1.1.1 Chronology, Including Political and Socio-Economic Developments

The history of the English language has traditionally been divided into three periods (based on inflectional characteristics), whose starting and end points are associated with historical landmark events that had a (long-term) effect on the development of the English language, notably the Old English, the Middle English and the Modern English periods. The historical events demarcating the periods did not have an immediate effect on the language, which is why the dates of the periods related to the history of English often vary somewhat (see Beal Reference Beal2004:1–2) and particularly Curzan (Reference Curzan, Bergs and Brinton2012 [2017]) for a detailed discussion of chronological divisions). As the traditional division has left a lacuna covering the twenty-first century, the current chapter, in line with Beal (Reference Beal2004:1–2), considers the Later Modern English period to last until the end of World War II.

In terms of landmarks, the Old English period starts in 449 ce with the settlement of the Angles, Saxons and Jutes, who had arrived in Britain from the north-western part of the European continent. The somewhat differing but closely related West Germanic dialects of these settlers served as the basis for the – retrospectively termed – Old English language or the Anglo-Saxon dialects. According to the Oxford English Dictionary (OED) online,Footnote ¹ the name Englisc was first recorded in relation to the language in an Old English translation of Bede’s Historia ecclesiastica gentis Anglorum. The earliest substantial texts written in Old English have been dated to the eighth century. As regards contact with other languages, the Old English period is marked by influence of Celtic, Roman and Old Norse (Viking incursions) that also left their mark on the vernacular language. The end of the Old English and the start of the Middle English period is marked by the Norman Conquest in 1066. The death of Edward the Confessor (c. 1003–1066) led to conflicts regarding his succession, from which William of Normandy (c. 1028–1087) emerged successfully. His becoming king of England in 1066 had a great effect on the vernacular language, particularly at the level of the lexicon, with many Norman French words entering the English language, as well as changed patterns of word formation and phonological effects on the language. The starting point of the Modern English period is linked to the introduction of the printing press with movable type by William Caxton in 1476. This period is sub-divided by language historians into the Early Modern English (1476/1500–1700) and the Later Modern English periods (1700–1945). The year 1700 as a dividing point can be linked to the Acts of Union (1707) that united the Kingdom of England and the Kingdom of Scotland into one kingdom that is Great Britain. Moreover, it lies before the Industrial Revolution, which had enormous effects on demography and introduced new manufacturing processes linked to technical inventions. Other important events that had a significant impact on the English language during the Early Modern English period are (a) the Reformation in the sixteenth century, when the Church of England broke with the Catholic Church and the Pope, replacing the latter with the king as the head of the Church, (b) the Renaissance during the Elizabethan era (1558–1603) that saw a renewed interest in classical learning and a change of perspective on science, (c) the first colonial ventures under the reign of Elizabeth I, (d) the publication of the King James Bible (1611), and (e) conflicts based on religio-political disagreements like the English Civil War (1642–1651), the Commonwealth (1649–1653, 1659–1660), the Protectorate (1653–1659), and the Restoration of the Stuart monarchy in 1660. During the Late Modern English period (1700–1900), the second Act of Union in 1801 saw the annexation of Ireland to Great Britain. The period is also particularly marked by the Industrial Revolution (c. 1760–1830), including the revolution in transportation, as well as demographic developments like continued urbanisation, which led to higher rates of literacy. A landmark event was the Elementary Education Act (1870), which set the framework for compulsory elementary schooling for children aged between five and twelve. The rise of literacy rates (in combination with technological developments) also had an impact on the publishing market, leading to the availability of mass media. In 1922, the Irish Free State became independent from the United Kingdom. Despite these developments, English has obviously continued to play an important role in the Republic of Ireland, as well as Scotland and Wales, given that it is the mother tongue of the majority of the population. A factor that has had more of an effect on English since World War II is the immigration of people from former British colonies and other countries, particularly in the private sphere. As Gramley (Reference Gramley2012:182) notes, in numerical terms, the most frequently spoken immigrant languages in the United Kingdom are ‘Punjabi, Bengali, Saraiki, Urdu, Sylheti, Cantonese, Greek, Italian, Black British English/Creole, Gujarati, and Kashmiri’ (see also Karatsareas, this volume).

1.1.2 Literacy Levels and Text Production in the History of English

The study of the history of the English language is largely based on written documents. As the production of documents is dependent on whether one was able to write, the texts and text types from different periods at our disposal for investigation reflect which members of society were literate at the time and what the contemporary socio-political concerns were. In the history of English, literacy rates were socially stratified, and determining these rates is challenging due to literacy being a rather imprecise concept: are we focusing on the ability to read and/or the ability to write, and at what level were people able to do so? Until the Elementary Education Act in 1870, writing skills were determined based on so-called signature literacy, namely the ability to write one’s name. Early statistical overviews were moreover based on signature literacy in marriage registers (that only capture part of the population), which already indicates that the evidence is incomplete and does not allow for generalisations to be made (cf. Lawson and Silver Reference Lawson and Silver1973:34; see also Reay Reference Reay1998). Despite not being accurate, the statistical overviews allow us to observe an increase in literacy rates (ability to write) over time.

In Anglo-Saxon England, two languages – Latin and Old English – were used for reading and writing. Texts written in the vernacular Old English can be found from the eighth century onwards and increased in volume in the following centuries. Godden (Reference Godden and Gameson2012:586) states that documents in Old English were produced soon after the conversion to Christianity and Old English remained associated with legal texts such as law-codes and charters in the first instance. There are indications of English having been used for education as early as the seventh century, with English glossing Latin in some early glossaries (Godden Reference Godden and Gameson2012:586). Old English was also increasingly used for religious and philosophical texts (many linked to King Alfred’s educational reform around 890); moreover, vernacular poetry exists from the ninth century and survives in manuscripts from the tenth and eleventh centuries. As far as the audience for these works and its implications for literacy is concerned, the bulk of the texts produced were aimed at clerics who had no Latin, as well as the landowning laity. The period is also characterised by an oral culture, as reflected in the reading aloud of Old English texts such as royal writs, charters, sermons and poetry (Godden Reference Godden and Gameson2012:589–90).

During the Middle English period (c. 1066–1476/1500), English vernacular literacy is also best considered in relation to other languages (i.e. Latin and French). After the Norman Conquest (1066), French was associated with the court and the aristocracy, as well as with literature (Wright Reference Wright2020a), while Latin served for governmental and administrative purposes. Throughout the period, education was still largely in the hands of religious houses. In the fifteenth century, public schools were established in all English counties which then existed, alongside education offered in religious houses and great households, and private learning, such as merchants training their apprentices (Orme Reference Orme2006). It may be argued that education was characterised by a wide range of different ‘literacies’. According to Rees Jones (Reference Rees Jones, Mostert and Adamska2014:220), literate behaviour in urban centres (that had higher levels of text production) was influenced by the complexity of social organisation and its occupations. It can generally be noted that literacy was socially stratified during the Middle English period, that is, the elite, the clergy, the gentry and merchants were able to produce different kinds of texts, such as administrative and legal texts, scientific and medical handbooks, philosophical and religious works, historical accounts, travelogues, fiction, and correspondence.

During the Early Modern English period (c. 1476/1500–1700), literacy continued to be socially stratified in that the variety of training opportunities (that did not include the lower social strata at the time) led to different levels of literacy. For instance, boys from the elite received a classical training where Latin was taught alongside the English vernacular. A common path thereafter was to attend a university like Oxford or Cambridge, or to train in the Inns of Court (Brooks Reference Brooks, Barry and Brooks1994:54; Lawson and Silver Reference Lawson and Silver1973:91–152). Boys from the middling orders tended to do an apprenticeship (Lawson and Silver Reference Lawson and Silver1973:122–5). In contrast, girls from well-off families received their education at home or, from the early seventeenth century onwards, also in private boarding schools. As regards literacy rates, Cressy (Reference Cressy1980:177) notes that in 1500, more than 10 per cent of the male population was able to read.Footnote ² Reay (Reference Reay1998:40) observes that literacy rates for men varied according to occupation and geography. While, in 1580–1700, the gentry and professionals were almost fully literate in many areas, yeomen, traders and craftsmen had higher literacy levels in the south than in the north (c. 70 per cent in London and Middlesex, c. 60 per cent in East Anglia, and c. 54 per cent in the North). Husbandmen and labourers had the lowest rates, ranging from 15 to 25 per cent. As regards text production, an increase in text types can be observed, such as trial proceedings, diaries, drama and biographies.

For the earlier part of the Later Modern English period (c. 1700–1945), the signatures in marriage registers were considered as signs of literacy by education historians (linked to the Hardwicke Marriage Act of 1753; cf. Houston Reference Houston1982:200).

Even though an increase in literacy rates can be determined in comparison to the Early Modern English period, as illustrated in Table 1.1, which presents male and female occupational literacy in England in 1700–70, literacy rates increased rapidly after 1840 (More Reference More2000: 58) as a result of the Elementary Education Act (1870) (see for instance Altick Reference Altick1957:171 for literacy rates in England and Wales 1841–1900).Footnote ³ As illustrated for the Early Modern English period, education opportunities varied according to sex and social class before this landmark event. While the education and training of the male elite was considered of great importance for the country, the opinion of the elite was that ‘too much literacy among the population at large was a danger to the established order’ (Lawson and Silver Reference Lawson and Silver1973:179). In other words, elementary education for the labouring poor should be limited and ‘designed to inculcate mainly practical religion, social obedience and low-level occupational skills’ (Lawson and Silver Reference Lawson and Silver1973:180). Nevertheless, the lower social groups were able to receive some schooling in Sunday and charity schools, as well as through self- and peer-schooling (for more details regarding education opportunities in Late Modern England, see Auer Reference Auer, Rutten, Vosters and Vandenbussche2014, Reference Auer, Auer, Schreier and Watts2015; Auer and Hickey in press). As regards text production, while a wide range of text types were produced by the elite and the middling sort, ego documents including diaries and letters (including petitions) were sometimes produced by the lower layers of societies. It is thus in the late eighteenth and early nineteenth centuries that we have autograph texts by the labouring poor.Footnote ⁴

1.1.3 Empirical Bases for Investigations of the History of the English Language

Investigations of the English language in different periods as well as language change over time have largely been based on electronic text corpora and databases, the creation and availability of which has been increasing at great speed since the late twentieth century. This is in line with the development of the field of corpus linguistics, which has had a great impact on the fields of diachronic and synchronic English linguistics. An important source for the history of the English language is the Helsinki Corpus of English Texts: Diachronic and Dialectal (HC), which covers the period c. 750 to 1710 and contains a range of different text types (that were available in different periods).Footnote ⁵ Another diachronic corpus that was constructed in a similar manner is ARCHER: A Representative Corpus of Historical English Registers. This is a multi-genre corpus containing British and American English texts from the period 1600–1999.Footnote ⁶ Many corpora available today focus on a particular genre and/or text type (e.g. newspapers, scientific texts, plays, depositions, novels, letters)Footnote ⁷ There is a recent trend to create corpora based on manuscript (rather than edited) material from archives, thereby emphasising the combination of historical linguistics, textual history and philology. Selected examples are The Middle English Grammar Corpus (MEG-C), a multi-genre corpus covering the period 1325–1500,Footnote ⁸ an Electronic Text Edition of Depositions 1560–1760 (ETED; Kytö, Grund and Walker Reference Kytö, Grund and Walker2011), and the Bluestocking Corpus containing letters by Elizabeth Montagu written between the 1730s and the 1780s.Footnote ⁹ As regards online databases, printed material can be accessed via Early English Books Online (EEBO),Footnote ¹⁰ Eighteenth Century Collections Online (ECCO),Footnote ¹¹ and Nineteenth Century Collections Online (NCCO),Footnote ¹² to name a few examples. As previously pointed out, literacy rates throughout the history of English determine the availability of written sources from different social groups. Within this context, historical sociolinguistic studies that cover the entire social stratum can only be carried out from the Late Modern English period onwards.

1.1.4 English from a Typological Perspective

The English language belongs to the Indo-European language family, which groups many languages in Europe and parts of Asia that have structural relationships. Within that family, English is part of the Germanic – and more precisely the West Germanic – branch. Other languages belonging to the latter branch are Frisian, Dutch, Afrikaans, German and Yiddish; these are thus the languages that are structurally most similar to English (cf. Henriksen and van der Auwera Reference Henriksen, van der Auwera, König and van der Auwera1994). A feature that can be found in these (and other Germanic) languages is, on a morphological level, the distinction between present and preterite forms in the verbal system. More precisely, strong verbs that form the past tense with ablaut, that is, a system of regular vowel variations in the same root (give ~ gave), contrast with a system of weak verbs whose past-tense forms contain an alveolar or dental suffix (like ~ liked). Concerning syntax, the Germanic languages have V2 (verb second) order in declarative main clauses. In line with the S(ubject)-V(erb)-O(bject) word order, prepositions (rather than postpositions) occur in the languages. As regards lexicon, while migration and contact with other languages have led to different lexicons, a commonality of the Germanic languages is the formation of new words by combining nominal, verbal and adjectival stems with derivational suffixes such as ‑dom and ‑ly in English today (cf. Hilpert Reference Hilpert, Heine and Narrog2011:709; for more detailed discussions, see Lass Reference Lass1994 and Harbert Reference Harbert2007).

