakers in modern Britain range from a paltry 3 per cent (Trudgill Reference Trudgill2001) to 5 per cent (Milroy Reference Milroy2001b) of the population. Clearly, on the basis of these figures, RP cannot be remotely construed as a mainstream accent (Milroy Reference Milroy2001b).
What is RP? Much like the notion of standard English, the concept of RP suffers from a certain degree of definitional imprecision, not least because ‘no two British phoneticians are likely to agree on where the line between RP and non-RP is to be drawn’ (Lewis Reference Lewis1985:247). Some of this indeterminacy derives from the fact that although RP has been extensively described, ‘there has been very little in the way of objective quantified investigation of its variability’ (Wells Reference Wells1982:279; but see for example Fabricius Reference Fabricius2000). Still, some variability in RP is implicit in the labels that have been associated with it: ‘General RP’, ‘Refined RP’ and even ‘Regional RP’ (Cruttenden Reference Cruttenden1994:80). Indeed, since its genesis in the nineteenth century, observers have conceded, albeit in restrained terms, that RP accommodates a certain amount of variation. In one of the earliest descriptions of RP, the phonetician Alexander Ellis (Reference Ellis1869:23) affirms, for example, that ‘we may … recognise a received pronunciation all over the country; not widely differing in any particular locality, and admitting a certain degree of variety’ (my emphasis).
The traditional hallmark of RP, as Ellis (Reference Ellis1869) makes explicit, is its non-localisability (Fabricius Reference Fabricius2000:43), or the absence of its affiliation with any particular region (Trudgill Reference Trudgill2001). Historically, the status that RP accrued as a ‘correct’ or ‘educated’ non-local accent derives in part from nineteenth-century antipathy towards regional speech, symbolically indexing uneducatedness, illiteracy and linguistic impropriety (see Mugglestone Reference Mugglestone and Mugglestone2006:362–3). As a powerful social symbol, RP acquired the trappings of a social class accent used by the ‘upper classes’ and was transmitted via high-status social networks rooted in the English public (i.e. private boarding) school system (Milroy Reference Milroy2001b).
RP’s putative aesthetic qualities, equated by one historian of the language with the ‘most pleasing and sonorous’ form of pronunciation (Wyld Reference Wyld1934:605), by no means elicit uniformly positive evaluations, despite its historical and contemporary ties with statusful speech. The findings of perceptual dialectology have shown that although RP speakers may be variously assessed as ‘competent’, ‘reliable’, ‘educated’ and ‘confident’, they are also rated less positively with regard to traits such as ‘friendliness’, ‘companionability’ and ‘sincerity’ (Giles Reference Giles, Mayor and Pugh1987).
There are indications that the ideological connotations of RP as a ‘high-status’ or ‘upper-class’ accent have been diluted by the apparent democratisation of English society, as witnessed, for example, by the greater tolerance shown towards regional accents in the broadcast media. Militating against this conclusion, however, are recent findings indicating that accent discrimination is alive and well in modern Britain. Watt, Levon and Ilbury’s (Reference Watt, Levon, Ilbury, Beal, Lukač and Straaijer2023:41) longitudinal investigation of British accent ratings from 1969 to 2019 revealed that although RP may ‘no longer stand head and shoulders above other accents’, the social hierarchy of accents in 2019 is not vastly different from what existed fifty years earlier.
Although RP may still occupy ‘perceptual high ground’ among British accents (Watt et al. Reference Watt, Levon, Ilbury, Beal, Lukač and Straaijer2023:34), it has not been entirely immune to dialect levelling processes affecting mainstream British varieties. Among the innovations used by younger RP speakers, Fabricius (Reference Fabricius2000:36–7) notes the following:
(19)
t-glottalling in word-internal syllable-final as well as word-final positions, involving the variable replacement of [t] by [ʔ], as in ‘button’ [ˈbʌʔn̩]
One identified source of these innovations is Estuary English, originally characterised as a mixture of non-regional and local south-eastern English pronunciation and intonation (Rosewarne Reference Rosewarne1994:3), and subsequently popularised as the possible ‘RP of the future’. Though dialect contact processes, driven by social mobility and population movements, are operating in the south-east of England (and elsewhere), and are implicated in the ongoing reduction of linguistic differences between geographically contiguous areas, these processes appear to be advancing in different counties surrounding London at different rates and to varying extents, depending on the linguistic feature in question (see Britain Reference Britain2009). This means that there is still considerable variation in the geographical area identified as the homeland of Estuary English, diminishing, rather than bolstering, its existence as an identifiable, unitary variety (Przedlacka Reference Przedlacka2002:97).
There are currently no grounds for arguing that Estuary English, as idealised as it may be, will eventually assume the functions of a standard ‘reference model’ of any kind, as inferred from Rosewarne’s (Reference Rosewarne1994:8) impressionistic observation that Estuary English is beginning to expropriate functions previously associated with RP. This is because standard varieties are institutionally imposed and sanctioned, whereas Estuary English lacks any official endorsement as a reference or ‘prestige’ model (see Watt and Milroy Reference Watt, Milroy, Foulkes and Docherty1999:43).
2.6 Concluding Remarks
A central idea foregrounded in this chapter is that the concept of standard English as a coherent entity is very difficult to pin down because it essentially refers to an abstraction. As has been emphasised throughout this chapter, the boundaries between standard and non-standard English remain plagued by a good deal of uncertainty, especially when the spoken language is factored into consideration (Milroy and Milroy Reference Milroy and Milroy2012:150). Herein lies the gulf between linguists’ characterisation of standard English and public understanding of this concept. Whereas many linguists currently resist ontological perspectives that reify the standard variety as an identifiable entity (Coupland Reference Coupland2000:632), the idea that there are no absolute and infallible norms of ‘correctness’ flies in the face of the unshakeable public faith in the existence of standard English as a concrete, enduring reality (see Cameron Reference Cameron2012:98).
It is the ill-defined nature of standard English that enables distinctions between written and spoken language, formal and informal speech, and vernacular and ‘correct’ usage to be so easily distorted. The elision of these distinctions paves the way to pitting standard English, as an idealised and unchanging monolithic entity, against everyday speech characterised by variation and change, popularly decried as signs of structural deterioration.
Though empirically grounded and scientifically accountable investigations of spoken English have gained traction over the past several decades, enabling many simplifications and popular stereotypes about the spoken language to be challenged, there remain numerous gaps in our knowledge. To address these, it is important to broaden the empirical investigation of the spoken language in its social context. Discourse structure and pragmatics are singled out by Crystal (Reference Crystal, Watts and Trudgill2002:244) as areas where it is currently problematic to effect systematic comparisons between standard and non-standard English. Advances in our understanding of discourse-pragmatic features in everyday spoken English such as discourse markers (like, you know, etc.) and general extenders (and stuff, and things like that, etc.) have long been hampered by persistent stereotypes, informed by prescriptive ideologies, which portray these features as meaningless fillers and superfluous hesitation markers. Increasing attention paid to structured variation in the use of these features, as well as the interactional functions they perform in discourse, has done much to debunk the myth that these are extra- or agrammatical components of language (Pichler Reference Pichler2010:582; see also Pichler and Cheshire, this volume).
In the final analysis, it is only via more empirically accountable investigations of naturally occurring discourse that we can hope to achieve a better understanding of what speakers actually do with English, offering a much-needed corrective to many prescriptive and theoretical accounts propagating grammatical fictions that have little basis in everyday usage.
