2.1.1 Introductory Remarks

As empirical work on intensifiers and many other linguistic features has so far shown, the systematic study of language use and its history benefits from access to large amounts of material in electronic form. Having the material annotated for extralinguistic factors and grammatical features further facilitates data collection and subsequent screening of raw data. Among recent advances in the field are smaller-size or large-scale stratified electronic multi-genre or specialized corpora, computerized and searchable text collections containing printed material, linguistic editions containing transcriptions based on manuscript material, and other computerized resources such as dictionaries, thesauri, and linguistic atlases (Reference Claridge, Lüdeling and KytöClaridge 2008; Reference Kytö, Jucker and TaavitsainenKytö 2010, Reference Kytö2011; Reference Curzan and KytöCurzan 2012; Reference Rissanen and KytöRissanen 2012). Such materials enable the researcher to pin down correlations between linguistic and extralinguistic variables with the help of sophisticated statistical analyses, observe the shifting frequencies of alternative or rival expressions, and trace back the process of change across time. Indeed, there are close affinities between corpus studies and the study of language variation and change via the lenses of historical sociolinguistics and historical pragmatics (Reference Jucker, Taavitsainen, Taavitsainen, Jucker and TuominenJucker and Taavitsainen 2014: 5–9); we return to these connections later when outlining some of the analytical frameworks which have proved central to our investigation of intensifiers in LModE.

For the purposes of the present study, the OBC provided a rewarding 24.4-million-word resource, which contains specimens of speech-based language annotated for both sociopragmatic features and part-of-speech (POS) information. Before looking into the structure and contents of the OBC, we briefly describe the even more extensive digitized text collection, The Proceedings of the Old Bailey, 1674–1913 (henceforth Old Bailey Online), from which the OBC text files were drawn. The publishing history of the original printed Proceedings of the Old Bailey (henceforth the Proceedings) is of particular interest to corpus linguists planning to use the OBC as it provides valuable clues to the question of how reliable these texts may be as a vehicle relaying features of past spoken interaction. Our following account is mainly based on Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker (2018a, Reference Emsley, Hitchcock and Shoemaker2018b, Reference Emsley, Hitchcock and Shoemaker2018c, Reference Emsley, Hitchcock and Shoemaker2018d, Reference Emsley, Hitchcock and Shoemaker2018e, Reference Emsley, Hitchcock and Shoemaker2018f, Reference Emsley, Hitchcock and Shoemaker2018g, Reference Emsley, Hitchcock and Shoemaker2018k) at the Old Bailey Online website (www.oldbaileyonline.org/), Reference Huber, Nissel, Maiwald and WidlitzkiHuber et al. (2012), Reference Huber, Nissel and PugaHuber, Nissel, and Puga (2016), and Reference WidlitzkiWidlitzki (2018), with references to a number of other prominent sources.

2.1.2 The Proceedings and Old Bailey Online

The web resource Old Bailey Online is one of the outcomes of the project ‘The Proceedings of the Old Bailey, 1674–1913’ carried out under the supervision of Tim Hitchcock and Robert Shoemaker (Reference Hitchcock, Shoemaker, Emsley, Howard and McLaughlinHitchcock et al. 2018). The original printed Proceedings, serving as the input of the web resource, comprise 2,163 volumes of the records, containing 197,745 criminal trials held at London’s central criminal court from 1674 to 1913, totalling some 134 million words (Reference Huber, Nissel and PugaHuber, Nissel, and Puga 2016: 2). It gives a unique view of late modern courtroom practices (Reference LangbeinLangbein 2003) and paints a multifaceted picture of the types of crimes taken to the court, the role played by the accused, defendants, witnesses, and legal professionals, and the sentences handed out. It also allows one to gain insights into the daily lives of ordinary people from all walks of life in late modern England and into aspects of domestic, commercial, entrepreneurial, medical, legal, and other domains of societal engagement between people. The Old Bailey served the City of London and the County of Middlesex up until 1834 when the Central Criminal Court Act extended its jurisdiction to cover metropolitan Essex, Kent, and Surrey (Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018a; Reference MayMay 2003: 147, referred to in Reference WidlitzkiWidlitzki 2018: 66–7). This has been taken to mean ‘that we know where the majority of speakers in the Proceedings lived’ (Reference WidlitzkiWidlitzki 2018: 67); of course, not all the individuals involved in a crime that was committed in the area necessarily had their domicile there.

The early editions of the Proceedings go back to the seventeenth-century popular tradition of crime accounts where notorious criminals’ lives were described in broadsides, chapbooks, and other ephemera. It was in the 1670s that the publication of crime literature exploded, and the first published collection of Old Bailey trials, entitled News from Newgate, appeared in 1674 as a pamphlet containing a selection of cases. The first edition describing all the trials at a single session appeared in October 1678, paving the way to the regular publication of the Proceedings of each meeting of the court. These early publications were commercially successful and consumed by a wide audience. The trials selected for the earlier Proceedings were often about rape, sodomy, and murder, which were considered shocking crimes, titillating the curiosity and sensibilities of upper-class and middle-class readers. Across the nineteenth century, newspapers and more expensive production costs caused the general public’s interest in the Proceedings to decline; the last edition dates to April 1913 (for further details, see Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018a; see also Reference ShoemakerShoemaker 2008; Reference Traugott, Pahta and JuckerTraugott 2011: 69–70).