Like other languages in use, the English language has changed since its beginnings. Factors that explain these changes have been categorised by Labov as (a) internal factors, such as linguistic reasons for change including chain shifts or lexical diffusion (Reference Labov1994), (b) social factors such as social class, age, gender, neighbourhood or ethnicity (Reference Labov2001), and (c) cognitive and cultural factors (Reference Labov2010). Further reasons for language change are dialect and language contact (for English see for instance Schreier and Hundt Reference Schreier and Hundt2013; see also Sharma, this volume; Fox, this volume). Language change can also be the result of a combination of these factors. The English language has undergone many changes that have affected its structure and sounds since the Old English period. For instance, while Old English is considered to be a synthetic language with a high number of inflections for case, number, gender, tense and mood as well as other grammatical categories, present-day English has become an analytic language with a much more simplified inflectional morphology. The loss of inflections has led to a greater importance of the role of prepositions and the establishment of a fixed word order (i.e. S(ubject)-V(erb)-O(bject)). Contact with other languages throughout history (e.g. with Latin, Old Norse and Norman French) has also had a great impact on the lexicon of English. The Germanic lexical basis has been expanded through the addition of words from Italic and Romance languages, and others, as well as the creation of new words (for details, see for instance Durkin Reference Durkin2014). Examples of changes related to different linguistic levels, notably phonology, morphology, syntax and lexicon, that have had an important effect on the language will be provided in the diachronic sections below.

1.2.1 General Background

Before the arrival of the Germanic tribes in Britain, Celtic tribes had settled on the island (c. first millennium bce), and the Romans invaded in 43 ce and expanded their control before leaving around 410 ce to defend their empire against invaders. The linguistic traces can particularly be found in place names, so-called toponyms, for example Kent ‘border’, Avon ‘river’ from Celtic, or Latin castrum ‘camp, fort’ that we find in Lancaster or Manchester (palatalised form). The West Germanic dialects that were brought to Britain in the fifth century were first documented in runic inscriptions. From the eighth century onwards, legal, religious, documentary and literary texts were written in the Roman alphabet, which was introduced by Christian missionaries. The majority of manuscripts containing Old English texts that have survived until today originated from the so-called West Saxon area. This is due to King Alfred’s educational reform which strongly supported the translation of texts from Latin into Old English, for political reasons (i.e. Viking raids), and other circumstances like the preservation of manuscripts. The raids by the Vikings that started in the late eighth century eventually led to them settling in the eastern part of England in the ninth century, where the so-called ‘Danelaw’ (i.e. the area where the laws of the Danes prevailed) was established. The contact between the Anglo Saxons and the Norsemen also had an effect on the English language, especially the lexicon (cf. Henriksen and van der Auwera Reference Henriksen, van der Auwera, König and van der Auwera1994:16).

The arrival of the Germanic peoples in England was followed by the rise of regionally different Old English dialects, which eventually replaced Common Brittonic and Latin. The Old English language, which is recorded from the eighth century onwards, can be divided into four main dialects that are associated with Anglo-Saxon kingdoms, notably (a) Northumbrian (north of the Humber) and (b) Mercian (between the Humber and the Thames), which are often grouped together as Anglian, (c) West Saxon (south) and (d) Kentish (south-east). As previously indicated, the majority of the extant texts are in the West Saxon scribal tradition. As this region did not come under Danish rule, which affected the Anglian areas, West Saxon was less influenced by Old Norse.

1.2.2 The Old English Language

1.3.1 General Background

The Norman Conquest in 1066 led to a greater Norman French presence in England, particularly at the upper level of society. It is noteworthy that due to Edward the Confessor’s (1041–1066) Norman origins, French had already been used at the Royal Court, which was then continued under the reign of William the Conqueror (1066–1087). According to the Domesday Book, which recorded landowners in 1086, the majority of feudal overlords at the time originated from Normandy. Similarly, a lot of high church leaders were French. Overall, however, the Normans did not exceed 5 per cent of the population (cf. Kibbee Reference Kibbee1991:9), and the majority of the inhabitants in England therefore remained English-speaking. Due to the association of Anglo-Norman (i.e. the variety into which Norman French developed in England after the Norman Conquest) with the upper layers of society, the variety gained the prestige that English had previously held. The effect of Anglo-Norman on English varied depending on the geographical location, for example Norman influence was stronger in the south and south-east, and on the type and amount of contact (for more details regarding Anglo-Norman, see for instance Ingham Reference Ingham2010, Reference Ingham2012; Timofeeva and Ingham Reference Timofeeva and Ingham2018). In addition to Anglo-Norman, continental French was also introduced through the arrival of groups of continental French speakers from the middle of the twelfth century onwards. At that point in time, the descendants of the Anglo-Norman conquerors had already started to shift to English, which may be taken as an indication that the status of English had changed. Another historical event that had an effect on the growing status of English was the loss of Normandy under the reign of King John (1199–1216) in 1204. As a result, Normans holding fiefs in both countries had to decide on one or the other country, and the connection to Normandy lessened severely, and the status of French was therefore also affected. The existence of different languages in the first half of the Middle English period is reflected in written records as the languages took on different functions. While French was firmly established as the language of the legal system and also of literature until the mid thirteenth century, English was gradually used in different domains from the early fourteenth century onwards. Schendl (Reference Schendl, Bergs and Brinton2012:508) notes that ‘the extent of English–French bilingualism in the thirteenth century is a matter of controversy’ but that ‘English had become the first language even for the vast majority of English–French bilinguals’ by the end of the century. Latin also played an important role during the Middle English period, particularly as the language of administration/recording, scholarship and the church. The shift to English in administration took some time, notably following a period of multilingualism that was maintained by professional clerks. The rise of English as the dominant language was fostered by demographic and social changes such as the increased urbanisation until the mid fourteenth century and the plague (Schendl Reference Schendl, Bergs and Brinton2012:508).

In the earlier part of Middle English, written sources displayed a lot of regional variation. The traditional distribution of the Middle English dialects is between Kentish (south-eastern part of England), Southern (west of Sussex, south and south-west of the Thames), Northern, West Midland, and East Midland, with the latter having had a stronger Old Norse influence due to the previously mentioned Danelaw. As a written standard did not exist at the time, dialectal differences are reflected in writing, with distinctions between the North and the South being particularly striking: for example Southern /tʃ/ (church) vs. Northern /k/ (kirk), the Southern third-person singular present tense verbal suffix ‑th (goeth) vs. Northern ‑s (goes), Old English ā being realised as /ɔ:/ in the South (stone) and as /e:/ in the North (stane), as well as third-person plural pronouns they/them in the North and hi/hem in the South. The gradual rise of supralocal norms from the fifteenth century onwards led to dialect levelling, and the variation of forms was further reduced (at different speeds in different text types) with the promotion of norms that would standardise written English. While the regional factor plays an important role in the Middle English period, social factors explaining language change become more relevant from the Early Modern English period onwards.

1.3.2 The Middle English Language

1.4.1 General Background

The many socio-political developments during the Early Modern English period, including the Reformation, the Renaissance and the beginning of colonisation, as well as important socio-cultural landmarks such as the introduction of the printing press with movable type in 1476 and the publication of the Book of Common Prayer (1549) and the King James Bible (1611), had a great impact on the English language. While the printing press, the Book of Common Prayer and the King James Bible had a normative effect on the written language, the Renaissance and colonisation introduced new lexicons from other languages. As previously discussed in relation to literacy levels and text production, the development of English vernacular literacy during the Middle English period (c. 1066–1500) is often viewed in relation to other languages, notably Latin, Anglo-Norman and French for the latter period. From 1500 onwards, the classical languages played an important role, particularly regarding lexicon, and in certain domains, Latin as the language of learning, for instance. This is also the period during which English became more dominant again and acquired a wide range of different functions, such as in administration and also learning. Moreover, the previously observed regional variation in English was gradually superseded by supralocal forms: the language underwent processes of standardisation.

1.4.2 The Early Modern English Language

1.5.1 General Background

This period was characterised by many important socio-political developments, notably political changes like the Acts of Union (1707), passed by the English and Scottish Parliaments, that led to the creation of the United Kingdom of Great Britain, as well as the second Act of Union (1801) that added Ireland to the United Kingdom. In 1922, Ireland left the United Kingdom and became a Free State within the British Commonwealth, followed by the creation of the Republic of Ireland outside the Commonwealth in 1949. As Beal (Reference Beal2004:10) notes, ‘the period between 1700 and 1945 saw the rise and fall of the British Empire, the American War of Independence [in 1776] […], and the rise of the USA as the most powerful nation on earth’. The late twentieth century saw political autonomy movements of Scotland and Wales that put a greater focus on Scots, Scottish Gaelic and Welsh (see Smith et al., Nance, and Willis, respectively, this volume).

Other important developments during the period were the Industrial Revolution (c. 1760–1830), including the revolution in transportation, as well as demographic developments like continued urbanisation. The effect of industrialisation led to labour-force mobility to the cities and towns. As the movement was more rapid than jobs available, an impoverished urban class developed that required support. London was particularly affected by the migration and therefore grew on a large scale during that period (cf. Hobsbawm Reference Hobsbawm1990:43; also see Lawton Reference Lawton, Langton and Morris1986). It is noteworthy that a large number of people resettled in the colonial territories, particularly North America and later the Southern Hemisphere. Industrialisation also saw the emergence of new centres in the North East (due to mining) and the North West/Midlands (due to textile manufacture and commerce) (cf. Beal Reference Beal2004:6–7). On a linguistic level, the movements also had an effect on the traditional rural dialects, notably dialect levelling and the development of new urban dialects. Nevertheless, general dialect distinctions, particularly in the North and the South, but also in the East and West Midlands, were retained (see for instance Wales Reference Wales2006; Kerswill Reference Kerswill and Wright2018). Other important developments, as previously already pointed out (see Section 1.1.1), such as the rise in literacy rates, in combination with the technological changes and the introduction of elementary compulsory schooling, had an effect on the publishing market and the development of mass media. Related to communications, another important development was the introduction of the Penny Post in 1840, which led to a steep increase in the sending of letters, both within Britain and to countries overseas. Further technical inventions and introductions concerned the electric telegraph in 1837 and the introduction of the telephone in 1876. Spoken communication was revolutionised through the invention of radio in 1895, as well as the establishment of the BBC in 1927 (Beal Reference Beal2004:9).

1.5.2 The Later Modern English 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.

1. (7)

   I think the speed in which we speak as well is – it’s a lot faster (DECTE/ Y07i006a/0665)

1. (8)

   I know people who have moved away but not who I’m really close with (DECTE/ Y10i024a/0057)

1. (9)

   that’s the lady that we’re doing this for (LIC/Mandy/278)

1. (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.

Data adapted from Hoffmann (Reference Hoffmann2005:81, Table 3.5).

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

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.

1. (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.

2.4.1 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:

1. (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.

Data adapted from Tagliamonte (Reference Tagliamonte1998:183).

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).

1. (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).

3.3.1 Stops

Probably the most widely studied variable in English phonology is (t), in particular with reference to its realisation as a glottal variant (Fabricius Reference Fabricius2002). Acoustic analysis shows that the percept of [ʔ] tends to be triggered by short periods of creaky phonation rather than full glottal closure (Docherty and Foulkes Reference Docherty, Foulkes, Hardcastle and Mackenzie Beck2005; Ashby and Przedlacka Reference Ashby and Przedlacka2014).

Glottal forms of (t) are common across much of Britain, although it is important to note that the frequency of [ʔ] is affected not only by social factors but by a complex set of linguistic factors, including phonological context, stress, word class, word frequency, and whether a word is repeated in discourse (Schleef Reference Schleef2013). Word-final pre-consonantal (t), as in get ready, is almost always produced as [ʔ] in England (a pattern dating back to the nineteenth century in Received Pronunciation (RP); Collins and Mees Reference Collins and Mees1996). In word-final prevocalic position (e.g. get off) glottals are now well established in rural as well as urban communities, displacing [t] (Marshall Reference Marshall2004). Most varieties tolerate glottals before syllabic nasals and/or laterals, and also in turn-final position (although in Newcastle a fully released oral [t] acts as a turn handover cue; Docherty et al. Reference Docherty, Foulkes, Milroy, Milroy and Walshaw1997). It is really in word-medial intervocalic position (e.g. water) where the ongoing spread is most noticeable and socially evaluated. In Derby, for example, speakers born before 1950 produced only 4 medial glottals in almost 1,000 tokens in unscripted conversation. By contrast, speakers born in the 1960s–70s produced 22 per cent glottals (Docherty and Foulkes Reference Docherty, Foulkes, Foulkes and Docherty1999). In the North East, the glottal stop is in fact replacing not [t], but a local form, [ʔ͜t] (a glottal stop, or glottal constriction, articulated with the [t]). For younger speakers across England, it appears that the glottal variant is now dominant (Smith and Holmes-Elliott Reference Smith and Holmes-Elliott2018). The use of [ʔ] remains a stigmatised feature for many self-appointed guardians of the language, although the stigma is undoubtedly waning fast for speakers themselves. Indeed, prevocalic [ʔ] is now found in RP (Fabricius Reference Fabricius2002), even in the speech of younger royals (Shaw and Foulkes Reference Shaw and Foulkes2015). Glottals can also be heard, but much less frequently, as realisations of /p/ and /k/, and occasionally of fricatives and /d/ (e.g. Tollfree Reference Tollfree, Foulkes and Docherty1999; Trudgill Reference Trudgill, Foulkes and Docherty1999a).

Deletion of /t/ and /d/ in word-final consonant clusters (e.g. last piece) is the norm in many accents, particularly in fast and casual speech. This phenomenon has been investigated in many US dialects (e.g. Guy Reference Guy and Labov1980), but few in the UK. Generally, deletion is most frequent in monomorphemes (e.g. mist), with lower rates in ‘semi-weak’ verbs (kept) and lowest in past-tense forms (missed). Deletion rates are also affected by preceding and following segments, and by word frequency (Baranowski and Turton Reference Baranowski and Turton2020). However, Tagliamonte and Temple (Reference Tagliamonte and Temple2005) found no morphological class effect in York.

A range of voiced variants of intervocalic (t) can be found across varieties, including [ɹ] and [ɾ]. The former, the ‘T-to-R rule’, is restricted to a small set of mostly grammatical words and word-final prevocalic position (e.g. what if, that is), but it also occurs medially in better and getting. This variant is mainly found in northern areas, is often stigmatised and seems to be receding (Docherty and Foulkes Reference Docherty, Foulkes, Foulkes and Docherty1999:51). The distribution of [ɾ] is less clear. It appears to be on the increase in Newcastle (cf. Watt and Milroy Reference Watt, Milroy, Foulkes and Docherty1999:29). Ashby and Przedlacka (Reference Ashby and Przedlacka2011) consider [ɾ] as one of several lenited variants in RP, reporting lenited forms to be the norm in non-initial contexts.

