Special issue on speech representation in Late Modern English text types: introduction

One of the key challenges in the field of historical linguistics is the lack of spoken records until the twentieth century. In order to reconstruct spoken language in the past, to investigate how spoken and written language have potentially influenced each other, and to shed light on language variation and change over time more generally, researchers have therefore had to focus on written records. As Kytö (1991: 29) aptly notes in relation to this matter, ‘we must ask which types of text would most reliably reflect the spoken language of the past, or, perhaps more to the point, what kind of spoken language the texts convey to us and how faithfully and accurately’. Depending on the historical period under investigation, different text types are available to historical linguists, reflecting language use in certain regions, situations associated with different levels of formality, as well as select authorship groups, notably those who were able to write at times when access to education and particularly literacy training were socially stratified. While the reconstruction of spoken language diachronically has been central to the fields of historical linguistics and historical dialectology (e.g. Dossena & Lass 2004, 2009; Boberg et al. 2018; Alcorn et al. 2019), the more recently established fields of historical pragmatics (e.g. Jucker 1995, 2000) and historical sociolinguistics (e.g. Auer et al. 2015; Nevalainen & Raumolin-Brunberg 2017) are similarly interested in reconstructing speech. Anchored in the observation that language histories to date have largely been based on written sources produced by the higher social strata who had the prerogative of receiving an education, an important goal of the field of historical sociolinguistics, which aims at studying the relationship between language and society in the past, is to give a voice to traditionally unheard social groups such as women and the lower classes. This so-called ‘language history from below’ approach (Elspaß 2007), which focuses on texts untouched by editors in addition to language produced by the non-elite (cf. Auer 2018: 15), similarly aims at investigating oral registers that can be reconstructed based on written sources, while also considering the interrelationship between speech and written language. Text types approximating speech, which are not too affected by writing conventions, are for instance so-called ego documents² like personal correspondence, diaries and travelogues, as well as text types like sermons, trial proceedings, proclamations, drama texts, and speech presentation in prose fiction (cf. Culpeper & Kytö 2010: 17–18; Auer et al. 2015). In relation to these types of written records, Schneider (2004: 68) observes that ‘a written record of a speech event stands like a filter between the words as spoken and the analyst’, which is why the researcher needs to ‘remove the filter’, i.e. to consider and carefully assess the recording process in order to be able to reconstruct the speech event itself. The current special issue ‘Speech representation in Late Modern English text types’ tackles this challenge for selected text types in the Late Modern English period, which is here considered to cover the period c. 1700–1900. This is in line with traditional divisions of the history of English into three periods, whose starting and end points are associated with historical events that had an effect on the development of English (see Curzan 2012 for a detailed discussion of chronological divisions). The starting point 1700 can be linked to the Acts of Union (1707) that unified the Kingdom of England and the Kingdom of Scotland, creating Great Britain, as well as the Industrial Revolution with its technical inventions and the great impact on demography such as urbanisation, which led to the rise of literacy levels. While the choice of c. 1900 as endpoint creates a lacuna in the twentieth century, which, for instance, Beal (2004: 1–2) remedies by extending the period to the end of World War II in 1945, c. 1900 may be taken as demarcation from revolutionary inventions linked to the recording of spoken language such the invention of the phonograph in 1877, the radio in 1895, and the establishment of the British Broadcasting Company (now Corporation, i.e. BBC) in 1922 (cf. Beal 2004: 9–10). Given the focus of the current special issue on speech reflections in different text types and a scarcity of spoken records in the nineteenth century, the traditional periodisation is therefore deemed appropriate.