Over its 239 years of publication history, the Proceedings changed in character from sensationalist audience-pleasing accounts into quasi-official reports. The year 1679 marked a milestone in the history of the Proceedings: from that year onward, the approval of the Lord Mayor of London was required for publication. This had consequences for language use, as all too judgemental and crowd-pleasing formulations started to be discouraged; later on, towards the end of the eighteenth century, language use was monitored occasionally to keep it respectable, for example, in trials where cases of sexual offence were dealt with. As Reference WidlitzkiWidlitzki (2018: 51) put it, the publishers ‘had to perform a delicate balancing act to satisfy their two major stakeholders: their paying customers and the City of London authorities’. After the sensationalist start and from the 1720s onward, the intended target audience of the Proceedings came to comprise ‘respectable’ readers from the middle and upper classes (Reference ShoemakerShoemaker 2008: 563, 565; Reference WardWard 2014: 25–6; Reference WidlitzkiWidlitzki 2018: 52). Even so, the Proceedings have also been thought to have reached ‘lower-class Londoners, including those accused of crimes’ (Reference ShoemakerShoemaker 2008: 575; Reference WidlitzkiWidlitzki 2018: 52).

Regarding competition in the market, in addition to the Proceedings, there were alternative publications reporting on trials at the Old Bailey. In the early days of the Proceedings, publishers sent scribes to the Old Bailey to take notes of ‘juicy’ trials that were thought to interest readers, but in the course of the eighteenth century, the Proceedings were expanded with verbatim testimony and advertisements, so as to meet the competition by alternative accounts and daily newspapersFootnote ¹ that were entering the market (Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018a; Reference WidlitzkiWidlitzki 2018: 52). Even one and the same publisher could print an account of a trial in the Proceedings and in an alternative publication. Reference HuberHuber (2007) discusses one of these cases, The tryal at large of John Ayliffe (1759), and shows how the two versions share overlapping stretches of text but also differ substantially regarding length, and lexical, morphological, and syntactic features. However, such overlapping publications were occasional phenomena rather than the rule. Further competitive publications comprised a number of volumes entitled Select Trials from the Old Bailey in 1718–20, 1734–35, 1742, and 1764, and collections of notorious crimes such as The Newgate Calendar, which entered the market in 1773, and the four-volume Annals of Newgate (1776). The trials in these collections were based on accounts from earlier times, as there was apparently little interest in contemporary criminals’ lives by the 1770s (Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018 g, Reference Emsley, Hitchcock and Shoemaker2018k).

It was in 1712 when instances of above-mentioned verbatim testimony started to appear in the trial accounts, instead of the brief summaries of trials that had been customary (‘especially in trials which were thought to be salacious, amusing, or otherwise entertaining’, Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018a; see also Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018b). Verbatim records increased in volume in the 1720s, when shorthand became more commonly used. More space was given to testimony delivered by witnesses and to contributions by defendants, prosecutors, and judges (for further details, see Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018a). Indeed, over the eighteenth century, the Proceedings became the legal record of the trials held at the Old Bailey, and, in 1778, the City ‘demanded that the Proceedings should provide a “true, fair, and perfect narrative” of all the trials’ (highlighting in the original) (Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018a). As a consequence, while trial accounts became more detailed, the Proceedings’ volumes became more expensive to publish (Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018a) and also slower to produce (Reference DevereauxDevereaux 2007: 19). The readership narrowed down to mostly comprise court officials, and, after full shorthand notes became mandatory to provide an official record in 1907, the publication was discontinued in 1913 (for further details, see Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018a and the succinct summary in Reference WidlitzkiWidlitzki 2018: 52–3).

Despite the Proceedings’ often detailed and convincing speech-like appearance, these glimpses at their complex production and publishing history make it clear that these documents should not be approached uncritically as one-to-one accounts of the courtroom events or even the language used by the speakers. On the whole, the scribes’ input in creating the record must not be underestimated, and the degree to which the published record might correspond to the original speech event remains difficult to assess (Reference Kytö and WalkerKytö and Walker 2003; Reference DotyDoty 2007, Reference Eckert2010; Reference GrundGrund 2007a, Reference Grund2007b). For instance, in one testimony a scribe could attempt to convey the speaker’s use of regional pronunciation features or lexical items but in another record mention is made of the scribe having rendered a speaker’s non-standard ‘broken English’ as standard English (for examples, see Reference HuberHuber 2007: section 3.2.2.2; Reference Traugott, Pahta and JuckerTraugott 2011: 72; Reference WidlitzkiWidlitzki 2018: 59–61). Information was omitted, and ‘the Proceedings are far from comprehensive transcripts of what was said in court’ (highlighting in the original) (Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018b). Comparing the text of the Proceedings with other manuscript and published accounts of the same trials can confirm that ‘what they did report was for the most part reported accurately’ but can also display considerable differences across the various versions (Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018b). Mentions made in newspapers about trials confirm this observation. For instance, the editors of Old Bailey Online refer to a trial in 1787 that, according to The Times, lasted six hours but that only comprises 468 words in the Proceedings (Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018b). Summarizing and omissions were also necessitated by economic concerns, as having everything included in print would have made the Proceedings too costly at the same time as newspapers were encroaching upon the ground and drawing audience from the Proceedings. Even when full shorthand notes were available, only condensed reports would be included in the published version. No original shorthand notes survive; we will return to the different stages the notes went through on their way to the printer (see Section 2.2.2). Systematic omissions, such as those based on the orders given in 1805 to omit argumentation by the defence counsel, and statements by defendants and character witnesses for the defence, have been interpreted as measures driven by the ideology aimed at making the general public trust in the straightforward and uncomplicated nature of the judicial system (Reference WidlitzkiWidlitzki 2018: 54; see also Reference ShoemakerShoemaker 2008: 570–2, 579; Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018b). Yet, against the dearth of audio records of early speech, and within the principles of historical pragmatics, the material does offer ‘the rare opportunity of analyzing everyday language in a period that has been neglected both with regard to the compilation of primary linguistic data and the description of the structure, variability, and change of English’ (Reference HuberHuber 2007: section 1). Such material is particularly well suited for the study of the history of intensifiers, as the texts are based on past speech events involving speakers of various socioeconomic and educational backgrounds recorded in a shared space, often in emotionally charged circumstances; we return to the issues involved in the study of historical speech-related records in Section 2.2.

