Background

Corpus and Data Collection

To develop an accurate assessment of an individual’s behaviour, and the pressures they experience, detailed social information is vital, as is a sufficiently broad coverage of each individual’s output as well as the community of which it forms part. To meet these needs, we make use of the Early Modern Multiloquent Authors corpus (EMMA), a 90-million word specialized corpus of prolific authors of Early Modern English (Petre et al. 2019). Comprising 50 authors spanning five generations of people born in the 17th century, the corpus is intended to be an in-depth collection of the published work of individuals writing in the Early Modern English period. In the present research, we focus on the language of 25 authors from the first three generations born between 1599-1616, 1621-1631, and 1635-1644, respectively. Their details are displayed in Table 6.1. The size of our dataset is amenable to statistical analysis of change in the lifespan and comparison between individuals’ usage over time. Detailed metadata of life events, genre and background information for all of our authors allow for an assessment of the interaction between these factors and individual linguistic behaviours. Our data span a period of 1623-1726, henceforth referred to as “our period”.

The data were searched using a regular expression capturing /Y-cleft structures of varying lengths and types and spellings. Relative markers were included that were separated by up to eight words from the copula. Data collection consisted of annotating the resultant hits (a total of 4,452 for the 25 authors discussed for this study) manually for a number of dimensions of variation, as well as excluding 20,263 instances of noise. Noise includes related constructions, primarily extraposition

(Continued)

sentences. The data were then analysed qualitatively as well as quantitatively, making use of Cosycat (Collaborative Synchronized Corpus Analysis Toolkit, Manjavacas and Petre 2017) to annotate the original texts, Excel to examine the resultant annotated corpus data, and R (R Core Team 2013) to conduct the Kendall Tau tests utilized in this chapter.