Mediumship and stylometry: Exploring a new way of attributing authorship to mediumistic writings

Abstract

Stylometry is still largely unexplored for applications in attributing authorship to mediumistic writings, as outlined in the subsequent literature review. Stylometry is a statistical method for analyzing authorship with a history to detect anonymous authors, e. g., J. K. Rowling as the real author of The Cuckoo’s Calling. Modern stylometry transcends literature analysis, applying to forensic inquiries, such as plagiarism detection, online security, and group identity characteristics. Historically, stylometric research unfolds in three phases: focusing on single numerical functions, statistical methods based on word frequencies, and contemporary use of machine learning. Authorship analysis includes three primary objectives: authorship attribution, verification, and characterization. It uses mathematical linguistics, information theory, and linguistic features like vocabulary richness, word frequency, and syntactic structures. Stylometric methods reliably detect authorship in imitation scenarios (pastiche) and have now also led to adversarial stylometry, which facilitates and identifies attempts at anonymization. Distribution patterns of features such as function words or punctuation can only be detected with high computing power and appear to contain too little content or emotion to be recognized by psi abilities of mediums. Imitation on a stylometric level has not yet been achieved by professional writers and does not appear to be one of the abilities of savants either, whose left hemisphere, including speech and language, is the least developed, which is why such genuine talent cannot be assumed in mediums either. Due to its undeniable successes, its quantitative and computational approach, and its potential to investigate a wide range of research questions, stylometry therefore appears to be a desirable auxiliary science for mediumship research.