The Old Bailey Online collection comes with an efficient search engineFootnote ² suited for work in many areas of historical research by way of using words (or keywords) as a gateway to target tagged information on surnames, crimes, and punishment sentences. Searches can be limited to certain periods of time by using information on the month and the year of trials as the search criterion. Importantly, the user is also given direct access to facsimile images of the original printed Proceedings, which is of invaluable help for checking possibly unclear readings in the transcriptions based on the use of optical scanners. However, as Old Bailey Online was never envisaged for linguistic research in the first place, the kind of sophisticated search and post-processing tools required for historical corpus linguistic research (e.g., the word-list function to catch spelling variants and kwic-concordances for screening relevant data) were not built in the search engine. Nor were the Old Bailey Online texts annotated for linguistic or speaker-specific sociolinguistic features which would have facilitated data withdrawal and post-processing. To enable linguistic research on Old Bailey Online, a large-scale corpus based on the trial texts was compiled by Magnus Huber and his associates at Justus Liebig University Giessen, Germany (Reference Huber, Nissel, Maiwald and WidlitzkiHuber et al. 2012; Reference Huber, Nissel and PugaHuber, Nissel, and Puga 2016; http://fedora.clarin-d.uni-saarland.de/oldbailey/).

2.1.3 The Old Bailey Corpus

The 24.4 million speech-based words in the OBC drawn from stretches of spoken language derive from a selection of 637 sessions included in Old Bailey Online; the spoken passages in the OBC were identified and tagged semi-automatically by using customized Perl and Python scripts and human intervention (for details, see Reference HuberHuber 2007: section 4.2 and appendix B; see also Reference Huber, Nissel and PugaHuber, Nissel, and Puga 2016: sections 2 and 5). The first texts included in the OBC date from 1720, as that was when verbatim recording started to gain ground (see Section 2.1.2); the last texts are from 1913, the final year covered by Old Bailey Online (and the Proceedings). Each decade is represented by some 1.1 to 1.6 million words of speech-based text apart from the two opening decades (1720–39), when verbatim recording was still in its early stages, and the last decade represented by only the first four years; for these three decades the word counts remain considerably lower (see Table 4.2 in Section 4.3.2 for overall word counts). The material comprises altogether a good half million individual utterances and ‘constitutes a fairly representative sample of spoken, rather formal LModE in the courtroom setting’ (Reference Huber, Nissel and PugaHuber, Nissel, and Puga 2016: 2).

The value of the corpus is further enhanced by the sociobiographical, pragmatic, and textual annotation added by the compilers for each turn, when such information has been available. The coding scheme adopted thus gives information on the speaker’s gender, age, occupation, and social class, and the person’s role in the trial (a witness, a judge, etc.). It also gives information on the scribe, printer, and publisher of each individual proceeding; reliable information about which reporters were also transcribers at the trials becomes consistently available only after 1773, when names of scribes and what they had been responsible for began to appear in the front matter (Reference HuberHuber 2007; Reference ShoemakerShoemaker 2008). However, the source of this information is unclear, that is, whether it came from the reporter or the printer. The values for the gender (male, female, unknown) and age variables were recovered from the internal clues in the records (e.g., references to titles such as Mr and Mrs) (Reference WidlitzkiWidlitzki 2018: 49–50). For recovering the information on the OBC speakers’ occupations, the compilers resorted to the speakers’ own words or, and mostly for the later trials, to the information added by the scribes; retired informants were assigned the professions they had had earlier on, and women with no information on their professions were assigned their husbands’ profession if known (Reference WidlitzkiWidlitzki 2018: 51). The information on occupations of speakers was based on the Historical International Standard Classification of Occupations (HISCO) (Reference van Leeuwen, Maas and Milesvan Leeuwen, Maas, and Miles 2002), and the micro group name and number assigned to an occupation in this system were included in the corpus annotation as part of speaker characteristics. The HISCO definitions have been used for the HISCLASS system (Reference van Leeuwen and Maasvan Leeuwen and Maas 2011), which comprises twelve social classes extending from ‘higher managers’ and ‘higher professionals’ through ‘medium-skilled workers’, ‘farmers and fishermen’, and ‘lower-skilled workers’ to ‘unskilled farm workers’, to which the OBC compilers added a thirteenth class of ‘unspecified workers’ (Reference WidlitzkiWidlitzki 2018: 50; Magnus Huber, personal communication). This thirteen-level HISCLASS system was applied by the OBC compilers in the form of a conflated seven-level classification. Such intricate thirteen-level or even seven-level classification schemes usually prove too detailed to implement in quantitative corpus studies, and the compilers of the OBC recommend a simplified two-class system with a higher-class category comprising non-manual professions (HISCLASS 1–5) and a lower-class category comprising manual professions (HISCLASS 6–13) (Reference Huber, Nissel and PugaHuber, Nissel, and Puga 2016: 9). This was the system applied when annotating the OBC for social class, with the addition of the unknown category to indicate missing information. The speaker-role classification adopted for the OBC distinguishes seven categories, that is, judge, lawyer, victim, defendant, witness, interpreter, and unknown. We return to this variable in Section 2.2 when discussing the historical pragmatics framework and the early modern courtroom setting.

