Voynich manuscript research

A source-traceable investigation of an undeciphered manuscript: visual observations first, interpretation only when evidence supports it.

Current research

Visual Transcription Pass 3 (v006) — sealed provisional visual inventory

Pass 3 consolidated v005 fragmentation without targeting an alphabet size: 42,762 provisional families (was 46,955), 39,213 singletons (was 43,296) and 42,883 shape modes. Fresh blind QC found 92.7% of singleton attachments and 93.9% of within-family pairs the same form; failed family-to-family merges were reverted. Families remain provisional visual groupings that partition continuous variation. Six cards need review; every family and occurrence is in Explore All Results.

Project summary, all results and review needed

Research history

Completed studies appear newest first. Earlier conclusions remain historical; later work may qualify them. Two legacy explorers have partial coverage, and some older summaries did not separately record every explanatory field. Their preserved findings and limits remain available. Visual Transcription Pass 3 is now sealed as a provisional visual inventory and is presented as the Current Project.

Global script fingerprint search

2026-09-29

Do the recurring visual forms of Voynich writing, taken together, resemble any known historical writing tradition more strongly than chance would produce?

Explore this research project

Zodiac semantic-anchor experiment

2026-09-29

Do original Voynich visual observations on the twelve zodiac diagrams show a reproducible sign/month-linked pattern that could serve as a semantic anchor, under permutation, source-view and held-out controls?

Explore this research project

Independent visual-transcription audit — v005

2026-09-29

What original visual writing observations survive the corrected independent full-source audit, with uncertainty and Yale provenance preserved?

Explore this research project

Full-source neutral visual transcription — v004

2026-09-29

What visually appears written across the 213 authoritative Yale views, with source and segmentation uncertainty preserved?

Explore this research project

Blind structural-language comparison

2026-09-29

Without assuming a language, which known language types does the structural behaviour of Voynich writing resemble, under blind, frozen, controlled scoring?

Explore this research project

MS 408 before Rudolf II: provenance and book networks

2026-09-28

Can an evidence-supported physical transmission path be reconstructed for approximately 1430–1599?

Explore this research project

Gallows/bench vs German cursiva construction study

2026-09-28

Do Voynich gallows/bench constructions share real construction principles with a German/Central-European scribal tradition, or were earlier matches generic looped handwriting?

Explore this research project

9-like form and scholastic Latin abbreviation

2026-09-28

Does the recurring Voynich 9/g-like construction share recognizable medieval Latin scribal machinery, and what did that machinery do?

Explore this research project

LJS 456 secret-script follow-up

2026-09-28

What is the unusual 9/6/3-like sign text in the 1394 manuscript LJS 456, and is its resemblance to Voynich meaningful?

Explore this research project

Historical medical-manuscript comparison

2026-09-28

Does Voynich's writing look more like medieval medical/pharmaceutical working notation than ordinary contemporary writing?

Explore this research project

Voynich–Cappelli visual fingerprint comparison

2026-09-28

Do distinctive Voynich constructions match medieval Latin abbreviation signs catalogued by Cappelli?

Explore this research project

Double-loop and open-bend form study

2026-09-28

Starting from f016r, does the double-loop / S-like form recur as one visual family across the manuscript?

Explore this research project

Davis vs. Timm handwriting diagnostic audit

2026-09-26

Do Yale pixels support Davis's proposed hands or Timm's single-writer view on the diagnostic forms both cite?

Explore this research project

Writer vs. context analysis

2026-09-26

Are recurring similar Voynich forms better predicted by proposed writer or by their visual surroundings?

Historical explorer has partial coverage; its limitations are preserved.

Explore this research project

Blind handwriting / authorship analysis

2026-09-26

Does the measured handwriting fall into stable writer groups when compared blind across the manuscript?

Explore this research project

Manuscript-wide internal 28-pattern survey

2026-09-25

Which visual shapes and combinations recur across all 213 Yale images, without assigning identities?

Explore this research project

Full neutral extraction and validation

2026-09-25

Can all 213 Yale TIFFs be extracted neutrally with every region traceable to source pixels?

Historical explorer has partial coverage; its limitations are preserved.

Explore this research project

f1r 16 consequential-case review

2026-09-25

Which 16 consequential f1r segmentation cases should be accepted, rejected or kept separate?

Explore this research project

f1r high-resolution calibration and review

2026-09-25

How does f1r extract from the native Yale TIFF compared with the earlier PDF-derived run?

Explore this research project

First f1r calibration and PDF source audit

2026-09-24

Does the supplied PDF provide more manuscript detail, and what does a controlled f1r rerun show?

Explore this research project

Browse the manuscript

Explore all 213 Yale source views using canonical folios, searchable thumbnails and larger display copies.

Open the manuscript browser

About this project

The aim is a language-neutral, human-verifiable visual transcription, followed by independent tests that do not assume a decipherment. Similar shapes and provisional families do not establish letters, sounds, words, language or authorship.

The current v006 inventory overlays the unchanged v005 observation and material ledger. It revises family, variant, singleton and segmentation layers after a corrected rebuild and blind QC. Families remain provisional visual groupings; writing recall, glyph identity and independent human agreement are unresolved. Earlier studies preserve their historical v005 limits. Engineering validation is not scientific acceptance.

This website presents completed research and its review evidence. It does not assign a translation. Web images are display copies; native coordinates and source hashes preserve the evidence chain.

Manuscript source: Beinecke Rare Book & Manuscript Library, Yale University, MS 408. View Yale’s catalogue.