Measuring the evolution of a revised document

Abstract

By analyzing two drafts of a single written piece, we open windows into that document's evolution that may not be knowable otherwise. However, existing document comparison tools generally do not facilitate scientific inquiry, as they are generally low-throughput and lacking in visual accessibility. Here, we introduce sequence homology analysis (SHA) as an alternative approach to measuring changes between two documents. SHA is a technique common to molecular biologists viz. studies of amino acid- and DNA sequences, and has been extensively validated. Whereas there is no known application of sequence homology analysis in writing researches, we give overview to its implementation and interpretation, and present a novel algorithm which incorporates SHA into the study of any number of documents in a semi-automated fashion. Additionally, we propose a method for visualization based on standard network analytic conventions. We illustrate SHA, the algorithm, and the network visualization via a publicly accessible dataset of historical significance: consecutive drafts of United States President Dwight D. Eisenhower's farewell speech (EFS). Additionally, we describe the parameterization of this routine, its potential for further automation, and its extension into other areas of writing research.