Innovative pre-aspirated variants have been reported in several British accents, including Newcastle. They occur mainly for pre-pausal (t) and are strongly associated with young women (Docherty and Foulkes Reference Docherty, Foulkes, Foulkes and Docherty1999). Clark and Watson (Reference Clark and Watson2016) discuss lenition of stops in Liverpool, and their spread to surrounding areas. Social or regional variation in voice onset time appears relatively rare. However, de-aspiration of voiceless stops occurs in Lancashire (Wells Reference Wells1982:370) and Bradford, particularly for Asian heritage speakers (Heselwood and McChrystal Reference Heselwood and McChrystal2000).

While realisational differences have received the most attention, dialects also differ in the connected speech processes they permit. Nolan and Kerswill (Reference Nolan, Kerswill and Ramsaran1990) found variable assimilation effects for schoolchildren in Cambridge. Kerswill and Wright (Reference Kerswill and Wright1990) showed that place assimilations common in standard varieties, such as red car [ɹeɡ kɑː], appear to a lesser extent in Durham. Parts of West Yorkshire, notably Huddersfield and Bradford, permit regressive voiceless assimilation, thus Hyde Park may be [haɪt̚ paːk] (or [haɪʔ paːk] via glottalling), although this feature seems to be receding (Whisker-Taylor and Clark Reference Whisker-Taylor and Clark2019). Regressive voicing assimilation is present in Durham, but not in RP (e.g. like[ɡ] bairns; Kerswill Reference Kerswill1987:44).

3.3.2 Fricatives

Two particularly striking changes are affecting the fricative system of English. The first concerns /h/ dropping, the deletion of phoneme /h/. In grammatical words (him, hers, etc.), /h/ dropping is common in all accents, but in content words (happy, help), it is more variable and socially marked. Many studies report /h/ dropping to be socially stigmatised, and, as is typical for such variables, more common in the speech of men and lower social classes (e.g. Baranowski and Turton Reference Baranowski, Turton and Hickey2015). Although /h/ dropping is cited by Wells (Reference Wells1982:60) as ‘perhaps the single most powerful sociolinguistic shibboleth in England’, it appears that its spread has been halted or even reversed (Cheshire et al. Reference Cheshire, Fox, Kerswill and Torgersen2008; Jansen Reference Jansen2021).

The second change concerns labiodental realisations of the dental fricatives (θ, ð). Use of [f, v] in, for example, think, mother is widely assumed to have originated in the South East but has now spread by geographical diffusion (Kerswill Reference Kerswill, Britain and Cheshire2003) such that it is present across the whole of England, as well as parts of Wales, Scotland and Northern Ireland (MacKenzie et al. Reference MacKenzie, Bailey and Turton2016). Initial (ð) in function words (the, they, etc.) is only occasionally affected but instead may be articulated as a dental or alveolar stop. Stops for (θ, ð) are also found in Liverpool (Watson Reference Watson2007, Clark and Watson Reference Clark and Watson2016). In Multicultural London English (MLE) (Cheshire et al. Reference Cheshire, Fox, Kerswill and Torgersen2008; Fox this volume), [d] for (ð) is frequent word-initially in function words, while [t] for (θ) is largely restricted to thing and words where ‑thing forms part of a pronoun (e.g. something). Drummond (Reference Drummond2018; this volume) reports the same features in social practices relating to grime music (hip-hop) among adolescents in Manchester.

The [ʍ] pronunciation in words such as which has all but disappeared from accents in England, having merged with [w] (or [h] before /u:/, e.g. who). Exceptions include conservative RP and parts of rural Northumberland (Wells Reference Wells1982:228ff.).

3.3.3 Nasals

Several studies have investigated variability in the realisation of (ŋ) in final unstressed syllables of polysyllabic words such as swimming (it is not variable in monomorphemes such as sing). The non-standard [n] realisation tends to outweigh [ŋ] in colloquial speech and is generally used more by male speakers from lower socio-economic groups (e.g. Trudgill Reference Trudgill1974). Tagliamonte (Reference Tagliamonte, Gunnarsson, Bergström, Eklund, Fidell, Hansen, Karstadt, Nordberg, Sundgren and Thelander2004) discusses language-internal constraints on variation in (ŋ) in York, showing that verbs favour [n] while nouns and adjectives favour [ŋ]. Another non-standard form, [ŋk], is widespread, but only in pronouns ending in ‑thing (e.g. something); here, forms with [n] are also common.

The ‘stopped’ variant [ŋɡ] occurs in the North West and Midlands (e.g. Bailey Reference Bailey2019). In fact, it appears to be emerging as a local prestige form in some areas (Schleef and Flynn Reference Schleef and Flynn2015). Bailey (Reference Bailey2019) finds [ŋɡ] to be conditioned by prosodic context. It is most frequent before pauses, where its rate is increasing in apparent time.

3.3.4 Liquids and Approximants

The phonetic quality and phonotactic distribution of (r) are both variable. The most widespread form used is the alveolar approximant [ɹ]. Other variants are restricted to particular phonological positions, regional dialects and/or certain speech styles. These include [ɾ], which is common in northern varieties (e.g. Jansen Reference Jansen, Jansen and Siebers2019a), especially in intervocalic position, but declining; trilled [r], which in England seems restricted to affected or theatrical styles; and the ‘Northumberland burr’, [ʁ], which is on the verge of extinction (Maguire Reference Maguire, Montgomery and Moore2017). A labial, or labiodental variant, usually transcribed [ʋ], has also long been recognised but has usually been dismissed as a feature of immature, defective or affected speech. In recent years, however, [ʋ] has become firmly established as an acceptable mature variant in many urban dialects. King and Ferragne (Reference King and Ferragne2020) explore /r/ variants via ultrasound and lip cameras, reporting a continuum of tongue postures from bunched to retroflex. Lip protrusion is also common but variable between speakers, and different from the articulatory posture for [w]. Variability in phonotactic distribution primarily concerns whether or not (r) is articulated in postvocalic positions, a division encapsulated in the descriptive labels rhotic vs. non-rhotic. While rhoticity is generally prestigious in North America, the reverse is true for England. Rhoticity is much less widespread now than in the SED data (Leemann et al. Reference Leemann, Kolly and Britain2018:12), but it is still in evidence for some speakers in East Lancashire (Barras Reference Barras, Braber and Jansen2018) and the South West (Blaxter et al. Reference Blaxter, Beeching, Coates, Murphy and Robinson2019). Rhoticity tends not to be simply present or absent for individuals but is variable as a function of factors such as phonological context and word frequency.

Variation in (l) centres on phonetic quality. RP is traditionally described as having ‘clear’ (alveolar) [l] in syllable onsets and ‘dark’ (velarised or pharyngealised) [ɫ] in codas. In acoustic terms, clear [l] has high F2 and low F1, while for [ɫ] both F1 and F2 are low. However, the distinction does not always hold in other varieties, or the acoustic distinction is less marked (Kirkham, Turton and Leemann Reference Kirkham, Turton and Leemann2020). Clear [l] is found in all positions in the North East (Watt and Milroy Reference Watt, Milroy, Foulkes and Docherty1999:31) while universal dark [ɫ] is reported in Yorkshire (Kirkham Reference Kirkham2017). British Asian speakers appear to have clearer variants than ‘Anglo’ counterparts in the same locale (e.g. Kirkham Reference Kirkham2017). Ultrasound has been used in several analyses of /l/, revealing gradient as well as categorical patterns of articulation (Turton Reference Turton2017).

Where coda (l) is dark, delateralisation may occur, resulting in a variety of vowel variants (Wells Reference Wells1982:258ff). Vocalisation seems to be spreading across the South East even though it is usually thought to be stigmatised.

In most accents the approximant (j) is variably deleted in consonant clusters with a preceding coronal segment (e.g. new; Wells Reference Wells1982:206ff), although it is often categorically absent in Norfolk in words where it is obligatory elsewhere, such as beauty or view (Trudgill Reference Trudgill, Foulkes and Docherty1999a:133). Among many young speakers, alveolar stop + (j) sequences are assimilating into affricates [tʃ] and [dʒ], thus tune is frequently [tʃʉːn]. Likewise, in triconsonantal clusters, initial (s) may also palatalise through assimilation, e.g. stupid [ʃtʃʉːpɪd] (Bailey et al. Reference Bailey, Nichols, Turton and Baranowski2022)

3.5.1 Vocal Setting and Voice Quality

Vocal setting is defined by Laver (Reference Laver1994:396) as the ‘tendency underlying the production of the chain of segments in speech towards maintaining a particular configuration or state of the vocal apparatus’. Speakers who tend to keep the velum lowered, for instance, will possess a habitually nasal voice quality. Extensive studies of the phonetic correlates of vocal settings can be found in Laver (Reference Laver1980), Nolan (Reference Nolan1983) and Esling et al. (Reference Esling, Moisik, Benner and Crevier-Buchman2019).

As already noted, there has been relatively little systematic study of the regional, social or stylistic correlates of vocal settings, although impressionistic comments on voice quality are more widespread than those on supra-laryngeal settings. For instance, Knowles (Reference Knowles and Trudgill1978) comments on the velarised voice quality and raised larynx setting used in Liverpool. San Segundo et al. (Reference San Segundo, Foulkes, French, Harrison, Hughes and Kavanagh2019) conducted detailed perceptual analysis of 100 Standard Southern British English (SSBE) speakers. The most frequently encountered settings were fronted tongue body (98%), nasality (92%), creaky phonation (83%) and breathy phonation (73%). See also Stuart-Smith’s (Reference Stuart-Smith, Foulkes and Docherty1999) study of vocal settings in Glasgow.

Creaky phonation has been associated with RP (e.g. Laver Reference Laver1980:4), especially in men. In more recent years, however, habitual creaky voice has more typically been portrayed, especially in the media, as a stylistic feature of the speech of young women in varieties of UK, US and Australian English, although the evidence base in support of this portrayal is far from robust (Dallaston and Docherty Reference Dallaston and Docherty2020).

3.5.2 Rhythm

The rhythmic pattern of speech is a perceptual effect ‘produced by the interaction in time of the relative prominence of stressed and unstressed syllables’ (Laver Reference Laver1994:152). Languages are often said to be either syllable-timed or stress-timed, although in reality, these constitute two poles of a complex continuum (Arvaniti Reference Arvaniti, Setter and Knight2021). In syllable-timed languages (e.g. French), syllables tend to be more regular in duration than in languages considered to be stress-timed (e.g. English). In the latter, a relatively regular interval between each stressed syllable is considered to be a key feature, while unstressed syllables are compressed in between.

These differences in rhythmic organisation may also occur across different dialects of a language. Syllable timing is suggested as a feature of Bradford Panjabi English (Heselwood and McChrystal Reference Heselwood and McChrystal2000) and non-Anglo speakers of Multicultural London English (MLE) (Torgersen and Szakay Reference Torgersen and Szakay2012). Trudgill (Reference Trudgill, Foulkes and Docherty1999a:124) notes the characteristic rhythm of Norwich and Norfolk speech, involving very long stressed syllables and very short unstressed ones.

3.5.3 Stress and Intonation

Although word stress in English is not wholly fixed, dialects differ relatively little in the location of main stress placement. This is true both for word stress and sentence stress. Thus, words like acquisition and telephone almost invariably take stress on the indicated syllable, no matter what the accent. Counter-examples can be found, which often appear to reflect differences between British and North American usage. North American dialects tend to opt for initial word- and phrase-stress (ice cream, cigarette, inquiry), some of which have filtered into British usage. By contrast, traditional initial stress is being lost in many words, including kilometre and controversy (Wells Reference Wells1999).

Differences may occur in the type of pitch pattern used to represent the main stress within a sentence. RP, for instance, tends to use a falling pitch pattern in declarative sentences, and a pitch rise for questions. Other accents differ: rising intonation is traditionally associated with declaratives in Birmingham, Liverpool and Newcastle.

One of the most noticeable innovations in recent years has been the development of rising intonation in statements. This has been found in the USA, Australia and New Zealand as well as in UK varieties, and is usually labelled high rising tone or uptalk (Warren Reference Warren2016; Levon Reference Levon2020). Uptalk serves to track the listener’s comprehension and attention, especially when the speaker is presenting new information. It also acts as a turn-holding mechanism in narratives. What also emerges in almost all studies, however, is a high level of intra- and inter-speaker variability in intonation patterns (e.g. Grabe et al. Reference Grabe, Kochanski, Coleman, Dziubalska-Kolaczyk and Przedlacka2005; Nance et al. Reference Nance, Kirkham, Groarke, Braber and Jansen2018). This highlights the challenges of characterising general patterns of form–function correspondence and how they vary across accents.

3.6.1 Dialect Levelling and Diffusion

The Milroys’ study of language and social networks in Belfast (Milroy Reference Milroy1987) shows how close-knit communities resist language change, while more open communities and more mobile individuals encourage the spread of linguistic features through face-to-face contacts. Britain (Reference Britain, Watt and Llamas2014) refines the network hypothesis, drawing on Giddens’s (Reference Giddens1984) theory of routinisation as a factor in the perpetuation and maintenance of social structures. Routine activities serve to promote the maintenance of patterns of behaviour, including language use. Social and political changes may cause breaks in routine, opening the window for changes in behaviour. Over recent years, social upheavals in Britain have come to cause major disruptions to patterns of routine. There has been a dramatic decline of many traditional industries and a fragmentation of the communities which they supported, a marked increase in leisure opportunities and a massive expansion in the use of electronic media. Concomitant with these developments has been a rise in geographical and social mobility – though it remains to be seen what, if any, the longer-term effects of the 2020–2022 Covid-19 pandemic will be on mobility and language.