Overall, even though there were gaps in the sociopragmatic information available for the participants in trials, the compilers of the OBC were able to recover a good deal of it, especially for the gender and speaker-role variables. Some 97 per cent of the words uttered are coded for gender, with female speakers represented by 16 percentage points and male speakers by 81 percentage points; the court officials were male, which partly accounts for the male dominance. The speaker role is also coded for the majority, or close to 90 per cent, of the words in utterances. The social class is known to a lesser extent but, even so, for 64 per cent of the spoken words in the corpus, with 44 percentage points represented by the higher social classes (Reference Huber, Nissel and PugaHuber, Nissel, and Puga 2016: 7; Reference WidlitzkiWidlitzki 2018: 49). The informants with a known value for each of the three variables are of particular interest to us in view of our sociopragmatic analyses. These individuals are 74,445 in number and produce 14,737,838 spoken words in the corpus. Note, however, that the same speaker across trials counts as more than one person, a feature brought about by the automatic coding system applied to the corpus texts.Footnote ³ The sociopragmatic annotation makes the OBC a highly valuable data source for our study of intensifier usage, where we are interested in, among other things, correlating development in intensifier usage and sociopragmatic variability in LModE in its social context.

To facilitate searches, this sociopragmatic and textual information was added in the 637 text files of the OBC by way of automatic and semi-automatic annotation (for details, see Reference Huber, Nissel and PugaHuber, Nissel, and Puga 2016: 8–9). In addition, the texts were annotated grammatically by adding POS-tagging, which further helps in data collection and enables the study of, for example, morphosyntactic and syntactic topics beyond mere lexical searches. We have profited from these annotation features in our data collection. Regarding searches for our intensifiers, we used the plain-text version of the OBC for most of our searches. However, we used the POS-tagged version for retrieving our instances of the extremely frequent forms so and very and the frequent form a little. As known, automatic POS-tagging comes with error rates which can range from minimal to more substantial, especially when the tagger used on historical material was originally designed for present-day language. The compilers of the OBC used the CLAWS tagging suite designed for Present-day English (PDE) for POS-tagging the files but spot-checked the corpus texts where possible for wrongly assigned tags and carried out the necessary corrections (Reference Huber, Nissel and PugaHuber, Nissel, and Puga 2016: 7–8). Even so, some errors likely went unspotted. Where we came across any coding errors of significance to our searches, we carried out manual patching rounds by checking the plain-text version of the corpus.

While advocating the use of historical electronic resources such as the OBC for linguistic study, we need also to mention some of the drawbacks inherent in the corpus linguistic methodology. Compared with corpora containing PDE or even recent English, historical resources tend to suffer from gaps in representation owing to gaps in historical data. In the case of the OBC, we have discussed the degree of consistency and exhaustiveness with which the Proceedings represent the original OBC trials (Section 2.1). Some of the imbalance issues seem unavoidable, considering the sociohistorical circumstances in the late modern period. Even if the OBC is clearly a large-scale corpus, looking for certain terms may show it is small in size for certain purposes just the same. For instance, in the category of maximizers, there were a number of potentially interesting terms represented by one or two relevant occurrences only (e.g., eminently, exceptionally, incomparably, passingly). However, this is often the case in lexical studies where the data tends to cluster around a number of items, leaving a group of items represented by only a handful of examples. The problem with scanty numbers of certain items is that any breakdown across, for example, the time period categories makes the data subsets too small for reliable statistical analyses (for such and other issues in the use of historical corpora for linguistic study, see Reference RissanenRissanen 1989, Reference Rissanen, Lüdeling and Kytö2008b; Reference Claridge, Lüdeling and KytöClaridge 2008; Reference KytöKytö 2011, Reference Kytö, Bergs and Brinton2012; Reference Durrell, Gippert and GehrkeDurrell 2015).