The Late Modern English period is of particular interest for the study of speech representation in written sources for several reasons:

1. (i) Autograph texts by writers across the social spectrum are available, in addition to texts that represent the speech of different social groups through a mediator such as an author, a narrator or a scribe. Due to increased literacy levels throughout the period and the Elementary School Act (1870) that provided a framework for compulsory elementary schooling for children aged 5–12, occupational literacy (for men and women) saw a gradual increase in 1700–70 (e.g. around 75% of male yeomen and husbandmen versus 32% of female yeomen in 1770, in contrast to 36% of male labourers versus 12% of female labourers in 1700; based on Houston 1985 in Sanderson 1995: 11), followed by a rapid rise after 1840. This increase had an impact on the text production in previously less literate social classes as well, which allows historical (socio)linguists to investigate data sources such as correspondence from different social layers, e.g. the recently compiled Mary Hamilton Papers (c. 1740 – c. 1850)³ and the ongoing pauper letter corpus project The Language of the Labouring Poor in Late Modern England.⁴

2. (ii) The Late Modern English period is strongly associated with standardisation processes of written and spoken English, notably the codification of norms and their ensuing prescription as the ‘correct’ variants. As a result, many normative works such as grammars and pronunciation guides were published, some of which enjoyed great popularity at the time, e.g. Robert Lowth's Short Introduction to English Grammar (1762) and John Walker's Critical Pronouncing Dictionary (1791). Both the written and the spoken varieties promoted were strongly linked to London; while the written variety was associated with the polite language of educated gentlemen from the metropolis, according to Sheridan (1780: Preface), the pronunciation to appeal to was the language spoken at Court in the early eighteenth century. The promotion of a spoken prestige variety necessarily led to the stigmatisation of forms that deviated from the latter. As societal changes during the Industrial Revolution (c. 1760–1850) provided opportunities for social climbing, which can be linked to linguistic insecurity, some normative works, e.g. the pronouncing dictionaries by Sheridan (1780) and Walker (1791), pointed out pronunciation ‘mistakes’, for instance by people from Ireland, Scotland and Wales. Some of these normative works thus provide insight into promoted and stigmatised linguistic features, and sometimes draw connections with the language use in different regions (example above) and different social groups. These sources can be of great value for the study of enregisterment, which Agha (2003: 231) defines as ‘a linguistic repertoire differentiable within a language as a socially recognised register’ that then indexes ‘speaker status linked to a specific scheme of cultural values’.

3. (iii) In addition to sociolinguistic differences reflected in normative linguistic works, certain stereotypes can also be found in contemporary literature, albeit represented by an author. As the Late Modern English period also encompasses the Romantic literary movement (c. 1798–1837), which promoted the focus on the ‘real language of men’, as noted by Wordsworth in the Preface to Lyrical Ballads (1800), this development may therefore also provide more insight into the spoken language of the middling sort and the labouring poor.

Within this context, the current special issue contains invited papers that were first presented at an international workshop that was organised by the Swiss National Science Foundation (SNSF)-funded research project The Language of the Labouring Poor in Late Modern England in November 2021. The seven contributions focus on a range of different data sources that approach the study of speech representation in late modern times in different ways, while also finding commonalities and complementing each other.

For instance, Beal's contribution, ‘“Practised among the common people”: “Vulgar” pronunciations in eighteenth-century pronouncing dictionaries’, focuses on a text type that is strongly associated with the codification stage of the English language, notably normative works called pronouncing dictionaries. More precisely, the article focuses on metalinguistic comments on pronunciations that were labelled ‘vulgar’ by John Walker (1791) and some contemporaries, with the aim to detect links between linguistic variants that were associated with the lower classes. Beal shows that metalinguistic comments played an important role in the enregisterment of the prestigious spoken variety of English and the stigmatisation of speech variants that were associated with the lower social classes (cf. Agha 2003, 2007). While the negative ‘vulgar’ label was often linked to lower-class pronunciations, Beal notes that the application of the label served as a warning to the middle-class readership of pronunciation dictionaries.