The principal linguistic effect of these social changes is homogenisation. Sociolinguists have generally described homogenisation processes as dialect (or accent) levelling (e.g. Watt and Milroy Reference Watt, Milroy, Foulkes and Docherty1999; Williams and Kerswill Reference Williams, Kerswill, Foulkes and Docherty1999; Kerswill Reference Kerswill, Britain and Cheshire2003; Torgersen and Kerswill Reference Torgersen and Kerswill2004; Britain Reference Britain, Llamas and Watt2010a; Schleef Reference Schleef2013; Butcher Reference Butcher2020). The overall phenomenon is better seen as one of supralocalisation (also known as regional dialect levelling), since there are a number of distinct but overlapping processes involved, and these are at a level greater than the local (cf. Britain Reference Britain, Llamas and Watt2010a). ‘Levelling’ in a stricter sense refers to the loss of minority or regionally restricted forms found in a given area in favour of variants shared by the majority of speakers there, the mechanism being linguistic accommodation between individuals involved in chains of contact (Kerswill Reference Kerswill and Hickey2018). The most intense levelling occurs in new communities, such as the New Town of Milton Keynes, where speakers of different dialects migrate to a single location (Kerswill and Williams Reference Kerswill and Williams2000, Reference Kerswill and Williams2005). Levelling occurs on a wider scale, too, with the spreading variants already well established regionally or nationally. An example is the glottal stop [ʔ], which, as we saw earlier, is replacing [t] intervocalically. In north-eastern accents, it is replacing the highly localised [ʔ͜t], bringing the region into line with the majority. Levelling is also not a new phenomenon, as witnessed for instance by research into dialect contact in the Fens resulting from migrations in the seventeenth century (Britain Reference Britain1997, Reference Britain and Hickey2010b). Levelling appears to be on the increase, as indicated by the abundance of evidence of the erosion of the traditional, local dialects (see for example Trudgill Reference Trudgill1999b; Britain Reference Britain and Hickey2012; Maguire Reference Maguire, Montgomery and Moore2017; chapters in Braber and Jansen Reference Braber and Robinson2018).

The process of dialect/accent levelling is more complex than the account above suggests, which further justifies the cover term supralocalisation. The first question to ask is whether the process takes place simultaneously right across the area affected, with no directionality, or if the changes come in from a particular direction, originating, say, in a big city. This latter process is known as geographical diffusion – and it may be that the incoming change does not in fact lead to levelling but simply the replacement of an old form with a new one without any reduction in the number of forms (Britain Reference Britain, Llamas and Watt2010a:195). The second question is whether the variant that is gaining ground is already present but is ousting others, perhaps because it is in a majority, or whether it is a new feature to the area, either arriving from outside or generated within the area (in the latter case, we can say it is endogenous; Trudgill Reference Trudgill, Foulkes and Docherty1999a). The third question is to consider whether there is a role for standardisation and prestige, not only for middle-class speakers but also working-class people, who are generally expected to speak in the most local way (cf. the SED’s earlier focus on ‘NORMS’): often, the outcome of a levelling process is the introduction of a variant that is recognisably closer to a normative, standard form, displacing a local, less standard form. As we will show, it is often difficult to tell which process, or combination of processes, applies.

It is not a given that levelling (or, better, supralocalisation) leads to the development of new, entirely homogeneous regional varieties (such as the North East or South East) which are clearly demarcated from other regional varieties. While phonetic (as well as lexical and grammatical) features cover increasingly large areas (Trudgill Reference Trudgill and Trudgill2001; Strycharczuk et al. Reference Strycharczuk, Brown, Leemann, Britain, Calhoun, Escudero, Tabain and Warren2019), it is better to see it as an ongoing process taking place at different speeds in different places, and not affecting all features at the same time. As we shall see, geographically defined varieties remain on a continuum across the country today as in the past; the change is that the differences between nearby varieties are being lessened.

In what follows, we cover four geographical areas in which accent levelling/supralocalisation has been described in detail. It is probable that similar phenomena are present elsewhere, notably the Midlands and the South West.

3.6.2 The South East

The best-known example of levelling (to the lay public) in England is centred on London, encapsulated in a variety popularly labelled ‘Estuary English’ (Rosewarne Reference Rosewarne1994a, b; Przedlacka Reference Przedlacka2002; Altendorf Reference Altendorf, Bergs and Brinton2017). The estuary referred to is that of the Thames, but the variety is said to be spreading across a much wider region. However, from the linguistic descriptions it is clear that it is a mix of optional features that is too heterogeneous to qualify as a ‘variety’ (Altendorf Reference Altendorf, Bergs and Brinton2017:179).

However, corroborating this idea of a new, levelled variety in the South East is the convergence of vowel systems across the region, as shown in comparative research in Reading, Milton Keynes and Ashford (Williams and Kerswill Reference Williams, Kerswill, Foulkes and Docherty1999; Kerswill and Williams Reference Kerswill and Williams2000, Reference Kerswill and Williams2005; Torgersen and Kerswill Reference Torgersen and Kerswill2004). These three towns are similar in size and lie some 50–70 km to the west, north-west and south-east of London, respectively. By comparing SED data and older and younger contemporary data from these towns, Torgersen and Kerswill (Reference Torgersen and Kerswill2004) showed that there was a convergence in the short-vowel system on formant values found in modern London vernacular (‘Cockney’), from quite disparate starting points. The diphthongs of price and mouth also showed convergence in Reading and Milton Keynes (Ashford was not investigated; Williams and Kerswill Reference Williams, Kerswill, Foulkes and Docherty1999). This time, the outcome was similar to London vernacular only for price in Milton Keynes ([ɑɪ]), with little generational change in Reading, which largely maintained [ɔɪ]. For mouth, both towns converged not on London vernacular [æː^ʊ], but closer to RP [aʊ].

A later study of a town much closer to London, Debden (Essex), compares this town to Reading and Milton Keynes (Cole and Strycharczuk Reference Cole and Strycharczuk2022). Debden contains a large estate newly constructed in 1947–52, and, as Milton Keynes would do fifteen years later, received a large number of people from London. The levelling pattern for price and mouth is broadly similar, though it has been slower, probably because of the strong London identity of its older residents. However, the authors note that the vowel changes are in the direction of what they call an unmarked ‘pan-regional standard’, which shares features with RP but not its strong elite associations. This, they argue, is the target for the levelling changes in Milton Keynes, Reading and Debden, and not ‘exclusively as a result of dialect contact or face-to-face interaction’ (Cole and Strycharczuk Reference Cole and Strycharczuk2022). This is a clear example of accommodation towards an external, socially more neutral, non-local model, not occasioned by direct contact.

3.6.3 The North East

It is unlikely that levelling ever leads to complete homogeneity. In the north of England, the main pattern, as in the London area, is the spread of features of a city’s accent to its hinterland, though, again, it is difficult to distinguish diffusion from supralocalisation. We can take the urban North East as an example, an area stretching from the major conurbation of Tyneside (including Newcastle upon Tyne) in the north to the smaller town of Middlesbrough, 65 km to the south. Intervocalically, /p/ is often realised as [ʔ͜p] in most parts of the region, particularly in the north. In Newcastle, it is pronounced this way by a majority of speakers, with males being almost categorical and females lying at around 58 per cent (Docherty et al. Reference Docherty, Foulkes, Milroy, Milroy and Walshaw1997:301; Beal et al. Reference Beal, Burbano Elizondo and Llamas2012:37–8). In Middlesbrough, older and younger men use the feature in a proportion similar to that of their Newcastle counterparts, while older females rarely use it. Younger females, however, have partly caught up with males, with proportions approaching those of Newcastle. At first blush, this looks like a case of geographical diffusion from Newcastle, with men (presumably) adopting the localised form before women. However, data from Sunderland, which lies between Newcastle and Middlesbrough, suggests otherwise: while women use [ʔ͜p] less often than men, its overall usage is much lower than in the other two locations, at less than 25 per cent (Beal et al. Reference Beal, Burbano Elizondo and Llamas2012:38).

If there is any ‘gravitational’ pull in the North East towards Newcastle, it is not consistent across features. Local pronunciations of the face and goat vowels are the centring diphthongs [ɪə] and [ʊə], respectively. These are being replaced, apparently more or less simultaneously across the region, by monophthongs which can be broadly represented as [eː] and [oː] in a still-ongoing process (Watt Reference Watt2002; Kerswill Reference Kerswill1984, Reference Kerswill, Britain and Cheshire2003; Beal et al. Reference Beal, Burbano Elizondo and Llamas2012). The lack of directionality suggests supralocalisation rather than diffusion from one of the urban centres. At the same time, a local Tyneside monophthong, a central or fronted variant of goat, [ɵː] or [øː], which is local to Tyneside, is not spreading, but is disappearing (Beal et al. Reference Beal, Burbano Elizondo and Llamas2012:31). This suggests the levelling out of forms which are geographically restricted, either at the regional level ([ɪə] and [ʊə]) or at the local level ([ɵː]/[øː]). More strikingly, while in the north of the region (Tyneside, Sunderland and Durham) the vowels of face and goat have the raised, half-close pronunciations [eː] and [oː], in Middlesbrough they are more open, typically in the region of [ɛ̈ː] and [ɔ̈ː] – both being centralised (Beal et al. Reference Beal, Burbano Elizondo and Llamas2012:30–1, 44–5). Unlike the case of the shift from centring diphthongs to monophthongs for face and goat, there appears to be no generational change involving the height of the monophthongs. This fact weakens the argument that there is levelling at the regional level.

Findings from the SED shed some light on the situation. For several face words, Middlesbrough can be seen to lie at the northern extreme of a large area stretching south through Yorkshire and Lincolnshire where most words containing face have a lowered diphthongal variant [ɛə] (viz. grave, bacon and April; Orton, Sanderson and Widdowson Reference Orton, Sanderson and Widdowson1978:Ph60–Ph65).Footnote ¹ However, some other words (including spade and naked) contain the same raised centring diphthong [iə] or [ia] as the north of the region (including Newcastle) in an area again stretching south through Yorkshire, but stopping at the border with Lincolnshire. Given that today the southern portion of the North East contains the more open realisation in all face items, it seems likely that this is the result of a levelling process involving the transfer of [iə] or [ia] to [ɛː]. Orton et al. (Reference Orton, Sanderson and Widdowson1978: Introduction [no page nos.]) state that [iə] or [ia] in spade and naked are ‘traditional Northern forms’ descended from Middle English ā, while these are absent for grave, bacon and April. We can interpret this as being an example of lexical diffusion – a phonological change gradually running through the lexicon, in this case now complete. This seems to be a different process from the more general and probably more recent monophthongisation of the whole face set in Newcastle and Durham. The outcome represented in the SED for the southern portion is also consistent with standardisation (see next section), by which what appear to be two lexical sets ([ɛə] and [iə]/[ia]) are merged into one set, corresponding to Standard English face. This would then have been monophthongised to present-day [ɛ̈ː], the mainstream northern form.

Further differences within the region remain. In Tyneside, the final unstressed vowel in the letter/comma sets, such as in paper or Emma, is likely to be an open [ɐ], while in Sunderland, Durham and Middlesbrough the quality is central or fronted to [ɛ]. There appears to be no recent change in this differentiation (Beal et al. Reference Beal, Burbano Elizondo and Llamas2012:36; Watt et al. Reference Watt, Llamas, French, Braun and Robertson2016).

For the North East we are left, then, with a complex picture. Levelling on a regional level is not consistent: while the north-eastern glottalling of /p/ may well be spreading out from Newcastle, albeit inconsistently, the monophthongisation of face and goat is taking place simultaneously across the region rather than diffusing from one population centre; for both vowels, it can be seen as levelling towards a general north-of-England monophthongal norm. On the other hand, there is no recent convergence in vowel height at a North East regional level; instead, the lowered face is continuous with the large region to the south. As we have seen, the past history of face appears to differ between the north and the south of the region, with the lexicon in the south split into two apparent lexical sets which have now been merged back into one – and this suggests an earlier phase of levelling at a regional level. (This is true to some extent for goat, too, though for this vowel the centralisation characteristic of much of Yorkshire is largely absent; Haddican et al. Reference Haddican, Foulkes, Hughes and Richards2013.) With the letter/comma vowel, there is similarly no recent change, but the boundary between the northern [ɐ] and southern [ɛ] lies just south of Newcastle, well to the north of that for raised vs. lowered monophthongal face and goat.

Let us attempt some explanations. For monophthongisation, we can note Watt’s (Reference Watt1998:7) observation that people want to avoid the old-fashioned ‘cloth cap and clogs’ image of a local northern identity, and so adopt more widespread forms, such as monophthongal face and goat. With respect to the increased use in Middlesbrough of glottalised /p/, Llamas (Reference Llamas2007) argues that speakers’ sense of a North East regional identity plays a crucial role, having been made more acute by administrative boundary changes, with Middlesbrough coming into the ‘orbit’ of the North East and away from its traditional association with Yorkshire. As for the height differences in face and goat, explanations such as these may be harder to come by: research by Llamas et al. (Reference Llamas, Watt, French, Braun and Robertson2019) and Watt et al. (Reference Watt, Llamas, French, Braun, Robertson and Kendall2019) suggests a role for perception at a local level. They examine the perception of accent differences across the North East, finding consistent differences in listeners’ awareness of the geographical distribution of the localised variants of vowels, depending on which town/city they came from. There were also differences in where they perceived phoneme boundaries to lie depending on what variety they used.

At the same time, the notion of standardisation (Section 3.7) comes into play in a way similar to developments already noted in the South East: the monophthongal variants were already in use by middle-class speakers in the North East. The change towards monophthongs was female-led, following a general pattern by which widespread and/or standard forms are favoured by them (Eckert Reference Eckert1989; Milroy et al. Reference Milroy, Milroy, Hartley and Walshaw1994; Queen Reference Queen, Chambers and Schilling-Estes2013).

