In 2013 the crime novel The Cuckoo's Calling appeared, by an unknown debut author named Robert Galbraith – until a computer-aided text analysis unmasked the real author: J. K. Rowling. The method behind it is called stylometry, and in the Zeitschrift für Anomalistik the lawyer Adrian Weibel has worked out a proposal as obvious as it is bold: apply this tool to mediumistic writings – texts that mediums say they received from the deceased. For if the authorship of a text can be measured, then in principle it can also be measured whose style resides in a dictation from the beyond.
The stylistic fingerprint
Stylometry is statistical authorship analysis: it measures not what a text says but how it says it – the distribution patterns of function words, punctuation, word frequencies and syntax. The field's basic premise: every author carries a quantifiable "stylistic fingerprint" that lies beyond conscious control. The successes are considerable – from the Rowling unmasking to attribution questions around Shakespearean collaborations and Frankenstein, to corpus studies of Victorian literature with accuracy rates around 75 to 85 percent; today the field works with machine learning and also serves plagiarism and forgery detection.
Why this matters for the afterlife question
Weibel's core argument is the imitation finding: at the stylometric level, imitation has so far not been achieved even by professional writers – pastiches, deliberate style imitations, are reliably exposed by the analysis. Savants also drop out as an explanation, since their exceptional abilities least concern the language domain. This gives the tool its edge in both directions: if a mediumistic text stylometrically matches the medium, earthly authorship is practically proven. But if it matched – cleanly tested – the deceased author rather than the medium, the imitation explanation would face a problem that, on present knowledge, no talent can solve. Exactly the kind of testable criterion this research keeps looking for.
The old tradition: style comparison by eye
Qualitative style comparisons have existed for over a century. Arthur Conan Doyle analysed the mediumistic texts of three "communicators" in 1927: in the medium Hester Dowden's Psychic Messages from Oscar Wilde he found Wilde's colour adjectives, his sentence rhythm and the unmistakable cynicism ("Being dead is the most boring experience in life…"); Dowden herself noticed writing quirks such as interspersed splittings (d-eath, vin-tage) that were later also found in Wilde facsimiles. Conversely, F. W. H. Myers showed as early as 1898 that the famous "Julia" messages of the journalist W. T. Stead read like a patchwork of Stead's own magazine prose. Weibel names the problem of this tradition clearly: conspicuous quirks – chapter titles, nicknames, colourful adjectives – can be copied by a good imitator. Only the inconspicuous patterns nobody consciously controls carry evidential weight. And those are precisely what stylometry measures.
The hard tests so far – honestly reported
Computer-aided testing has happened only twice so far, and both times the result went against the mediums. Goto (2018) tested the famous "afterlife sequel" to Charles Dickens' unfinished The Mystery of Edwin Drood, published in 1873 by the printer Thomas P. James as a dictation from the dead Dickens: stylometrically the sequel is clearly not Dickens – it even lies closer to Thackeray. Schöberlein (2017) examined poems the medium Lizzie Doten attributed to the dead Edgar Allan Poe: they were unambiguously attributed to Doten herself – "Poe's author signal", the dry conclusion runs, "apparently died with him". Anyone coming from our article on the stylistic analysis of Xavier's psychographies can see: the method can cut – and it does not automatically cut in the mediums' favour. As a curious counterpoint, the study by Saad and de Medeiros (2022) deserves mention: four automatically written signatures of deceased communicators resembled their lifetime signatures strikingly – in a script that did not correspond to the medium's natural handwriting.
Weibel's programme – and its limits
Weibel proposes stylometry as an auxiliary science of mediumship research – alongside the established approaches, from blinded readings à la Beischel to the cross-correspondences. He fairly names the opposing position too: the skeptic Joe Nickell dismissed stylometric medium analyses in 2007, arguing that statistical style differences fluctuate from passage to passage. And there are practical hurdles – one needs sufficiently large text corpora of the deceased and of the medium, ideally in the same language and genre. But there are also obvious candidates: the vast œuvre of Chico Xavier, for instance, with hundreds of books in the styles of dozens of deceased Brazilian authors, would be a dream corpus for a modern stylometric examination.
The most honest limit, however, is this: the paper is a proposal, not a piece of research – not a single new corpus was analysed, no medium tested, no result produced. Telling for the state of production of German-language anomalistics is also who supplied the most interesting methodological impulse of recent ZfA volumes: not the full-time, foundation-funded research enterprise, but a lawyer publishing on the side. The actual work – such as putting Xavier's corpus to a real stylometric test – still waits for someone to do it.
Context
The finest thing about Weibel's proposal is its symmetry: stylometry is no tool of believers and none of skeptics – it is a referee who has so far whistled twice against the mediums and might whistle differently next time. A field that submits to such tests instead of avoiding them wins either way: either further pious legends fall – or one day a text is found whose unconscious fingerprint belongs to a dead man. Both would be a gain in knowledge. And that this proposal appears in the Zeitschrift für Anomalistik fits everything we have written about it: dispute as a format – this time as an invitation to the data themselves.
Sources: Adrian Weibel: "Mediumship and Stylometry. Exploring a New Way of Attributing Authorship to Mediumistic Writings", Zeitschrift für Anomalistik / Journal of Anomalistics 24 (2024), no. 1, pp. 55–79, DOI 10.23793/zfa.2024.055, full text (PDF) (containing the studies discussed: Goto 2018 on the Drood sequel; Schöberlein 2017 on the Doten/Poe poems; Saad & de Medeiros 2022 on the signatures; Doyle 1927; Myers 1898; Nickell 2007); supplemented by our articles on the stylistic analysis of Xavier's psychographies, Chico Xavier, Julie Beischel and the cross-correspondences.
For more, see our curated knowledge collection and the research link directory.