When a single person writes over 450 books across seven decades, attributing them to dozens of different deceased authors — and when these texts show remarkable stylistic consistency within each "spirit author" while clearly differing from one another — an empirically tractable question arises: can this be explained by that person's ordinary abilities? This is precisely the question that Brazilian forensic document examiners and linguists have tried to answer, methodically, in the work of Chico Xavier.
A Unique Corpus
From a scientific point of view, Chico Xavier is a special case. Few mediums leave behind a textual record large enough for statistical analysis; Chico left over 450 books, plus thousands of handwritten letters to bereaved family members. These texts are not anonymous but systematically attributed:
- Emmanuel — Chico's mentor spirit, philosophical-religious works, from 1931 to 2002
- André Luiz — a supposedly deceased physician, a series of books on life in the spirit world with medical and anatomical detail
- Humberto de Campos / "Irmão X" — the Brazilian writer who died in 1934, ironic, elegant literary prose
- Hundreds of individual spirit authors in the letters to bereaved families — deceased children, young adults, the elderly, each with their own diction and biographical detail
This makes Chico's output, unlike that of most mediums, genuinely accessible to methodical investigation — through both language and handwriting.
Graphoscopy and Stylometry — What Is Examined?
The two disciplines differ:
- Graphoscopy (or forensic graphology) is the scientific-comparative study of handwriting, as applied in criminal proceedings to verify wills or signatures. It compares features such as line tracing, pressure, letter form, connecting strokes and rhythm of writing.
- Stylometry is the statistical analysis of linguistic style features (word frequencies, sentence length distribution, function word use). The same procedures are used today to assign anonymous historical texts to an author or to detect plagiarism.
Both methods ask the same basic question: does this text/handwriting come from the stated person — or from someone else?
Perandréa's Graphoscopy Study (1991)
The most important single finding comes from Carlos Augusto Perandréa, a court-recognized Brazilian forensic document examiner with many years of practice. In 1991 he published the book A Psicografia à Luz da Grafoscopia ("Psychography in the Light of Graphoscopy"), in which he applied the methods of his discipline to a corpus of psychographed letters.
His procedure was simple and consistent: he obtained original handwriting samples of deceased people from their lifetimes — letters, diaries, signatures on documents — and systematically compared them with the corresponding passages from Chico's psychographies attributed to those deceased. He also analysed Chico's own private handwriting as a third reference.
His finding: the psychographed samples in a number of cases showed substantially higher agreement with the deceased's lifetime handwriting than with Chico's private hand. This is noteworthy above all because a simple imitation can be recognized by a trained examiner through characteristic slowdowns, pressure fluctuations and "construction marks" — features that were absent in the psychographed writings. Instead they showed the fluent, automated ductus that an experienced examiner expects only from lifelong, habitual handwriting.
"The psychographed writings examined show the characteristic features of authentic, habitual handwriting — not those of an imitation." — Carlos Augusto Perandréa, paraphrased from A Psicografia à Luz da Grafoscopia
Perandréa was not the only examiner to reach such conclusions; comparable expert reports were occasionally produced by other Brazilian graphologists. A systematic international replication, however, has yet to appear.
Stylistic Consistency Across Decades
In addition to the handwriting, the stylistic consistency is striking:
- Emmanuel remains faithful to the same calm, didactic-doctrinal tone for 71 years — from the first works of the early 1930s to the last of the early 2000s. Theological categories, imagery and sentence structure are stable across decades.
- The André Luiz series (16 volumes, beginning with Nosso Lar, 1944) follows a technical-descriptive diction with borrowings from anatomy, physiology and pharmacology. The vocabulary of this series differs measurably from that of the Emmanuel works.
- The Humberto de Campos texts (later "Irmão X") show the ironic, light short prose of the original writer; Brazilian literary critics considered them stylistically credible.
- The letters to bereaved families shift tone with every supposed sender: an eleven-year-old girl writes differently from an elderly grandfather, a student differently from a soldier.
This range of voices over such a long period is also explainable without supernatural assumption — for example through extraordinary literary talent and consistently sustained "character theatre" — but it is at least an empirical observation that belongs in an overall assessment.
The André Luiz Case: Medical Knowledge Beyond Primary School?
Chico Xavier attended school only to the primary level and worked all his life in modest jobs (a weaving factory, a grocery store, later as a secretary). Yet the André Luiz series contains passages with medical terminology, anatomical descriptions and physiological concepts that were by no means freely available in the 1940s and 1950s.
Proponents see this as a hint at a knowledge source beyond Chico's ordinary abilities. Critics rightly point out that:
- Chico was an unusually attentive observer and self-taught reader (a regular user of the spiritist library in Pedro Leopoldo);
- some of the medical terms remain vague and could be picked up from textbooks without specialist training;
- the medical language of the series occasionally contains inaccuracies that a practising physician would presumably not produce.