3.6.4 Liverpool

A rather different pattern applies in the Liverpool area, in the north-west of England. Liverpool’s accent is highly distinctive in that it uses fricatives/affricates for intervocalic /k/ and /t/, giving pronunciations such as [wɛːxən] and [lɛt^sə] for working and letter (Watson and Clark Reference Watson, Clark and Hickey2017). Clark and Watson (Reference Clark and Watson2016) investigate the pronunciation of /t/ in word-final position, where the main alternatives are [θ̳] (sounding somewhat like [s]) and [h]. The latter may occur in a set of high-frequency words with a preceding short vowel, as in get, got, bit, what, that, it and not. For this variant, there is ‘little evidence’ of levelling in favour of the widespread, supralocal [ʔ] found in most other accents (Clark and Watson Reference Clark and Watson2016:44). In addition, this feature has spread to nearby Skelmersdale, where it had barely existed until the generation born around 1945. Skelmersdale, then, has not taken part in the supralocal dialect levelling found across the country – at least not with respect to this feature. This contrasts with St Helens, which, despite being roughly the same distance from Liverpool, has not adopted this Liverpool feature, nor indeed several others. Perhaps even more striking is the spread of the [h] variant to a new phonological environment, that of di- or polysyllabic words ending in an unstressed syllable, such as market and certificate – a change which has spread to Skelmersdale, too (Clark and Watson Reference Clark and Watson2016:36). This is an example of the divergence of a local accent away from widespread norms. The new feature has spread to at least one other town, but there are clear geographical limits to the spread.

3.6.5 Manchester

The third northern conurbation we discuss in this chapter, Greater Manchester, again illustrates levelling within its region, while maintaining and even making a phonetic boundary sharper. In much of the north-west of England, the vowels of nurse and square are merged, resulting in a vowel whose realisation lies on a continuum from central [əː] (most of Lancashire, parts of Greater Manchester) and a front [ɛː] (Liverpool) (Watson and Clark Reference Watson and Clark2013; Baranowski and Turton Reference Baranowski, Turton and Hickey2015). In Manchester itself, however, there is no merger. This means that there is likely to be a pull in two directions – towards the merged west (Lancashire and Liverpool) and the rest of the north, with no merger. Research based on an extensive online dialect survey using university students and their networks (MacKenzie et al. Reference MacKenzie, Bailey and Turton2016, Reference MacKenzie, Bailey and Turton2022) reveals a change in the boundaries of the merger during the last fifty years. This survey asked, among other variables, whether or not the respondent claimed to make a distinction between the vowels in fur and bear. The results were displayed in terms of postcode areas for the whole of Britain, but here we focus on the Liverpool–Manchester area. This is shown in Figure 3.1, with the boundary of Greater Manchester indicated by a heavy line. Lighter areas within the area under discussion show there is a merger, darker areas the absence of a merger, for older and younger respondents. The lighter the shade, the closer to categorical the merger is. (Regions outside the Liverpool–Manchester area are non-merging; they are shown in pale shading.) From our point of view, two points are of interest. The first is that both the merging and non-merging areas are becoming more cohesive, with fewer outlying merged (light-coloured) areas in the north and east of Manchester. The second is that both Liverpool and Manchester have become clearer focal points for mergers and non-mergers, respectively. Lighter shaded areas are particularly noticeable around Liverpool, even though the merged area itself is contracting.

Figure 3.1 The nurse~square merger in postcode areas in the north-west of England for older and younger speakers. Merged areas are shown in lighter shading, non-merged in dark shading. Shading represents z-scores. Regions outside the Liverpool–Manchester area are not under consideration. (See text for details.).

Maps courtesy of Danielle Turton and George Bailey, based on data from the Our Dialects project, www.ourdialects.uk

The Liverpool–Manchester data show the linguistic effect of two significant dialect areas in close proximity: in this case, there is only a weak tendency for the influence of one, Liverpool, to be reduced, while the feature itself is strengthened within its own area. An explanation for the loss of the merged areas on Manchester’s periphery may well be found in increased contacts with the city through commuting, commercial operations and leisure. Furthermore, the spread of non-mergers is relatively rare and is cognitively complex, and depends on high levels of contact with non-merged speakers.

4.2.1 Verbs

4.2.2 Negation

4.2.3 Adverbs

Inflectionless forms of adverbs, as in (49) and (50), are attested right across the country (Hughes and Trudgill Reference Hughes and Trudgill1979, Cheshire Reference Cheshire1982:80). Their vitality, however, appears to depend both on where they are used and on the adverb in question. Tagliamonte and Ito (Reference Tagliamonte and Ito2002) showed a decline in York in the use of forms without ‑ly, but this was almost entirely because the use of intensifier real, as opposed to really, had fallen. For other adverbs the decline in the use of ‑ly-marked adverbs was present, but less dramatic (Reference Tagliamonte and Ito2002:252–3), and young less well-educated speakers still retain inflectionless forms over 25 per cent of the time (Tagliamonte Reference Tagliamonte2018:118). For other locations in (largely rural) England, Tagliamonte (Reference Tagliamonte2018:127) finds higher levels of inflectionless forms: over 40 per cent in Maryport in Cumbria, and over 50 per cent in Wheatley Hill near Durham and Tiverton in Devon. Across England, frequently occurring adverbs, such as real(ly), near(ly) and quick(ly) are more likely than less common adverbs to have ‑ly (Reference Tagliamonte2018:129).

1. (49)

   It does come round quick, you know (Tagliamonte and Ito Reference Tagliamonte and Ito2002:237, York)

1. (50)

   They aren’t nice up at that physio, they treat you bad (Tagliamonte Reference Tagliamonte2018:115, Tiverton, Devon)

A number of other researchers have specifically focused on variation in adverbial intensification of the kind in (51–54). Hudson and Holmes (Reference Hudson and Holmes1995:14) pointed to Liverpool as the focal site of the use of intensifier dead. Barnfield and Buchstaller (Reference Barnfield and Buchstaller2010:267) show, for Newcastle, that both dead and real(ly) are rapidly becoming popular intensifiers, while very is declining rapidly in use. Palacios Martínez and Núñez Pertejo (Reference Palacios Martínez and Núñez Pertejo2012:780) also find that very is declining in London, with really and so on the increase, but dead and well barely used at all. They also point to the use of other forms among their younger speakers, such as uber, mega and super. Stenström, Andersen and Hasund (Reference Stenström, Andersen and Hasund2002:151) show that real, unlike in York, as well as right and well are favoured by middle-class youngsters in London.

1. (51)

   he was very young (Coggeshall, Essex)

1. (52)

   I know when I go to the gym and I see my six pack I feel so proud (Norwich, Norfolk)

1. (53)

   you’re dead close to the beach (Barnfield and Buchstaller Reference Barnfield and Buchstaller2010:257, Newcastle)

1. (54)

   it was bloody peaceful while you were away (Palacios Martínez and Núñez Pertejo Reference Palacios Martínez and Núñez Pertejo2012:783, London)

4.2.4 Prepositions

Variability in prepositional use seems to take a number of forms:

1. 1. The use of a complex preposition where the standard has a simple one, as in (55) (see Cheshire et al. Reference Cheshire, Edwards, Whittle, Milroy and Milroy1993:77; Braber and Robinson Reference Braber and Robinson2018:101)

2. 2. The use of a simple preposition where the standard uses a complex one, as in (56) (see Cheshire et al. Reference Cheshire, Edwards, Whittle, Milroy and Milroy1993:77).

We know little about the regional distribution of either of these.

1. 3. The omission and reduction of to. Watts (Reference Watts2006:322) finds to often totally omitted after go by working-class speakers in Cheshire, as in (57). Shorrocks (Reference Shorrocks1999) reports the reduction of to to some form of glottal stricture or devoicing in Manchester and Lancashire (see also Braber and Robinson Reference Braber and Robinson2018; Ojanen Reference Ojanen1982:252; Vasko Reference Vasko2005:168–74). Stevenson’s analysis of tweets – http://nwdialectatlas.uk/ – also finds a concentration of prepositionless forms in Cheshire and around Liverpool.

2. 4. A range of non-standard usages of particular prepositions (Shorrocks Reference Shorrocks1999; Clark and Asprey Reference Clark and Asprey2013). Vasko (Reference Vasko2005), for example, examines in detail the use of up in Cambridgeshire. Willis et al.’s Tweetolectology project, meanwhile, notes that the use of down in ‘down the pub’ (versus ‘at the pub’ or ‘in the pub’) is most often found in East Anglia and the South East, but is barely found at all in the North.Footnote ⁴

1. (55)

   We’d been getting a lot of stick off of the Manchester United fans (Braber and Robinson Reference Braber and Robinson2018:101, Leicester)

1. (56)

   If you went down London (Braber and Robinson Reference Braber and Robinson2018:101, Mansfield, Nottinghamshire)

1. (57)

   My dad needs to go the opticians (Watts Reference Watts2006:323, Wilmslow, Cheshire).

4.2.5 Plural Marking

Many non-standard varieties do not mark plurality overtly on a number of measurement nouns (see Cheshire et al. Reference Cheshire, Edwards and Whittle1989; Hughes and Trudgill Reference Hughes and Trudgill1979, etc.). Examples can be found in (58–62). There has been virtually no empirical research, however, to examine the vitality of this phenomenon, or, for example, the extent to which it has expanded its scope to metric measurements.

1. (58)

   We were going probably like fifty mile an hour (Norwich, Norfolk)

1. (59)

   I want a piece four foot by eighteen inches (Colchester, Essex)

1. (60)

   He spends a hundred pound a week on taxis (Brentwood, Essex)

1. (61)

   It was a lovely old car. I had it for six year, never gave me any trouble (Holbrook, Suffolk)

1. (62)

   What’s happened in my life in the last six to eight month? (Colchester, Essex)

From inner London, Cheshire (Reference Cheshire2013) reports a set of new plural forms of man, as in (63) and (64).

1. (63)

   I wanna be with the mandem, innit, your friends who you grew up with (Cheshire Reference Cheshire2013:616, London)

1. (64)

   If you put all the mans together, make them fight a cause together (Cheshire Reference Cheshire2013:616, London)

4.2.6 Pronouns

4.2.6.5 Pronominal Word Order

In clauses with both a direct and indirect object pronoun, geographically marked word order variability exists, with (91), (92) and (93) all widely attested.

1. (91)

   She gave it to me

   Verb + Direct Object (DO) to Indirect Object (IO)

1. (92)

   She gave me it

   Verb + IO + DO

1. (93)

   She gave it me

   Verb + DO + IO

Drawing from data in the Survey of English Dialects, Kirk (Reference Kirk, Kirk, Sanderson and Widdowson1985:135) shows that Verb + DO to IO (as in (91)) was reported as the majority form only in the English South West, whereas Verb+DO+IO (as in (93)) was dominant in the Midlands and Lancashire and Verb + IO + DO (as in (92)) most common in the far North and North East, London and East Anglia.

There has been considerable exploration of this geographical variability using internet and social media platformed surveys. Figure 4.1 shows the geographical reach and use of Verb+DO+IO in the results of the 2016 smartphone-based English Dialects App (EDA) survey of over 50,000 users (Leemann et al. Reference Leemann, Kolly and Britain2018). The EDA found this to be the majority form in East Lancashire, Blackpool, Blackburn, Bolton, St Helens, South Yorkshire, North Derbyshire, Derby, West Nottinghamshire, North Staffordshire, Stoke-on-Trent and parts of the Black Country (especially Dudley and Walsall). Verb+IO+DO was the majority form in Liverpool, West Yorkshire, especially Leeds, and the East coast from Northumberland down to North Lincolnshire. While this form was common in East Anglia at the time of the Survey of English Dialects, it was not reported at high levels from there in the 2016 survey. Elsewhere Verb+DO to IO was reported as the majority form.

Figure 4.1 The distribution of ‘give it me’ (Verb+DO+IO) in the English Dialects App.

MacKenzie et al. (Reference MacKenzie, Bailey and Turton2014) examined variability through an internet survey of over 8,000 responses, which asked users the extent to which they find the Verb+DO+IO – give it me – form ‘acceptable’. The area of acceptability is extremely similar to the pattern found in the EDA. The same pattern is also found in Stevenson’s (Reference Stevenson2016) investigations of this feature in Twitter feed. The geographical spread of this form, then, appears relatively resilient, despite it being the formally non-standard variant of the three (see also Siewierska and Hollmann Reference Siewierska, Hollmann, Hannay and Steen2007). The Verb+IO+DO form, on the other hand, appears to have undergone some geographical shrinkage since the Survey of English Dialects. Stevenson’s Twitter analysis (http://nwdialectatlas.uk/) matches that of the EDA and shows that it remains the majority form in Liverpool, Leeds and the far North and North East, but is only now found at modest levels – 15–30 per cent – in the South and South East.

4.2.7 Comparison

A number of varieties report ‘double comparison’ forms, using both the appropriate analytical marker (more or most) and the inflectional ending (‑er for comparatives and ‑est for superlatives), as in (102) and (103) (e.g. Hudson and Holmes Reference Hudson and Holmes1995:20; Stenström et al. Reference Stenström, Andersen and Hasund2002:134). There appears to have been little empirical investigation of their contemporary distribution and use, however.

1. (102)

   no I’d rather have the bigger planes, they seem more safer (Nordelph, Norfolk)

1. (103)

   I think she was most prettiest when she was about eighteen (Colchester, Essex)

4.2.8 Definite and Indefinite Articles

Commonly reported across many parts of the North of England is Definite Article Reduction, whereby the is reduced to [t] or [ʔ] (Mark Jones Reference Jones2002; Lodge Reference Lodge2010; Rupp and Page-Verhoeff Reference Rupp and Page-Verhoeff2005; Rupp Reference Rupp2008; Tagliamonte and Roeder Reference Tagliamonte and Roeder2009), as in (104). Stevenson’s analysis of tweets (http://nwdialectatlas.uk/) shows concentrations of this in Leeds, Sheffield and Hull. Rupp and Tagliamonte (Reference Rupp and Tagliamonte2019a) show that in York the definite article can be totally absent, especially in discourse-new, hearer-old contexts, in other words where the noun had not been referred to recently in the conversation but is known to both participants in the conversation (see also Braber and Robinson Reference Braber and Robinson2018:79).

