Research guides
Missing leaves, copied errors, translation, reordering, and editorial choice explain why witnesses under one grimoire title often disagree.
A book is often a family of witnesses
A grimoire title rarely guarantees a single stable text. A scribe may combine excerpts from several works, translate a passage, redraw a seal, or add a local note for a new reader. Later collectors can bind separate booklets together and assign them one title. The Folger Shakespeare Library’s sixteenth-century Book of Magic illustrates the physical side of this history: parts were separated, later matched by distinctive blue pagination, and reunited after an auction acquisition. Its missing and recovered leaves matter to any account of what the “book” contains. A digital image of one opening shows only one stage of its life. Historians use shelfmarks, handwriting, paper, watermarks, bindings, and patterns of error to reconstruct relationships between witnesses. The result may be a branching history rather than a recoverable original master copy.
Translation is interpretation
Printed editions widen access but also stabilize one editor’s decisions. Mathers’s Key of Solomon and Abramelin translations are well-known examples. A translator chooses how to render specialized words, while a publisher decides how to set tables, circles, and images on a page. An edition drawing on multiple manuscripts may combine readings that never appeared together in one historical copy. Some older editions omit passages; some modern transcriptions silently modernize spelling. A digitized nineteenth-century book is a primary source for that century’s occult reception and an edited witness to an older textual tradition. Its publication date is distinct from the dates of the ideas and manuscripts it presents.
Before and after: twelve Almadel altitudes become four
The Almadel tradition provides a visible before-and-after example of transmission changing structure. In a Latin branch, twelve altitudes coordinate zodiacal signs, months, directions, colors, divine names, and angelic offices. The small table or altar is one component in that twelve-part celestial map. The English Ars Almadel incorporated into the Lemegeton reduces the sequence to four colored altitudes. Its first two groups correspond fairly closely to the first two Latin altitudes, its third resembles the Latin eighth, and its fourth has no straightforward partner. The English version also describes a wax apparatus and introduces a crystal, while older accounts expect a manifestation above an altar through which fumigation rises. The drift reaches structure as well as spelling. Selection compresses twelve indexed regions into four; translation alters instructions for the object; compilation places the result beside three other Lemegeton books. A modern redrawing made from the English text can be perfectly accurate to that witness and still be a poor image of the wider Almandal family. “Before” and “after” name related textual arrangements, not a pristine original and a merely corrupt copy.
Why sigils change between witnesses
A grimoire image can come from a manuscript, printed edition, modern redrawing, or generated illustration. Titles, editors, publication dates, and manuscript identities separate those objects. Library catalogues sometimes describe an object differently from a website caption, and another witness may preserve a different spelling, number, or visual feature. Sigils are especially sensitive to copying: one altered stroke can produce a visibly different character. These variations reveal the history of a particular version rather than a single timeless design. Page layout can cause changes too. A copyist may rotate a sign to fit a margin, separate a label from its figure, or mistake a faded line for damage. Printing introduces another set of decisions when an engraver translates pen strokes into a reusable block or plate. Comparing the complete pages often explains differences that an isolated digital icon conceals.
Questions to explore
- How can one title represent several physical manuscripts?
- What changes when an editor combines or normalizes witnesses?
- Why can a manuscript date differ from an edition date?