Normative works, notably the Collection of Nineteenth-century Grammars ⁵ (CNG; see Anderwald 2016), also serve as the basis for Wiemann's contribution, ‘Representations of phonological changes in goat and /r/ in the Collection of Nineteenth-century Grammars (CNG)’, that investigates the evidence provided by grammarians for early diphthongisation in goat words as well as for changes in the distribution of /r/ variants in prestigious nineteenth-century accents. Wiemann observes in the norm corpus that monophthongal goat variants and a twofold distinction of /r/ sounds are featured throughout the nineteenth century. While this may be taken as an indication that these were the pronunciations of educated speakers, it is next to impossible to verify this in a usage corpus, particularly in the language of the educated elite as they largely adhered to grammar rules and therefore do not contain any speech reflections. The frequent use of labels like ‘improper diphthong’ to refer to goat suggests that many Late Modern English grammarians were copying from each other; at the same time, the normative works also contained much variation regarding their terminology. Generally, while a good amount of insight into nineteenth-century pronunciations can be gained, the data have to be treated with caution.

While the articles by Beal and Wiemann are based on metalinguistic comments on pronunciation in normative works, the contributions by Cooper, Hodson and Ruano-García use literary texts – dialect literature and literary dialect – as the basis of their investigations. A characteristic that normative works and literary texts share is that they do not directly reflect actual language usage; rather, we are dealing with the writers’ perceptions and representations of the spoken language used in a specific region and/or by members of certain social groups. For instance, Cooper's article, ‘Yorkshire folk versus Yorkshire boors: Evidence for sociological fractionation in nineteenth-century Yorkshire dialect writing’, focuses on nineteenth-century discourses on enregistered Yorkshire dialect and identity as reflected in dialect literature, literary dialect and dialect poems, ballads, songs and dialogues, as well as dialect use by Yorkshire characters in novels and plays. The study observes that literary writers, described as the outgroup, depicted Yorkshire characters using generic enregistered dialect as boorish, i.e. a stereotype that emerged in the eighteenth century and was maintained in nineteenth-century texts. In contrast, local writers, i.e. the ingroup, deemed the representations of the characters and the dialect use inaccurate, which led them to include a wider and more authentic range of dialect and character representations, including more occupations.

Like Beal and Cooper, Hodson's and Ruano-García's articles illustrate the important role of enregisterment and how it results in social values becoming attached to the speech used in certain regions and/or by social groups. In her contribution, ‘Talking to peasants: Language, place and class in British fiction 1800–1836’, Hodson focuses on the representation of the language of the labouring poor, based on 100 novels included in the Dialect in British Fiction 1800–1836 database⁶ (see Hodson et al. 2014). Hodson finds that the lives and voices of the labouring poor do not seem to be of great interest during the period under investigation. Nevertheless, the chronological observation shows a development of the voices of the labouring poor being sometimes represented in fiction at the start of the period, followed by a time during which a clear increase in the fictional representation of the language of the labouring poor in novels set in rural Scotland and Ireland can be observed. The end of the period sees a continued representation of the rural Scottish and Irish labouring poor, and also an extension into extraterritorial varieties. Generally, the article provides insight into changing attitudes towards, as well as understandings of, the language of the labouring poor. In addition, the systematic study of the works resulted in an extensive list of non-standard linguistic variants that can serve as the basis for future studies, e.g. further developments regarding the representation of the language of the labouring poor in fiction or a combination of findings in fiction with those of other text types.

Ruano-García's contribution, ‘“Well, taakin about he da bring inta me yead wat I promised var ta tell ee about”: Representations of south-western speech in nineteenth-century dialect writing’, uses data from the Salamanca Corpus,⁷ a digital archive of English dialect texts (c. 1500 – c. 1950; see García-Bermejo Giner et al. 2011–), to shed light on the language of south-western speakers in the nineteenth century. More specifically, Ruano-García investigates the distribution and frequency of periphrastic do and pronoun exchange in the data, showing that while neither feature occurs very frequently, both the distribution and use of the forms largely pattern with later findings. From the perspective of third-wave sociolinguistic models, the findings also give insight into the social meanings of the linguistic features, notably their being employed as indexicals to create and represent south-western characters. Thus, Ruano-García's article contributes to studies on enregisterment and also the reconstruction of the development of selected linguistic features.