1. (104)

   They had a baby, and as soon as t’baby arrived he got jealous (Rupp and Page-Verhoeff Reference Rupp and Page-Verhoeff2005:326, Quernmore, Lancashire).

The formal spoken standard demonstrates allomorphy in the article system, sensitive to the following phonological environment: the [ðə] and a [ə] before consonants, and the [ði] and an [ən] before vowels. There is sporadic evidence from a number of traditional dialects that this allomorphy is not always present (Britain Reference Britain, Britain and Cheshire2003:203, Ojanen Reference Ojanen1982:126 and Peitsara Reference Peitsara, Klemola, Kytö and Rissanen1996:288 for the East, Piercy Reference Piercy2010 and Wagner Reference Wagner, Kortmann, Burridge, Mesthrie, Schneider and Upton2004:155 for the South West, etc) (see (105)). Fox provides detailed and dramatic empirical evidence of this erosion of allomorphy – [ðə] and [ə] before vowels – in Multicultural London English (see Fox Reference Fox2015; Britain and Fox Reference Britain, Fox, Filppula, Klemola and Paulasto2009; Gabrielatos et al. Reference Gabrielatos, Torgersen, Hoffmann and Fox2010), more advanced for the definite than the indefinite article, and with the glottal stop serving to break the hiatus in prevocalic environments investigated (as in (106)). Britain, Guzzo and Fox (Reference Britain, Guzzo and Fox2007) find the same development among older male adolescents of Italian heritage in the East Midland town of Bedford.

1. (105)

   And naturally her father was a older man when she was a young girl (Wagner Reference Wagner, Kortmann, Burridge, Mesthrie, Schneider and Upton2004: 155, Cornwall)

1. (106)

   He was like a animal (Britain and Fox Reference Britain, Fox, Filppula, Klemola and Paulasto2009:194, Hackney, London)

4.2.9 Conjunctions

A few studies report the use of non-standard conjunctions. Trudgill (Reference Trudgill and Palmer1995) and Peitsara (Reference Peitsara, Klemola, Kytö and Rissanen1996), for example, discuss the use in East Anglia of ‘consecutive conjunctions’ as in (107) and (108).

1. (107)

   Don’t you climb that tree, do you might get hurt (Claxton Reference Claxton1968:13, Suffolk)

1. (108)

   You lot must have moved it, do I wouldn’t have fell in (Trudgill Reference Trudgill and Palmer1995:139, Norwich, Norfolk)

5.3.1 Clause- or Utterance-Final Items

General extenders (GEs), such as and stuff, or something like that and whatever, prototypically occur in clause-final position, mark the preceding referent (‘roasts’in (1)) as a member of a set (‘English food’), and contain one or more of these components: connector + modifier + generic noun/pronoun + similative + deictic (1) (Pichler and Levey Reference Pichler and Levey2011:448).

1. (1)

   I quite like the English food actually I love roasts and things like that (Cheshire Reference Cheshire2007:174)

The variability of GEs has been investigated in Reading (Cheshire Reference Cheshire2007), London (Aijmer Reference Aijmer2002; Levey Reference Levey2012; Palacios Martínez Reference Palacios Martínez2011; Secova Reference Secova2017; Stenström, Andersen and Hasund Reference Stenström, Andersen and Hasund2002), Hertfordshire (Stenström et al. Reference Stenström, Andersen and Hasund2002), and Milton Keynes (Cheshire Reference Cheshire2007) in the southern region; and in Hull (Cheshire Reference Cheshire2007), York (Denis Reference Denis2010, Reference Denis2011), Maryport (Tagliamonte Reference Tagliamonte2013), and Berwick-upon-Tweed (Pichler and Levey Reference Pichler and Levey2010, Reference Pichler and Levey2011) in the northern region. Normalised GE frequencies vary greatly across locations and social groups. In Berwick, GEs are markedly higher among (young) males. No consistent gender effects were detected among adolescents in Hull, Reading and Milton Keynes; in London, GEs are (marginally) more frequent among females. Social class and age effects are similarly varied. Frequencies are higher among working-class adolescents in Hull and among middle-class adolescents in Reading and Milton Keynes. The frequency of GEs increases with decreasing age in York and Berwick in the north, but in London it is lower among adolescents than adults and pre-adolescents. Focusing on adolescent data only (to rule out confounding age effects), we can discern no clear geographical trends in GE frequency. But we acknowledge that such trends may be masked by effects of social class and interactional context, predictors not consistently controlled across studies.

There are cross-variety similarities in the diversity and distribution of GE forms. Large numbers of individual forms (as many as 84 and 94 in multi-generational data from York and Berwick) exist across varieties, typically distributed unevenly between a small number of frequent and a large number of infrequent forms. Short and adjunctive GE forms (e.g. and that, and stuff) tend to outnumber longer and disjunctive forms (e.g. or something like that). Or something is consistently the most frequent disjunctive GE. In post-1990 datasets, the most frequent adjunctive GEs are and that, and everything, and things (like that), and stuff (like that). Among adolescents, and things may be more frequent in the south than the north.

In Reading, London, Milton Keynes, Hull, York and Berwick, the use of stuff-forms (2) sharply increases among young speakers, especially among highly gregarious adolescents and those with very diverse networks (Denis Reference Denis2011; Secova Reference Secova2017:12). They are less frequent in working-class data from Maryport and Berwick, suggesting either that these forms thrive in mainstream (rather than peripheral) varieties, or that they are innovated by middle- (rather than working-) class adolescents. Robust social class distinctions are also evidenced for and that (3): across both northern and southern varieties, and that is characteristic of working-class sociolects. In contrast to innovating stuff-forms, the widespread use of and that is a retention of a conservative dialect feature (Tagliamonte Reference Tagliamonte2013:176). In Berwick, and that is also used more by male and young speakers. In London, and that is marked for heritage background (white British), and and everything is marked for speaker sex (female). Selected GE forms are marked for region (e.g. or summat and or (n)owt (like that) seem restricted to the north).

1. (2)

   I’ve never been one to be distracted by boys and stuff but I was distracted by man (Secova Reference Secova2017:1)

1. (3)

   she normally goes there and sits on the swings and that (Levey Reference Levey2012:258)

Across southern and northern varieties, GEs with the generic things are more likely than other GE forms to occur with comparative elements and with antecedent referents that match the syntactic-semantic properties of things [+inanimate, +plural], and to perform a set-extending function (see (1) above). Across most varieties, stuff-GEs resemble things-GEs in regularly implicating a general set; they differ in that there is usually a mismatch between the properties of stuff ([+inanimate] [+mass]) and those of the antecedent referent (see (2) above). The properties of and that and and everything also differ from things-GEs in a manner consistent across varieties. They tend to be short and are more likely than other forms to perform interpersonal and textual (rather than set-marking) functions. For example, in (4) and everything functions to reinforce the preceding information about the speaker’s work on a brochure. Unlike and that, which tends to be preferred after nominal referents across most varieties (see (3) above), and everything has consistently high rates after non-nominal referents. Despite these cross-variety similarities, conclusions about the regional distribution of GE syntactic-semantic properties and their functions must remain tentative because of lack of uniformity in methods of analysis.

1. (4)

   And I designed it and did the layout and everything (Aijmer Reference Aijmer2002:241)

Like GEs, grammatically dependent negative-polarity question tags (NPQTs) are frequent, multifunctional and variable across varieties in England. NPQTs consist of an auxiliary, negator and pronoun; the auxiliary and pronoun tend to match the syntactic-semantic properties of the subject and verb of their anchor clause and reverse its polarity (5).

1. (5)

   It’s lush, isn’t it? (Childs Reference Childs2021:427)

   And we could baffle them, couldn’t we? (Pichler Reference Pichler2013:170)

The distribution of (a)in(t)- and isn’t it-tags is analysed across regions in BNC1994 by Anderwald (Reference Anderwald2002) and Krug (Reference Krug1998). Selected NPQTs have also been investigated in: London (Andersen Reference Andersen2001; Cheshire and Fox Reference Cheshire and Fox2009; Erman Reference Erman and Haukioja1998; Kimps et al. Reference Kimps, Davidse and Cornillie2014; Palacios Martínez Reference Palacios Martínez2015; Pichler Reference Pichler2021a; Torgersen et al. Reference Torgersen, Gabrielatos, Hoffmann and Fox2011), Reading (Cheshire Reference Cheshire1981; Cheshire et al. Reference Cheshire, Kerswill, Williams, Auer, Hinskens and Kerswill2005), and Milton Keynes (Cheshire et al. Reference Cheshire, Kerswill, Williams, Auer, Hinskens and Kerswill2005) in the southern dialect region; in Salford (Childs Reference Childs2021) in the central region; in Bolton (Moore and Podesva Reference Moore and Podesva2009) on the northern–central border; and in Hull (Cheshire et al. Reference Cheshire, Kerswill, Williams, Auer, Hinskens and Kerswill2005), York (Tagliamonte Reference Tagliamonte1998), Tyneside (Childs Reference Childs2019, Reference Childs2021; McDonald and Beal Reference McDonald and Beal1987) and Berwick-upon-Tweed (Pichler Reference Pichler2013) in the northern region.

In all varieties examined, NPQTs are multifunctional. The frequency of individual functions fluctuates across comparable studies in London, Salford, Tyneside and Berwick, but tags soliciting listener responses always predominate. In London English, the functional profile of NPQTs is unstable, possibly because of the rise and evolution of innit here. NPQTs not soliciting listener responses are on the rise in London, where selected NPQT forms also guide narrative interpretation (6).

1. (6)

   We all got pizza innit. We beat up the pizza man. Then we took his bike innit, joyriding it. And the police come innit. (Pichler Reference Pichler2021a:729)

In some varieties, function conditions the choice between NPQT forms. In Tyneside English, NPQTs seeking information contain only an isolate negator (7a); those seeking confirmation contain both clitic and isolate negators (7b). Across the central and northern varieties investigated, phonetically non-reduced NPQTs (7) are linked with seeking verification, confirmation and involvement. In adolescent Reading English, ain’t-tags always solicit a listener response while in’t-tags can also convey aggression or hostility toward listeners (8). Krug (Reference Krug1998:151–2) suggests a widespread link between disyllabic NPQTs and non-response elicitation.

1. (7)

   a.	She can’t come, can she not?
   b.	She can’t come, can’t she not?
   (McDonald and Beal Reference McDonald and Beal1987:53)

1. (8)

   You’re a fucking hard nut, in’t you? (Cheshire Reference Cheshire1981:376)

(A)in(t)-tags are infrequent in BNC1994, especially in southern dialect regions and among younger speakers. Their low and decreasing frequency is also evidenced in London, Salford, Bolton, Hull, Tyneside and Berwick. Unlike (a)in(t)-tags, innit is increasing in frequency across England. In BNC1994, innit has a higher frequency south of Trudgill’s (Reference Trudgill1999) central-southern dialect boundary, a finding broadly supported by studies of private corpora.

Methodologically comparable studies exploring the sociolinguistic distribution of innit show regional effects on innit use. In Salford, Tyneside and Berwick in the central and northern dialect regions, innit is outnumbered by, but competes vigorously with, its presumed source form isn’t it. Here, as well as in Reading, Milton Keynes and Hull, innit is largely restricted to third-person singular neuter present tense be-anchors (9a). In London, innit has all but replaced isn’t it and come to be used across anchor contexts (9b). The functional and social profile of innit varies across individual central and northern varieties as well as across northern, central and southern varieties. In northern and central varieties, innit never guides narrative interpretation; in London, it regularly does so among adolescents (see (6) above). Outside London, innit is used more by male (adolescent) speakers and working-class adolescents. In London, male and female adolescents from all class and ethnic backgrounds use innit frequently across all anchor contexts and for all the NPQT functions; amongst those from ethnic minority backgrounds and in multiethnic boroughs, innit has virtually replaced other NPQT forms. Geographical variation in innit use may reflect its longer history in London English, with its spread in London possibly accelerated by widespread language contact in the community.

1. (9)

   a.	Home’s always home, innit? (Pichler Reference Pichler2013:189)
   b.	You get dazed, innit? (Pichler Reference Pichler2021a:724)

Other phonetically reduced NPQT forms are also prevalent across England. For example, the auxiliary-negator string din(t) (for didn’t, don’t, doesn’t) (10a) has been attested in London, Salford, Bolton, and Berwick; and the reduced string wun(t) (for wouldn’t) (10b) in London and Salford. Overall, the frequency of reduced (relative to non-reduced) NPQTs varies across regions and social groups. Notwithstanding potential interviewer effects (Childs Reference Childs2019), reduced forms are more frequent in Salford than in Tyneside and Berwick, among male than female Tynesiders, and among younger than older Londoners.

1. (10)

   a.	You went all weird, dint you? (Moore and Podesva Reference Moore and Podesva2009:470)
   b.	You just fail, wun you? (Pichler Reference Pichler2021a:757)

Non-standard weren’t-tags (11) occur across most dialect areas in BNC1994 but with vastly differential frequencies – highest in East Anglia and lowest in the north-west Midlands. Other studies show that in central and northern England, non-standard weren’t-tags are more frequent in Bolton and York than in Tyneside and Berwick. In York and Outer London, their use is associated with young females, but only in Outer London is weren’t it regularly attached to anchors that do not contain third-person singular neuter subjects and/or past tense forms of be (11b).

1. (11)

   a.	He was shorter and stockier, weren’t he? (Tagliamonte Reference Tagliamonte1998:164)
   b.	Cos I stopped bunning, weren’t it? (Cheshire and Fox Reference Cheshire and Fox2009:25)

Finally, Tyneside English and Berwick English have NPQTs with divn’t (for don’t) and the isolate negator no (for not) (12), which are not typically found outside the north(-east) of England (or Scotland).