The sociolinguistic features and linguistic production of language users are in focus in historical sociolinguistics (e.g., Reference Raumolin-Brunberg, Nevalainen and Raumolin-BrunbergRaumolin-Brunberg 1996; Reference Conde-Silvestre and JuanConde-Silvestre and Hernández-Campoy 2005; Reference Nevalainen, Raumolin-Brunberg, Hernández-Campoy and Conde-SilvestreNevalainen and Raumolin-Brunberg 2012, Reference Nevalainen and Raumolin-Brunberg2017), consolidated in the aftermath of the advent of sociohistorical linguistics (Reference RomaineRomaine 1982). Within this framework, researchers aim to account for variation and change in language in its social and historical context. Do speakers/writers use language in conservative terms or are they keen to pick up and spread new expressions they have come across, or are they bold innovators themselves who find ways of enriching the linguistic repertoire? For our study of intensifiers, all those avenues are of interest to investigate, innovation being a particularly intriguing area. As mentioned previously, it is only via systematic empirical study of language data that the interplay of the various linguistic and extralinguistic factors influencing processes of change can be mapped reliably. In addition, as our language data derive from spoken interaction in the past, we also need powerful frameworks such as historical sociolinguistics and historical pragmatics aimed at helping one make sense of past language use in its sociolinguistic and sociopragmatic context.

2.2.1 Historical (Socio)pragmatics

Historical pragmatics has been defined as ‘historical linguistics combined with pragmatics’ (Reference Taavitsainen, Susan, Fitzmaurice and TaavitsainenTaavitsainen and Fitzmaurice 2007: 12) and envisaged as an empirical branch of linguistics targeting authentic language use in the past, investigating ‘how meaning is made’ (Reference Taavitsainen, Susan, Fitzmaurice and TaavitsainenTaavitsainen and Fitzmaurice 2007: 13). In other words, historical pragmatics ‘focuses on the linguistic inventory and its communicative use across different historical stages of the same language’, with linguistic variation playing a central role in diachronic form-to-function and function-to-form mapping (Reference JespersenJacobs and Jucker 1995: 3, 13–25). Traditionally pragmaticians study verbal interaction with the aim of uncovering communicative functions and patterns in specific speech situations, but in the 1990s this approach was broadened to cover written texts, and especially speech-related texts as legitimate historical data. Such text can be investigated not just as ‘imperfect renderings of the real thing’ but ‘as communicative manifestations in their own right, and as such […] amenable to pragmatic analyses’ (Reference Jacobs, Jucker and JuckerJacobs and Jucker 1995: 9; see also Reference RomaineRomaine 1982: 14–21; Reference CollinsCollins 2001: 17; Reference Culpeper and KytöCulpeper and Kytö 2010: 9). In our trial texts, face-to-face spoken interactions are presented as written texts, mediated by the scribe and the printer from shorthand notes. This results in a functionally rich input for analyses: not only does the original discourse in the speech situation have its functions but another layer of interactive functions is there between the record texts and their readers (Reference Culpeper and KytöCulpeper and Kytö 2010: chapter 3). Even if we only make use of the original discourse level in our analyses, the multilayer approach can be expected to afford ample opportunities for future work in the field.

As with historical sociolinguistics, the consolidation of historical pragmatics was closely connected with the advent and spread of historical language corpora and corpus linguistic techniques enabling relatively fast data collection and post-processing. To start with, corpus linguistic techniques were mostly used to retrieve lexical and morphosyntactic elements in electronic texts, and historical pragmaticians were not very keen on applying the potential of this methodology to the study of dialogues and dialogic features in historical texts (Reference Jucker, Fritz, Lebsanft, Jucker, Fritz and LebsanftJucker, Fritz, and Lebsanft 1999: 19). But since the mid-1990s, as demonstrated by, for example, many of the contributions in Reference JuckerJucker (1995), the systematic study of historical speaker–hearer and writer–reader interaction has benefited from the use of electronic resources (Reference Jucker, Fritz, Lebsanft, Jucker, Fritz and LebsanftJucker, Fritz, and Lebsanft 1999: 13–20; Reference Rissanen, Lüdeling and KytöRissanen 2008b: 54). The rationale of this approach is that hypotheses about the spoken interaction of the past can be formulated and tested on texts which obviously convey conversational or oral features in notes taken down in authentic speech situations or even in fictive speech found in literary texts; comparisons across genres representative of these texts further help to assess what past spoken interaction was like in its sociocultural contexts (Reference Rissanen and LassRissanen 1999a: 188; Reference Rissanen, Lüdeling and KytöRissanen 2008b: 60–1; Reference Culpeper and KytöCulpeper and Kytö 2010; Reference Kytö, Jucker and TaavitsainenKytö 2010).

