A quantitative study of spelling variation in William Caxton’s printed texts

This thesis presents new quantitative methods for the exploration of spelling variation in early printed texts, and applies these methods to the books published by William Caxton (c. 1421-1492). Combining the study of historical spelling variation with methods from computer science and mathematics, this work presents the first use of spellings as data for the exploration of Caxton’s editorial approach to his texts and the early printing processes used in his printing house. I examine the use of orthographic variation at the level of idiolect for researching fifteenth-century print, and focus on the spellings of the individuals involved in the printing process.
This research takes a bottom-up quantitative approach to Caxton’s texts, and the methods developed in this thesis are designed to simultaneously analyse all the spellings contained in any given text. In Section One I establish through quantitative analysis that the printing process has no influence over the orthography of the printed text, and from that point onwards this thesis focuses on the spellings used by the individuals whose language can be found in the printed texts—the compositor who set the type and the scribes who wrote the copy text. Building on this argument, Section Two develops and tests methods for the exploration of the orthographic idiolects that are used in printed texts. In Section Three I apply these methods to spelling data drawn from a selection of Caxton’s texts, and in doing so explore the layers of spellings present in the texts to develop our understanding of Caxton’s practices as an editor and translator.
This thesis demonstrates the value of focusing on idiolects when researching orthography in early printed texts, and expands our knowledge of the extent to which spelling variation operates at the level of the individual in the fifteenth century. The innovative methods I have developed for use with spelling data in this thesis—cluster analysis and similarity measurements—demonstrate the value of quantitative methods in exploring historical texts and historical spelling variation, and provide avenues of research into historical spelling variation using quantitative methods. Through exploring the use of data-driven quantitative approaches to analysing spelling variation in early print, this thesis provides new insights into our knowledge of historical spelling variation, the quantitative study of spelling in early print, and the history of the early printed book.