1. (12)

   a.	I mean, you get drugs everywhere, divn’t you?
   b.	I think kecks really is underpants, is it no?
   (Pichler Reference Pichler2013:183)

Invariant tags (IVTs) occur across anchor types (13a), and some also occur in pre-finite position (13b, 13c). Here, we focus primarily on their use in clause- or utterance-final position.

1. (13)

   a.	it was more like a rifle, it hadn’t got a stand, you know, more like a rifle butt, you know (Beeching Reference Beeching2016:111)
   b.	This geezer from Bedlam yeah got stopped the other day in his car yeah (Stenström et al. Reference Stenström, Andersen and Hasund2002:172)
   c.	In that episode right she gets pregnant (Stenström et al. Reference Stenström, Andersen and Hasund2002:181)

Although IVTs occur throughout England, their distribution has not been widely investigated. Beeching (Reference Beeching2016) examines you know in BNC1994 but without considering regional effects. Selected IVTs have been studied in Reading (Cheshire and Williams Reference Cheshire and Williams2002), Liverpool (Herat Reference Herat2018), Hull (Cheshire et al. Reference Cheshire, Kerswill, Williams, Auer, Hinskens and Kerswill2005), Tyneside (Bartlett Reference Bartlett2013; Beal Reference Beal, Kortmann, Schneider, Burridge, Mesthrie and Upton2004; Bueno-Amaro Reference Bueno-Amaro2022), and Maryport (Tagliamonte Reference Tagliamonte2013). Several studies analyse (adolescent) data from London (Andersen Reference Andersen and Pichler2016; Erman Reference Erman2001; Pichler Reference Pichler and Pichler2016, Reference Pichler, Beaman, Buchstaller, Fox and Walker2021b; Oxbury Reference Oxbury2021; Stenström et al. Reference Stenström, Andersen and Hasund2002; Torgersen et al. Reference Torgersen, Gabrielatos, Hoffmann and Fox2011; Torgersen, Gabrielatos and Hoffmann Reference Torgersen, Gabrielatos, Hoffmann and Friginal2018). Their variable distribution across London boroughs suggests that IVTs are susceptible to geographical variation. As such, they warrant discussion here.

Despite reported drops in frequency, you know remains (one of) the most frequent IVT form(s) among London working-class adolescents. The functions of final you know may vary across social groups (e.g. Erman Reference Erman2001). In BNC1994, it typically evokes consensual truths or seeks agreement, depending on intonation, and is correlated with unskilled working-class speakers and speakers aged 60+. The age correlation may be widespread; it is discernible in working-class data from Berwick (Pichler and Levey Reference Pichler and Levey2010) and tallies with our informal observations of working-class data from London, Salford and Tyneside. Other IVT forms with you+know, specifically (do) (you) know what I mean and (do) you know what I’m saying, are increasing in London working-class speech, with females, those of white British heritage and those from Outer London in the lead. If you know what I mean is also significantly more frequent in Outer than Inner London.

Other IVT forms that differentiate Inner and Outer London – and possibly other – varieties are: (do)/(if) (you) get me, (do)/(if) you get what I mean, (do)/(if) you get what I’m saying (14). These (you) get (me)-forms are negligible in Outer London. In Inner London, they are innovated by male adolescents from ethnic minority backgrounds, with selected (you) get (me)-tags increasingly adopted by female adolescents from ethnic minority backgrounds and adolescents of white British heritage with highly ethnically diverse friendship networks.

1. (14)

   I don’t care bruv.. you get me? (Torgersen et al. Reference Torgersen, Gabrielatos, Hoffmann and Friginal2018:181)

   that’s how I see it if you can’t stick up for your mum then you’re worth nothing do you get what I mean? that’s how I see it (Andersen Reference Andersen and Pichler2016:38)

The rise of long you know-tags in Outer London and of (you) get (me)-tags in Inner London affects the frequency and distribution of more long-standing IVTs in London English, such as yeah, right, okay and eh (see (13b) above). Although among the most widely used IVTs in London, use of yeah is falling in some areas. It is used more by adolescents who are male, working class and from Outer London. Right and okay are now the least frequent IVTs in London adolescent speech. Both are ethnic minority features; the former is also more frequent among working-class adolescents and in Inner London. Overall, the use of IVTs in London English is markedly higher among adolescents than other age groups. Only the form eh occurs more frequently among adults.

Qualitative analyses of IVTs in London English tend to include innit, on the basis that – like other IVTs – it lacks syntactic-semantic usage constraints in this variety. These analyses reveal that individual IVT forms are not fully interchangeable. For example, okay – unlike innit, yeah, eh and right – is not used to signal uncertainty or solicit extended listener responses; eh – unlike innit, yeah, okay, right – is not used to guide narrative interpretation; and innit performs functions not performed by yeah, right, okay, and vice versa (see Stenström et al. Reference Stenström, Andersen and Hasund2002:184). London male adolescents’ choice of IVT form may also be determined by the story worlds and stances they create and adopt in narratives (Pichler Reference Pichler, Beaman, Buchstaller, Fox and Walker2021b).

We noted earlier that some IVT forms are positionally flexible. They can occur after clause-initial PPs and subject or left-dislocated NPs (13b,c, and 15), where they take narrow scope over the preceding PP or NP and request corroboration of its activation (and identification). For innit, this flexibility has been noted in London English but not in comparable analyses in northern varieties.

1. (15)

   Cos the export people, innit, they sprayed the spray yeah. (Pichler Reference Pichler and Pichler2016:72)

Clause-final like broadly functions as an emphatic device with declaratives and as a signal of interest or surprise with interrogatives (16). It is typically associated with traditional dialects and with north-east England (and Scottish and Irish) varieties, but it is also prevalent in Liverpool and Maryport. In Liverpool and Tyneside, clause-final like is most frequent among young males; across England, it indexes working-class masculinity (Beeching Reference Beeching2016:155). While not exclusive to central and northern varieties, it is generally rare in the south of England and London (e.g. Levey Reference Levey2006:425).

1. (16)

   This shit is pretty new to me, like. (Herat Reference Herat2018:101)

   Do you know what I mean, like? (Bueno-Amaro Reference Bueno-Amaro2022:17)

Final but is noted in Tyneside and southern regions in BNC1994 (Hancil Reference Hancil, Hancil, Haselow and Post2015). In Reading, middle-class male adolescents use both final but and final so more often to focus on referential meanings, and females on affective meaning. Final still is used by adolescents in London (and perhaps more widely in the south) to emphasise or express speaker stance towards preceding propositions.

5.3.2 Clause- or Utterance-Initial and Stand-Alone Items

Initial or stand-alone discourse-pragmatic items such as those in (17) tend to broadly function as focus markers, backchannel responses or attention signals. They have been studied in BNC1994 (Beeching Reference Beeching2016), in London in the south (Andersen Reference Andersen2001; Ilbury Reference Ilbury2021; Oxbury, Hunt and Cheshire Reference Oxbury, Hunt and Cheshire2023; Pichler Reference Pichler and Pichler2016; Sebba and Tate Reference Sebba and Tate1986; Torgersen et al. Reference Torgersen, Gabrielatos, Hoffmann and Friginal2018), and in Bradford (Sebba and Tate Reference Sebba and Tate1986) and Teesside (Snell Reference Snell, Moore and Montgomery2017) in the north of England.

1. (17)

   a.	and I you know they are doing some really amazing things out there (Beeching Reference Beeching2016:101)
   b.	B:	if it was up to them they wouldn’t let no – this would be a strictly white school
   	A:	you know?
   	(Sebba and Tate Reference Sebba and Tate1986:171)

In BNC1994, you know occurs regularly in initial position (17a), where its rightward scope over following propositions draws listeners’ attention to new information. Our own observations suggest that initial you know is not restricted either socially or geographically (though frequencies may vary). This may not be the case for stand-alone you know (what I mean), which was first recorded in the 1980s in creole and non-creole varieties of English spoken by British-born, Caribbean-heritage adolescents in London and Bradford. These tokens scope over and signal agreement with previous speakers’ propositions (17b). Given its origins, we assume that stand-alone you know (what I mean) with agreement-signalling function may be widespread in areas of England with high proportions of Caribbean-descendent residents.

In London, you get me, is it and innit have recently developed into stand-alone items with scope over previous speakers’ propositions, suggesting perhaps a more general capacity for items originally used as IVTs to develop positional, scopal and functional flexibility. In stand-alone or initial position, you get me functions as a general backchannel (18a), is it conveys surprise (18b), and innit signals agreement (18c). Stand-alone is it and innit are limited to working-class adolescents in London and are more frequent among ethnic minority speakers. In the 1990s, they were not found in Hertfordshire, but in the mid-2010s, stand-alone innit was recorded among adolescents in Manchester (Drummond Reference Drummond2018). Perhaps stand-alone innit is now more widespread – at least in multiethnic urban centres in England. Because of the tendency to analyse individual discourse-pragmatic items in isolation, it remains unclear to what extent stand-alone innit competes with or has replaced agreement-signalling uses of stand-alone you know (what I mean).

1. (18)

   a.	Roshan:	he buns it down with man
   	Robert:	no
   	Roshan:	you get me
   	(Torgersen et al. Reference Torgersen, Gabrielatos, Hoffmann and Friginal2018:192)
   b.	Charlotte:	You know what, she’s probably a lesbian.
   	Orgady:	Is it?
   	(Andersen Reference Andersen2001:150)
   c.	Truno:	He’s he’s smart. That man is smart.
   	Josie:	Innit. When he takes the man’s face he puts it ah it was so bad!
   	(Andersen Reference Andersen2001:139)

In London, (ethnic minority) adolescents strategically recruit innit and selected phonetically reduced NPQTs (e.g. in he) to clause- or utterance-initial position in order to draw attention or secure the conversational floor before presenting propositions to be confirmed by (inattentive) interlocutors (19). This recent innovation may not be widespread geographically. It has not been identified in comparable studies of innit and other NPQTs in contemporaneous data from central and northern varieties.

1. (19)

   They always target (.) innit, they always target everyone around here. (Pichler Reference Pichler and Pichler2016:69)

Instead of the established Cockney form oi, male adolescents in east London use ey (pronounced [ʌɪ]~[eɪ]) to attract attention and deploy a dominant or confrontational stance that exerts authority over or displays solidarity with others (20). Other London innovations include utterance-initial and stand-alone wallah(i), originally Arabic and used in a west London multiethnic area by young people of Muslim heritage to express high commitment to propositions. Here, and also in east London, adolescents use I swear, swear down and (on) X’s life in a similar way (21a); in east London, I swear also expresses low commitment. In both locations, say swear and say mums challenge or ask for clarification, as does say wallah in west London. Say wallah and wallah also occur as adjacency pairs, with say wallah giving addressees the option of continuing the talk or replying wallah to close the sequence (21b).

1. (20)

   ey you stepped on man’s huarache’s cus (Ilbury Reference Ilbury2021:625)

1. (21)

   a.	Chantelle:	mother’s life he did, he picked up the white ball (ennit) (Oxbury et al. 2023:829)
   b.	Ali:	they moved and they came back cos the house BURNT. but they fixed it
   	Khadir:	say wallah
   	Ali:	wallah
   	(Oxbury et al. 2023:835)

More established discourse-pragmatic items in initial position include howay, a directive meaning ‘come on’ and limited in use to working-class speakers in north-east England, specifically Teesside and Tyneside.

5.3.3 Clause-Internal Items

Most studies of intensifiers analyse adverbials that scale up the meaning of following adjectival heads (22). We therefore focus on these here.

1. (22)

   They were extremely good (Ito and Tagliamonte Reference Ito and Tagliamonte2003:258)

In the south, intensifiers have been investigated only in London, as far as we know (Núñez-Pertejo and Palacios Martínez Reference Núñez–Pertejo and Palacios Martínez2018; Paradis Reference Paradis and Kirk2000; Stenström Reference Stenström and Kirk2000; Stenström et al. Reference Stenström, Andersen and Hasund2002); in the north, they have been investigated in Manchester (Drummond Reference Drummond2020), York (Ito and Tagliamonte Reference Ito and Tagliamonte2003), and Tyneside (Barnfield and Buchstaller Reference Barnfield and Buchstaller2010; Bueno-Amaro Reference Bueno-Amaro2022; Childs Reference Childs2016; Pearce Reference Pearce2011).

In York and Tyneside, the same three intensifiers are the most frequent: really, very and so. In both locations, really has overtaken very relatively recently, but it entered the intensifier system in different ways. In York, its use spread across a gradually increasing number of semantic categories of adjective before becoming the most frequent intensifier overall. In Tyneside, really occurred with all semantic categories of adjective from the outset but, until 2007 at least, it was the most frequent intensifier with only three categories. By 2017–19, however, really had become the most frequent intensifier for Tyneside teenagers with all categories of adjective. Comparison of BNC1994 and BNC2014 suggests that the rapid replacement of very by really and the increase in the use of so may be country-wide (Aijmer Reference Aijmer, Brezina, Love and Aijmer2018:65). In Outer London, however, the second most frequent intensifier for adolescents is well (23a); very is rarely used. Intensifiers currently attested only in London are enough and bare, the former used only by female speakers from multiethnic communities in Inner London, mainly as nuff (23b). Bare is increasing in both Inner and Outer London, as is well. Both well and totally may be increasing in England more generally (see Aijmer Reference Aijmer, Brezina, Love and Aijmer2018, Reference Aijmer2021, Reference Aijmer, Peterson, Hiltunen and Kern2022 for a comparison between BNC1994 and BNC2014; Bueno-Amaro Reference Bueno-Amaro2022 for Tyneside).

