Early Modern Multiloquent Authors (EMMA): Designing a large-scale corpus of individuals’ languages

Abstract The present article provides a detailed description of the corpus of Early Modern Multiloquent Authors (EMMA), as well as two small case studies that illustrate its benefits. As a large-scale specialized corpus, EMMA tries to strike the right balance between big data and sociolinguistic coverage. It comprises the writings of 50 carefully selected authors across five generations, mostly taken from the 17th-century London society. EMMA enables the study of language as both a social and cognitive phenomenon and allows us to explore the interaction between the individual and aggregate levels. The first part of the article is a detailed description of EMMA’s first release as well as the sociolinguistic and methodological principles that underlie its design and compilation. We cover the conceptual decisions and practical implementations at various stages of the compilation process: from text-markup, encoding and data preprocessing to metadata enrichment and verification. In the second part, we present two small case studies to illustrate how rich contextualization can guide the interpretation of quantitative corpus-linguistic findings. The first case study compares the past tense formation of strong verbs in writers without access to higher education to that of writers with an extensive training in Latin. The second case study relates s/th-variation in the language of a single writer, Margaret Cavendish, to major shifts in her personal life.