Various models have been developed to depict the complex relationship between spoken and written language. In his foundational 1988 study of PDE spoken and written texts, Biber presented empirical corpus-based evidence that linguistic variation across speech and writing should not be approached from a dichotomous perspective. While Biber used factor analysis in his study of dimensions of variation, Reference Schneider, Chambers, Trudgill and Schilling-EstesSchneider (2002) took the proximity of a genre to speech as his starting point in his classification of early American English texts and distinguished ‘recorded’ (trial records, interview transcripts), ‘recalled’ (ex-slave narratives), ‘imagined’ (letters by semi-literate writers, diaries), ‘observed’ (contemporaries’ statements) and ‘invented’ (fictitious utterances by fictitious speakers) categories (Reference Schneider, Chambers, Trudgill and Schilling-EstesSchneider 2002: 72–3). Among the very influential models has been the one proposed by Reference Koch, Jucker, Fritz and LebsanftKoch (1999) and Koch and Oesterreicher (1985–1986, Reference Koch and Oesterreicher1990), which envisages this relationship as a continuum with phonic-medium features at the one end and graphic-medium features at the other. The language of ‘communicative immediacy’ is characterized by physical immediacy, privacy, familiarity (or intimacy) of partners, high emotionality, deictic immediacy, dialogue, communicative cooperation, and spontaneity, while the language of ‘communicative distance’ displays the opposite features. Yet, the language of immediacy is not restricted to the phonic code but may also occur in the graphic code areas, and similarly, the language of communicative distance may encroach in the areas of the phonic code (for a summary and discussion, see Reference Jucker, Di Martino and LimaJucker 2000: 19–25; Reference Culpeper and KytöCulpeper and Kytö 2010: 10–12).

There are of course differences in the degrees to which some written texts come closer to spoken language than others, and most speech-based genres such as courtroom proceedings, debates, town meeting records, public speeches, and sermons display a mixture of communicatively distant and immediate features. As for the Proceedings, in addition to dialogues proper, rendered in the form of question–answer adjacency pairs, there are also dialogues where the questions have been omitted and where we have rows of answers segmented by hyphens or dashes. Within constructed speech presentation, a playwright or the author of a novel may have mimicked colloquial, everyday speech in the utterances assigned to a character, while within authentic speech presentation, a real speech event underlies transcribed speech and may convey colloquial features in, for instance, the words reported by a witness informing the court of a previous speech event (Reference Biber, Finegan, Rissanen, Ihalainen, Nevalainen and TaavitsainenBiber and Finegan 1992: 689; Reference Kytö and WalkerKytö and Walker 2003: 222–4; Reference ArcherArcher 2005: 10–11; Reference Culpeper and KytöCulpeper and Kytö 2010: 17–18). The communicative distance approach has been applied in empirical terms to the genre classification adopted for an investigation of EModE dialogues and resulted in a three-way categorization of speech-related texts distinguishing speech-based texts (e.g., trial proceedings), speech-purposed texts (e.g., plays), and speech-like texts (e.g., personal correspondence) (Reference Culpeper and KytöCulpeper and Kytö 2010: 16–18). In this classification our late modern trial records in the OBC fall in the category of speech-based texts, and, while being limited to the graphic code, they can nevertheless be expected to convey features of communicative immediacy from various points in the scalar speech–writing continuum. Among such features likely to be found in the OBC texts are, for instance, rapid turn taking in the question–answer adjacency pairs, and the use of emotionally charged expressions, pragmatic markers, and, interestingly, intensifiers. Further ‘involved’ features include first- and second-person pronouns, direct wh-questions, that deletion, pragmatic and modality markers, and private verbs (cf. Dimension 1, the ‘Involved versus Informational Production’, in Reference BiberBiber 1988). Example (1), with relevant features (apart from that deletion) given in italics, gives a brief impression of this:

1. (1)

   Mr. Phipps. Although I am prosecutor in this trial, I can give but little light into it. I can only inform the court I did suspect the prisoner to be with child, and very strictly charged her with it.

   What was she?

   Phipps. She was my servant.

   Q. Can you tell the time you charged her with being with child?

   Phipps. I can’t exactly tell the time. I believe it might. be about two months before she was brought to bed.

   Q. What was her answer?

   Phipps. She very boldly denied it, and assur’d me with great truth and veracity it was a very unjust censure; she shew’d no signs of guilt at all. She answer’d in such a manner that I was really a good deal inclined to think she was innocent.

   Q. Did she appear big?

   Phipps. She did, which was the cause of suspicion. She told me it was a distemper, and was inclinable to think it was the dropsy. (t17570526-22)

If historical pragmatics, broadly speaking, aims to describe conventions of language use – and especially characteristics of spoken interaction – and to account for changes in them across time in past communities (Reference Jacobs, Jucker and JuckerJacobs and Jucker 1995: 6), its sub-branch historical sociopragmatics ‘concerns itself with any interaction between specific aspects of social context and particular language use that leads to pragmatic meanings’, with focus on ‘language use in its situational context, and how those situational contexts engender norms, which participants engage or exploit for pragmatic purposes’ (Reference Culpeper, Jucker and TaavitsainenCulpeper 2010: 76). For historical sociopragmatic analyses, the sociohistorical, sociocultural, and cognitive contexts influencing the interaction are of importance (Reference ArcherArcher 2005: 7, with reference to Reference ArcherArcher 2002: 3 and Reference Raumolin-Brunberg, Nevalainen and Raumolin-BrunbergRaumolin-Brunberg 1996: 11). Our courtroom context and access to information on speakers’ sociolinguistic properties and their roles allow us to highlight intensifier usage from the sociopragmatic perspective.