South-western speech is also the focus of Anne-Christine Gardner's article, which is entitled ‘Speech reflections in Late Modern English pauper letters from Dorset’. In this contribution, the text type of pauper letters serves as the basis for historical dialectological research. More precisely, based on thirty-one poor-relief application letters sent by ten individuals to parishes in Dorset, Gardner investigates linguistic variation by focusing on spelling and morphosyntax. The linguistic features found in the pauper letters are compared to modern dialect surveys as well as more contemporary sources. The systematic investigation of linguistic features provides insight into the provenance of some of the letters. The study, which combines a focus on the lower social orders with regional variation, thus shows convincingly that pauper letters can contain reflections of regional speech that allow us to gain more knowledge about older dialect stages in England.

Finally, Peter Grund's article, ‘Disgusting, obscene, and aggravating language: Speech descriptors and the sociopragmatic evaluation of speech in the Old Bailey Corpus’, investigates the representation of spoken language through speech descriptors that bring together speech representation and evaluation. The text type under investigation is the proceedings of London's central criminal court, the Old Bailey, covering the period 1720–1913, thus speech-related texts that are sociolinguistically, pragmatically and textually annotated in the Old Bailey Corpus ⁸ (see Huber et al. 2016). Grund's study of speech descriptors combined with the lemma language shows that the use of the descriptors avoids the repetition of offensive and/or inappropriate utterances; instead, the focus is deflected from the original speech and focused on the evaluation and impact of the descriptor wording. This wording will in turn affect the representation, be it mitigation, deflection or disproval of an accusation. Generally, the article shows that studies of spoken language in the past also need to consider the complex choices of language users in terms of wording, i.e. what they choose to represent in their speech, how this is done, and how it reflects the evaluation of the spoken language and also the speaker.

The seven contributions encompass a range of different text types that shed light on spoken language in different ways, notably (i) normative works like pronouncing dictionaries and grammars that contain metalinguistic comments about language use and the associated speakers, (ii) literary sources that can range from plays and novels to poems, ballads and songs, that may be labelled as and/or contain dialect literature or literary dialect, (iii) trial proceedings that represent speech in the context of the criminal court, and (iv) pauper application letters that can reflect the speech of the writers, particularly if they have not received (much) grammatical schooling and are therefore not familiar with some established norms of written English. In most of the studies, the representations of spoken language have been carried out by grammarians, lexicographers, orthoepists, literary authors and court scribes, thus professions that require a certain level of literacy and special training, therefore also indicating a certain social standing. Through their comments on and linguistic choices in speech representation, they also provided an evaluation of the represented speech, attaching social and moral values to certain speakers or regions. In contrast, in the case of the pauper letters, it is precisely the lack of literacy training (linked to a lower social standing) that provides insight into spoken language. The more educated the writer of a pauper letter was, the more difficult it is to determine his or her origin and to gain insight into spoken language and thus data that can contribute to historical dialectology. While the historical reconstruction of speech will remain challenging and cannot be perfectly accurate and reliable, the combination of different text types, linked to a particular awareness of the accuracy and faithfulness of the source as well as the importance of removing the filter, i.e. ‘the relationship between the speech event and the record’ (Schneider 2004: 68), as has been done in contributions to this special issue, seems promising for the field of historical (socio)linguistics and dialectology. Due to the greater availability of text types in the Late Modern English period that can contain metalinguistic comments on language use, represent writers’ or scribes’ perceptions and evaluations of language use, and can illustrate language use approximating speech (particularly at a lower social level), the period still provides many uninvestigated sources and therefore more opportunities for the reconstruction of speech in the future.