1. (23)

   a.	she was like “oh yeah that boy’s well nice” (Núñez-Pertejo and Palacios Martínez Reference Núñez–Pertejo and Palacios Martínez2018:129)
   b.	his mum looks nuff different though (Fox Reference Fox, Bergs and Brinton2013:2027)

Stratton (Reference Stratton2020), one of the few studies to analyse discourse-pragmatic items in BNC2014 in terms of region, finds proper occurring throughout England. It is very frequent amongst working-class teenagers in Outer London and also occurs, though less frequently, in multiethnic areas of Inner London. For Tyneside teenagers, it is the fourth most frequent intensifier, after really, so and very. Super and fucking increased in BNC2014 relative to BNC1994; in BNC2014, both are used more often by younger speakers, and slightly more often by females (Aijmer Reference Aijmer, Brezina, Love and Aijmer2018). Totally and fucking are also slightly more frequent amongst female working-class teenagers in Manchester.

Intensifiers are well known to be subject to constant change and renewal; those that have declined in use include real (already noted by Britain Reference Britain and Britain2007) and dead.

Characteristically north-eastern intensifiers are geet (or git) (24a) and canny (24b). Canny as an intensifier is a recent development, used more frequently by young males, perhaps reflecting the covert prestige of the traditional north-east dialect where it has long been used as an adverb and adjective. Canny is rare, however, among Tyneside teenagers, who claim to also use it as a downtoner.

1. (24)

   a.	your songs on here are geet good (Pearce Reference Pearce2011:4)
   b.	I’m canny happy he did do that (Childs Reference Childs2016:239)

The overall frequency of intensifiers in England is higher among younger than older speakers, and among female than male speakers. In Tyneside, male teenagers use intensifiers and downtoners at about the same rates, but females use intensifiers far more often than downtoners. These sociolinguistic differences in the overall frequency of intensifiers could suggest more general potential differences in discourse style, both across time and across the country.

Like as a discourse-pragmatic item is extremely flexible in its position, but younger speakers, in most locations, use it more frequently clause-internally (25). It has been studied in the southern dialect region in Reading (Cheshire et al. Reference Cheshire, Kerswill, Williams, Auer, Hinskens and Kerswill2005), London (Andersen Reference Andersen2001; Levey Reference Levey2006; Cheshire and Secova Reference Cheshire and Secova2018:222–3), and Milton Keynes (Cheshire et al. Reference Cheshire, Kerswill, Williams, Auer, Hinskens and Kerswill2005); in the central region in Liverpool (Herat Reference Herat2018); and in the northern region in Hull (Cheshire et al. Reference Cheshire, Kerswill, Williams, Auer, Hinskens and Kerswill2005) and Maryport (Tagliamonte Reference Tagliamonte2013).

1. (25)

   I got like a scar under my eyebrow (Levey Reference Levey2006:414)

Studies comparing overall frequencies across real or apparent time in England show a dramatic increase within the last few decades. For most speakers in Liverpool, like is the most frequent discourse-pragmatic item overall. It is more frequent among adolescents in Hull than in Reading or Milton Keynes, perhaps reflecting its well-established use in clause-final position in northern varieties (see (16), section 5.3.1). The age distribution in BNC1994 indicates that the rapidly increasing frequency represents a language change rather than age-grading (Beeching Reference Beeching2016:144). In Outer London, 7–8-year-olds use like just as often as 10–11-year-olds.

There are no clear sociolinguistic patterns across varieties with either social class or speaker sex. In BNC1994, clause-internal like is used more often by 15–24-year-old middle-class females. In Outer London, pre-adolescent boys use like more often before NPs where it can alert listeners to the introduction of a new referent (25), perhaps reflecting a more general orientation for boys to attend more to referential meanings in the construction of talk, whereas girls use like, with other discourse-pragmatic items, to negotiate interactional alignments and establish solidarity. In Liverpool, on the other hand, like occurs more often as a focuser in female speech, as well as a ‘metonymic exemplifier’ and quotative.

Be like (26) and other quotative expressions have, to our knowledge, been analysed only in London (Cheshire et al. Reference Cheshire, Kerswill, Fox and Torgersen2011; Fox Reference Fox, Buchstaller and van Alphen2012) in the south; and York (Tagliamonte and Hudson Reference Tagliamonte and Hudson1999; Durham et al. Reference Durham, Haddican, Zweig, Johnson, Baker, Cockeram, Danks and Tyler2012) and Tyneside (Buchstaller Reference Buchstaller2011, Reference Buchstaller2015; Pearce Reference Pearce2011) in the north. These studies show a recent dramatic increase overall in the frequency of be like across real and apparent time, although Buchstaller (Reference Buchstaller2015) finds a move towards more conventional quotative forms in Tyneside as speakers approach middle age. Her nuanced analysis of change across the lifespan of individuals takes account of social and ideological factors associated with the use of be like.

The widespread emergence of be like is presumably influenced by the equally rapid increase in the use of clause-internal like. Older speakers in London and Maryport who use discourse-pragmatic like relatively infrequently, albeit in all clause positions, do not use be like at all.

1. (26)

   And I’m like “Oh my God! Oh my God! Oh my God!” (Tagliamonte and Hudson Reference Tagliamonte and Hudson1999:157)

In the early 2000s, overall frequencies of be like in both London and Tyneside were lower than in York. This may reflect a social class difference: in London, the speakers analysed were working class, while in York they were university students; and in Tyneside, young middle-class speakers used be like more than twice as often as working-class speakers of the same age. Female speakers used be like more often in the earlier York study and in Inner London, whereas in Tyneside and Outer London young male speakers were in the lead.

Quotative go (27) was frequent in Tyneside in the 1990s but had declined by the 2000s, probably in response to the rise of be like. In York, however, be like mainly usurped only say; go occurred at the same rate as be like in the earlier study and had declined only slightly by 2006. In Outer London and, especially, Inner London, go is robust for young working-class speakers.

1. (27)

   She goes “I might not wear them” (Buchstaller Reference Buchstaller2011:59)

Be like is not the only innovation to have entered the quotative system in England. Git/geet in north-east England occurs not only as an intensifier (see above) but also, less frequently, with a function comparable to discourse-pragmatic like (28) and in the quotative expression be git/geet (29), prompting Pearce (Reference Pearce2011) to suggest that be geet may have developed by analogy with be like.

1. (28)

   I was geet working it out in my head (Pearce Reference Pearce2011:5)

1. (29)

   I was geet “ehhhhh” (Pearce Reference Pearce2011:5)

Another recent localised quotative is this is + speaker (30), found by Cheshire et al. (Reference Cheshire, Kerswill, Fox and Torgersen2011) in multiethnic areas of Inner London. It was used more frequently by female speakers, and often to quote direct speech at particularly dramatic peaks in performed stories (Fox Reference Fox, Buchstaller and van Alphen2012). It is not attested in subsequent London studies (e.g. Oxbury Reference Oxbury2021) and seems, therefore, to have been a short-lived innovation. Still more fleeting is give (it), mentioned in Harris (Reference Harris2006:114) and occurring just once in Cheshire et al.’s (Reference Cheshire, Kerswill, Fox and Torgersen2011) Outer London data.

1. (30)

   this is my mum “what are you doing? I was in the queue before you” (Cheshire et al. Reference Cheshire, Kerswill, Fox and Torgersen2011:172)

Some researchers explain the recent emergence of new quotatives as resulting from changes in narrative style. For be like, the reporting of narrators’ inner thought processes (26) may have created a niche in the quotative system that be like then filled to create a more dramatic effect than think or say (Tagliamonte and D’Arcy Reference Tagliamonte and D’Arcy2007). This would account for the widely attested initial preference for the use of be like in first-person contexts to quote reported thought. Fox (Reference Fox, Buchstaller and van Alphen2012) suggests that as be like comes to introduce reported speech as well as thought, a new quotative is needed to highlight dramatic peaks in narratives. In Tyneside, the need was met by the resurgence of quotative go, which was latent in the system, while in Inner London the brand-new quotative this is + speaker emerged. The functions of geet as intensifying adverb and adjective (see above) may similarly make it appropriate for quoting moments of high drama in performed narratives.

In Tyneside, the addition of two new quotative expressions sparked off more subtle changes to the system. In the 1960s Tyneside corpus, say was the main quotative, used in a wide range of tenses. By 2000, however, adolescents used say mainly in the conversational present tense, with be like and go now used mainly in the simple past tense. Furthermore, young people used be like and go more often in narratives, reserving say for other contexts, when the speech they were reporting was not part of a story. Older speakers continued to use say in all types of context.

6.2.1 Overview

Anyone hearing English in Scotland is presented with a range of accents which vary by location and social context. In peripheral areas, such as the Borders, the North East and the Northern Isles, it is usual to find strongly divergent accents, Broad Scots spoken by speakers of all ages and backgrounds, and SSE, with some speakers only really having command over one or the other. This polarised accent situation also occurs in some working-class Central Belt communities between those who use highly consistent Broad Urban Scots phonology (Stuart-Smith Reference Stuart-Smith, Corbett, McClure and Stuart-Smith2003) and educated, mainly professional, middle-class groups favouring more or less exclusively the accepted, ‘proper’, SSE accent (Abercrombie Reference Abercrombie, Aitken and McArthur1979; Johnston Reference Johnston and Gorlach1985). But within and between the large Edinburgh and Glaswegian conurbations, there are also very many, especially in upper-working and lower-middle-class areas, whose relative command of Scots and SSE phonology is overlapping and variable in daily use. Understanding Scottish English phonology is made easier by becoming comfortable with the notion of Scottish speech as a locus of similarity and difference at the same time (Maguire Reference Maguire and Hickey2013).

Standing back, the overall impression is that these accents are similarly ‘Scottish’, and this is because all share common phonological characteristics which bind them together and contrast them with other English accents. These include: a largely shared segmental inventory, rhoticity of some kind, a high degree of overlap in lexical selection, including one vowel phoneme (cat, boot, cot) where others show two, largely monophthongal face and goat vowels and, most distinctively, operation of the phonologically and morphologically induced durational patterns known as the Scottish Vowel Length Rule. According to this – largely synchronic – perspective, SSE can be considered as the outer rim of concentric spheres about a core of increasingly Scots accents.

At the same time, focusing in on any point immediately reveals systematic differences in terms of pockets of lexical incidence, interconnected prosody and patterns of segmental realisation, constituting urban/rural Scots and (regional) SSE accents, each carrying associations of identity, place and social class, especially in the Central Belt, and each continuing recognisably different historical developments. This more diachronic perspective accentuates a bipolar accent continuum (Aitken Reference Aitken and Trudgill1984b; Johnston Reference Johnston and Britain2007), effectively made up of overlapping spheres, with Scots larger in speaker numbers, accent diversity and time-depth, and SSE smaller, relatively less diverse and shallower. For purposes of accent description, the phonological constructs of Scots and SSE serve as useful abstractions.

Another key related phonological dimension is linguistically how accent can be ‘carried’. As for any traditional dialect (Wells Reference Wells1982), Broad Scots words and grammar must be pronounced with Scots phonology, whereas Standard English forms can show Scots or SSE phonology. This means that an impression of ‘broad’ Scots can be achieved in two ways: through a higher proportion of Scots lexis and grammar with remaining forms produced with Scots phonology (‘denser’ Scots), mainly but not exclusively found in rural areas; or less Scots forms and more Standard English forms, but all produced with Broad Scots phonology, typical of Central Lowland communities.

Our brief outline concentrates on Urban Scots in this latter situation, partly because we know most about it from sociophonetic research (e.g. Macafee Reference Macafee1994, Reference Macafee and Jones1997; Macaulay Reference Macaulay1977; Stuart-Smith Reference Stuart-Smith, Corbett, McClure and Stuart-Smith2003; Schützler Reference Schützler2015; Stuart-Smith et al. Reference Stuart-Smith, Lawson and Hickey2017), and because – alongside, and in variable use with SSE features – this kind of accent is typical of most Scottish English speakers. As we will show, the largely common ‘Scottish’ phonological core is cloaked by systematically divergent, yet intersecting, layers of Scots and SSE phonological variation (cf. Aitken Reference Aitken and Trudgill1984b:519f.; Aitken Reference Aitken and Trudgill1984a:94f.). We also draw on recent analysis of approximately 1,100 speakers from the eleven Scottish corpora from the Speech Across Dialects of English (SPADE) (https://spade.glasgow.ac.uk/); for more detail on regional Scots phonology, see Johnston (Reference Johnston and Jones1997, Reference Johnston and Britain2007).

6.2.2 Vowels

Scots and SSE share the following vowels: /i ɪ e ɛ a o ɔ ʉ ʌ ɔi ae oe ʌʉ/. Table 6.1 shows how the vowels are lexically distributed for Scots (column 1) with broad phonetic realisations for Urban/Central, Southern and Northern Scots (columns 2–4), and broad phonetic realisations with lexical equivalents for SSE (columns 5 and 6). Comparison of Urban/Central with Southern and Northern Scots (columns 2–4) shows some of the dialectal diversity within Scots. Comparison of Urban/Central Scots and SSE vowel qualities (columns 2 and 5) illustrates how the ‘same’ vowel phonemes with similar lexical incidence show systematically different phonetic qualities across the Scottish English continuum. Table 6.1 also indicates some common lexically specified alternations for Urban/Central Scots (e.g. out with /ʉ/ or /ʌʉ/ for oot/out, hoose/house), though again, the quality of /ʌʉ/ is phonetically different in Scots and SSE (Eremeeva and Stuart-Smith Reference Eremeeva and Stuart-Smith2003). The salience of these socially stigmatised, alternating forms probably arises because they are vigorously maintained by younger working-class adolescents and because they are so lexically infrequent, which then increases their prominence. out is the most common alternating set, but Macafee showed that this only occurs to any extent in seven different words, and overall the number of possible tokens for the alternation comprised less than 2 per cent of her whole corpus (Macafee Reference Macafee1994; Stuart-Smith Reference Stuart-Smith, Corbett, McClure and Stuart-Smith2003, also for more alternations, e.g. heid/head, aff/off, etc.).