2.2.2 The Courtroom Setting

The courtroom context and trial practices as such are worth some closer scrutiny, as they constrained the circumstances where spoken interaction took place in the courtroom. In particular, the power structure among defendants, witnesses, victims, and judges and other court officials influenced the range of manoeuvring available to speakers. Crucially, over the nearly 200-year period included in our study, there were changes not only in the publication practices of the Proceedings (see Section 2.1) but also in the make-up of the physical courtroom, the session procedures, and the roles of the interactants. We find it illuminating to look into changes in this physical space more closely. The Old Bailey was rebuilt a number of times over the period covered by the Proceedings. In 1673 an open-air building solution was opted for, and the ground floor housing the courtroom was open to weather so as to prevent infectious diseases from spreading. In 1737, the ground floor was again fronted and the courtroom enclosed. The following description of the seating and other arrangements is given at the Old Bailey Online website (for sources, see Reference HowsonHowson 1970; Reference JacksonJackson 1978; Reference RumbelowRumbelow 1982).

Changes responding to new requirements in how trials were conducted in the courthouse meant continuous albeit slow changes to its design. In the 1774 reconstruction project the access and view of the general public to the courtroom was further restricted by a brick wall built in front of the building. A table was added for the lawyers to plead from, indicative of the increasing role of the counsel, and in 1824, a second courtroom seating for, among others, attorneys, counsel, prosecutors, witnesses, law students, officers of the court, and spectators. The new and larger courthouse built in 1907 was able to provide separate rooms for male and female witnesses, lawyers, and barristers’ clerks as well as for witnesses of upper echelons of society. These building-expansion projects reflected the increase in the number of trials and in their length (Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018c).

Indeed, early trials were quick and only lasted half an hour on average. When the death penalty was abolished for many crimes in the 1820s, they could last even a shorter time; according to a contemporary estimate recorded in 1833, a trial took no more than eight-and-a-half minutes on average (Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018d). Trial procedures were under continuous transformation regarding the legal status, rulings, and judicial discretion applied. The pace of change quickened from the early nineteenth century onwards (Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018d). Until the last decades of the nineteenth century, lawyers were seldom present in criminal cases, and it was up to the defendant to explain away the evidence brought against them. Despite the presence of the counsel becoming more common at the turn of the nineteenth century, defendants remained at a disadvantage. They had no right to remain silent and until the early 1800s, they were expected to establish their innocence on their own. They were not informed beforehand of the kind of evidence to be brought against them, so they could not prepare their cases in this respect. They conducted cross-examinations of witnesses, along with the judges, up until the practice of using defence lawyers became more common; it was only as of 1836 onwards that defence lawyers were allowed to address their summary of the case to the jury. The general source of evidence was witness testimony, which must have made the language in it an even more important tool in the past courtroom than it is in the modern. While prosecutors could call in their witnesses and compel them to testify, defendants, who could call in their own witnesses from 1702 onward, were not able to force their witnesses to attend trials (Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018d). Indeed, witnesses’ testimony comprises the largest portion of the speech-based utterances found in the OBC, altogether 59 per cent in version 2.0 (we report on the shares contributed to our data by our speaker groups in Chapter 4); an illuminating discussion of the speaker groups in version 1.0 of the OBC (2012) can be found in Reference WidlitzkiWidlitzki (2018: 66–73), where this version of the corpus was used as the main data source.

If defendants were in the underdog position in the late modern courtroom, judges were the driving force, examining the accused and witnesses and summing up the cases. Even though much of what they said was abridged or omitted in the Proceedings, they clearly played a leading role in the courtroom; when their contributions to the discussions were recorded, they were referred to as ‘The Court’. The increasing participation of lawyers on both the prosecution and defence sides meant changes to the role of the judges, who were charged with arbitrating between them and providing expertise in settling legal arguments. In the eighteenth and early nineteenth centuries, reforms were carried out with the aim to rectify the power balance in the courtroom so as not to have it weigh so heavily against defendants (Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018d).

A further indication of imbalance between speaker groups in the late modern courtroom power structure was the limited participant roles assigned to women. As judges, lawyers, jury, interpreters, and court officials were men, women were represented only in the roles of defendant, victim, and witness; they could also attend the trials as spectators in the gallery. Women’s testimony was more likely to be considered less reliable than the evidence presented by men, and it could be omitted from the Proceedings more easily than men’s testimony (Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018e). Giving testimony has also been assumed to have been an intimidating experience for women in the noisy and male-dominated courtroom (Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018e; Reference WidlitzkiWidlitzki 2018: 70). Some linguistic evidence pointing to women and men witnesses having been treated differently has been presented for early modern trials where male witnesses were exposed to heated cross-examination and women witnesses were construed as narrative-givers, expected to comply as the less-empowered party (Reference Culpeper, Kytö, Kastovsky and MettingerCulpeper and Kytö 2000: 81). Women also appear significantly less as victims or prosecutors of crime in the Proceedings: as husbands were the owners of marital property, they were also the ones to be the main parties in cases of theft; theft in all its various forms was the most frequent offence prosecuted across the full Proceedings collection. Nor did women usually carry weapons that they could have used to harm their fellow citizens if provoked (Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018e, Reference Emsley, Hitchcock and Shoemaker2018f). These circumstances are quantitatively reflected in the OBC, where only some four million words out of the 24.4 million words in the corpus are assigned to female speakers.