A strict quantitative stylometric analysis comparing the vocabulary of the André Luiz volumes against contemporary specialist literature has yet to be undertaken. The material is there — the investigation is not.
Forensic Application: Court Cases
The forensic relevance of graphoscopy and stylometry shows in the fact that psychographed texts were actually admitted as evidence in Brazilian courts:
The Humberto de Campos lawsuit (1944) is described at length in the main blog on Chico Xavier: the deceased writer's family sued Chico and the FEB to establish authorship (and thus royalty rights). Judge Joao Frederico Mourao Russell ruled in 1944 that the state could not authoritatively determine whether a spirit had written a book or not — a careful, methodologically well-founded declaration of non-competence. Chico subsequently attributed the volumes in question to the anonymous "Irmão X".
The Goiânia case (1979) is forensically even more interesting: in the murder trial of José Divino Nunes, a letter psychographed by Chico, supposedly from the victim Maurício Garcez Henrique, was filed as evidence. The letter described the death as an accident and exonerated the accused. Judge Orimar de Bastos admitted the letter as evidence — remarkable in a secular legal system — and acquitted the accused. The verdict triggered international debate; it is one of the few modern examples of a secular criminal court incorporating a psychographic document, at least as a contributing factor, into its weighing of evidence.
Critical Assessment
What do graphoscopy and stylometry actually show — and what not?
They show an empirical puzzle that leaves at least three explanations open:
- Extraordinary literary talent: Chico possessed a rare combination of mnemonic capacity, linguistic mimicry and long-perfected character consistency.
- Unconscious dissociation: different personality fragments producing stylistically distinct texts, comparable to dissociative phenomena — although Moreira-Almeida's studies argue precisely against pathological dissociation in spiritist mediums.
- External source: the texts indeed originate from deceased personalities — the hypothesis the Spiritists themselves work with.
Which of these explanations one prefers depends on one's prior framework. What remains noteworthy is that the data — contrary to what is often claimed — constitute a serious empirical trail and cannot be dismissed by a blanket appeal to fraud or self-deception. The SPECT study by Peres, Moreira-Almeida and Newberg has additionally shown that the neurophysiological state during psychography differs measurably from normal writing — which connects to the question of the relationship between consciousness and brain.
What Remains to Be Done
A systematic, computational stylometric study of the entire Chico Xavier corpus does not yet exist. The tools are available — for example the established literary methods of John Burrows (Delta method) or n-gram analysis — and could be applied to the digitized works. The hypothesis "different spirit authors are stylistically clearly distinguishable" would be empirically falsifiable; so would the counter-hypothesis "everything is Chico's style with cosmetic variation". Such a study would be methodologically uncontroversial and substantively interesting — regardless of the result it produces.
With the current state of AI-assisted language processing the effort would, in fact, be surprisingly modest. Beyond classical Burrows-Delta statistics, modern methods offer considerably finer instruments: embedding models such as BERT or Sentence-Transformers capture stylistic features beyond mere word frequencies — sentence rhythm, semantic association patterns, typical constructions — in high-dimensional vectors that can be compared across corpora. Classifiers trained on authorship attribution reach hit rates above 90 % given sufficient material. A concrete study design would be straightforward: digitize the corpus, label by "spirit author", run cluster analysis in stylistic space, compare against control corpora of contemporary Brazilian literature and against Chico's personal letters. Using Python libraries such as stylo, pyplot-stylo or embedding models directly, the work could be done in a few weeks. That this has not yet been carried out cleanly is, given the importance of the corpus and the accessibility of the tools, more a research gap than a real obstacle.
Sources
- Perandréa, C. A. (1991): A Psicografia à Luz da Grafoscopia. São Paulo: Editora Jaboticaba (original graphoscopy study on Chico Xavier's psychographies).
- Lewgoy, B. (2004): O grande mediador: Chico Xavier e a cultura brasileira. Bauru: EDUSC (anthropological study of the work).
- Stoll, S. J. (2003): Espiritismo à brasileira. São Paulo: EDUSP/Orion (cultural-studies framing).
- On the Humberto de Campos lawsuit (1944) and the Goiânia acquittal (1979): trial reports and spiritist literature; see sources in the main blog on Chico Xavier.
- On the neurophysiological side: Peres, J. F., Moreira-Almeida, A., Caixeta, L., Leuchsenring, F., Newberg, A. (2012): Neuroimaging During Trance State: A Contribution to the Study of Dissociation. PLoS ONE 7(11): e49360.
- On stylometric methodology: Burrows, J. F. (2002): Delta: A Measure of Stylistic Difference and a Guide to Likely Authorship. Literary and Linguistic Computing 17(3): 267–287.