Yet another source of imbalance in the OBC is the speakers’ social class. Compared with the speaker-role and gender parameters, social class is elusive and often non-recoverable information (almost nine million words in the OBC have no class coding for speakers). As a consequence of higher-class speakers (HISCLASS 1–5 in the thirteen-tier scheme) mostly being higher-class men, lower-class speakers (HISCLASS 6–13) remain underrepresented (for numerical information, see Chapter 4). The informants coded for all three sociopragmatic parameters, that is, gender, speaker role, and social class, comprise half of the speakers. These parameters are heavily intertwined, with representation in one group being dependent on representation in another group. Overall, the power structure embedded in the late modern courtroom dynamics not only gave the upper hand to judges and lawyers as against the other participants in trials but also left women and lower-class speakers underrepresented in the Proceedings. It must also have influenced spoken interaction in the courtroom and the speech habits of individual participants in the trials.

What could we conclude from the above regarding approaches to the study of spoken interaction in late modern courtroom? Considering the speaker roles and the seriousness of the subject matter at hand, interaction in a courtroom must have been, and still is, necessarily formal in nature, owing to procedural and institutional practices. However, as pointed out, this does not need to exclude the presence of language of communicative immediacy. Many of the cases dealt with criminal acts related to everyday events and can be expected to prompt spontaneous, uninhibited language and instances of colloquial usage, especially by the substantial number of lay or non-professional speakers; conversely, the professionals’ language, which could be expected to be marked more by distance, remains underrepresented due to omissions. During cross-examinations, rapid exchange of turns may give reason for heated emotions, while testimonies delivered in a conversational manner may encourage personal engagement and efforts to evoke and convey the immediacy of the events and acts once experienced in the past. A defendant or a witness wishes to appear as a credible person, and for that purpose language features such as intensifiers can conveniently be used to play up or down circumstances and the role the person played in the crime in question. The interlocutors find themselves in the same physical locality and their dialogues can be both governed by a question–answer frame and include spontaneous contributions to the volume of speech.

However, multiple levels of speech presentation are usually involved, covering the dialogue between the interactants, the speaker’s reporting of earlier speech events, and the scribe recording what he was able to or chose to take down from the flow of speech mostly in shorthand (Reference HuberHuber 2007: section 3.2; Reference Culpeper and KytöCulpeper and Kytö 2010: 51–2; Reference Emsley, Hitchcock and ShoemakerEmsley, Hitchcock, and Shoemaker 2018b). Further levels are added to the setting when these notes were then developed into proceeding text and sent to the printer for subsequent processing. Huber distinguishes a number of consecutive stages from the speech event during a trial at the Old Bailey to the printed Proceedings (Reference Huber, Helbig and SchalleggerHuber 2010: 344–5; see also Reference HuberHuber 2007: section 3.2.1, and Reference WidlitzkiWidlitzki 2018: 59–60):

The language used in the original speech event may well have undergone changes during each of these stages, and, in the end, we do not know for certain how exactly the text was processed through the different stages. For instance, Huber’s studies of Thomas Gurney’s influential shorthand system based on his Brachygraphy or short-writing (1752, with multiple editions and reprints down to the nineteenth century) showed that while certain speech features could well have been taken down reliably, no mention was made of some others (e.g., the differentiation between the negative forms cannot and can’t). Yet, as no original shorthand manuscripts have been found to enable comparisons with the printed Proceedings, it is impossible to draw conclusions on the accuracy of speech presentation over those stages of production (Reference HuberHuber 2007, Reference Huber, Helbig and Schallegger2010). Some evidence of the accuracy of the records can be found in the scribes’ responses to questions asked by, for example, the judges regarding the formulations taken down from the witnesses’ and other speakers’ mouths, but, on the whole, such commentary is rare (for further discussion on the role of scribes, see Section 9.2.1 in the present volume). What has been considered to add to the accuracy of the written record as the representation of the original speech event has been ‘the temporal distance between the speech event itself and the time of recording’ (Reference Schneider, Chambers, Trudgill and Schilling-EstesSchneider 2002: 72). The sooner the written record can be finalized, the less danger there is for memory loss and other such damage to interfere. In the case of the Proceedings, according to Reference HuberHuber (2007: section 3.2.1), the expanding of the shorthand notes must also have followed shortly after, as the Proceedings were usually published within a fortnight to a month after every session (for further details and discussion, see Reference HuberHuber 2007, Reference Huber, Helbig and Schallegger2010).

Features of communicative immediacy may have been lost at various stages of the original speech being transcribed and prepared for publication, and features of communicative distance were likely introduced by the recorder or the printer to follow various recording conventions (Reference Kytö and WalkerKytö and Walker 2003; Reference Culpeper and KytöCulpeper and Kytö 2010). Linguistic evidence has been presented supporting the impression one has of a standardization process underway across the history of the Proceedings; for instance, Huber in his Reference Huber2007 and Reference Huber, Helbig and Schallegger2010 studies on the rates of contractions in the OBC and the Corpus of English Dialogues 1560–1760 (CED) concluded that there was a shift to an increasingly formal recording style after 1760 under the growing control by the City of London in the mid-eighteenth century. Consequently, our intensifier data is available to us in the court record, which largely follows the conventions of communicative distance while at the same time containing a report of instances of spoken interaction in the courtroom offering opportunities for language of immediacy to emerge in the flow of the discourse (Reference Jucker, Di Martino and LimaJucker 2000: 24–7; Reference ArcherArcher 2014: 261).