How to Read an Old Grimoire Without Fooling Yourself

Most practitioners who pick up a grimoire are already at least one step removed from what they think they are reading. The book is a translation of a translation, based on manuscripts copied centuries after the period the title page implies, assembled from sources the compiler didn’t always understand, and named after a figure — Solomon, Agrippa, Honorius — who almost certainly had nothing to do with it.

None of that makes the text useless. It makes it something specific: a document with a dateable context, a traceable transmission history, and a set of claims that can be evaluated rather than accepted wholesale. What follows is a working method for doing exactly that.

What “grimoire” means, and what it does not mean

The word is French, from grammaire — grammar, in the sense of learning. It arrived in English roughly in the eighteenth century as a term for a book of magic instruction, particularly one containing conjurations, talismans, and spirit lists. Owen Davies, in Grimoires: A History of Magic Books, illuminates the many fascinating forms these recondite books have taken and exactly what these books held. His framing is useful: Davies focuses on conjuring and talismans, including the diabolic but treating divinatory material differently, while texts like the Grand and Petit Alberts, the Key of Solomon, and books attributed to Agrippa crop up repeatedly throughout his account.

The word is not a synonym for “old book of spells.” A seventeenth-century herbal with charm receipts tucked between the poultice instructions is not a grimoire in the genre sense. A handwritten notebook of folk cures is not a grimoire. The grimoire proper is a text in a specific genre: structured conjuration or talisman-making procedure, often with accompanying spirit hierarchies, astronomical or astrological timing conditions, and ritual prescriptions. This matters when you are researching, because the secondary literature uses the word precisely, and slipping into loose usage muddies the sources you are drawing on.

The pseudepigrapha problem

The single most common misunderstanding about historical grimoires is what their attributed authors mean.

“The Key of Solomon” does not mean Solomon wrote it. “Agrippa’s De Occulta Philosophia” does mean Agrippa wrote it — that text is genuinely his — but texts “attributed to Agrippa” or “in the manner of Agrippa” circulating after 1535 often are not. “The Sworn Book of Honorius” has nothing to do with any pope named Honorius. The pattern is so consistent it has a technical name: pseudepigrapha, meaning writing falsely attributed to a prestigious name.

The mechanism is not fraud in the modern sense. Claiming a text descended from Solomon or was revealed to a biblical patriarch was, in medieval manuscript culture, a way of asserting that the knowledge was old enough and authoritative enough to be worth preserving. The claim was conventional, like a modern book citing peer-reviewed research to establish legitimacy. Readers of the period understood the convention. Modern readers who don’t mistake it for biography.

The practical consequence is this: the attributed author tells you almost nothing about when or where a text originated. The manuscript evidence does.

Three texts, three lessons in transmission

The Key of Solomon

The Clavicula Salomonis, known as the Key of Solomon, is a grimoire dating to at least the fourteenth century. The earliest reference to it appears in Peter of Abano’s Lucidator dubitabilium astronomiae of 1303–10, and the oldest known manuscript fragment dates to approximately 1380–1410. That fragment is two leaves, recently discovered by Matteo Cova at the Biblioteca capitolare di Trento; before its identification, the oldest known copy was an Italian manuscript (BNF Ital. 1524) dated 1446.

Most contemporary practitioners encounter the text through S.L. MacGregor Mathers’s 1888 translation. Here is the problem with that encounter: Mathers’s translation is almost entirely based on French Colorno manuscript exemplars dating to the eighteenth century, represented by the Kings 288, Harley 3981, and Sloane 3091 manuscripts. Kings 288 and Harley 3981 indicate they were translated by Abraham Colorno. In other words, the Victorian translation that most people treat as the authoritative text is a rendering of eighteenth-century French copies — several additional steps removed from the medieval Latin tradition.

The manuscripts Mathers used were from the British Museum, but were actually written around the sixteenth century in Renaissance Italy. The textual gap between “King Solomon’s original” and what Mathers translated and published runs approximately two and a half millennia in the romantic version, and more accurately spans several centuries of manuscript copying, at minimum two intermediate languages, and the specific decisions of scribes and translators whose agendas we can only partially reconstruct.

None of this means the text is worthless. It means the Mathers translation is a useful document of what a nineteenth-century British occultist thought the Clavicula meant, filtered through eighteenth-century French manuscripts copying a sixteenth-century Italian tradition of a medieval Latin text. Each of those steps is real intellectual history. None of them is Solomon.

The Picatrix

The Picatrix is a 400-page Arabic book of magic and astrology, which most scholars assume was originally written in the middle of the eleventh century, though an argument for composition in the first half of the tenth century has been made. Its Arabic title is Ghāyat al-Ḥakīm — “The Aim of the Sage” or “The Goal of the Wise.” The work was translated into Spanish and then into Latin during the thirteenth century, at which time it received the Latin title Picatrix.

The Arabic text was translated into Castilian Spanish and eventually into Latin in 1256 for the Castilian king Alfonso the Wise. David Pingree, whose edition of the Latin text underpins the most recent scholarly translation, described it as “the most thorough exposition of celestial magic in Arabic.”

The transmission history introduces a specific kind of distortion. Researcher Martin Plessner noted that the Latin translator omits many theoretical passages throughout the work. The English-speaking practitioner drawing on the Picatrix is therefore working with: an eleventh-century Arabic compilation of earlier sources, filtered through a thirteenth-century Spanish translation, then a Latin translation that systematically trimmed theoretical content, then centuries of manuscript copying, then — until 2019 — no reliable English translation at all.

Dan Attrell and David Porreca’s English translation, published by Penn State University Press in 2019, is based on Pingree’s edition of the Latin text. The original Arabic text was compiled from over two hundred sources. The introduction to that edition traces the transmission history in detail and is the right place to start if you are working with the Picatrix seriously.

The practical takeaway: when someone cites “the Picatrix” as a source for a specific claim, ask which Picatrix — the Arabic original? the Latin version? which manuscript of the Latin? Attrell and Porreca’s English translation? The answer changes what the citation actually means.

The Ars Notoria and the Sworn Book of Honorius

These two texts illustrate the same problem from a different angle: later contamination being mistaken for early stratum.

The Ars Notoria — a text for attaining mastery of arts and sciences through angelic figures and prayers — circulates in manuscripts from the thirteenth century onward. The Sworn Book of Honorius (unrelated to any historical Honorius) is usually dated to roughly the same period. Both were copied, adapted, merged with other material, and recopied across several centuries. By the time a printed edition appears in the early modern period, it may contain interpolations from other texts entirely, corrections that introduced new errors, and redactions that removed passages a copyist found dangerous or unintelligible.

The social life of these texts is part of what they are. Around 134 trials are known to have occurred in one former Danish territory, and nearly a third of them involved grimoires, written spells, or runes and symbols derived from them. The grimoire as an object — handled, passed between people, annotated, abbreviated, reconstructed from memory — accumulated variation at every exchange. The ideal clean “original” version that modern readers often imagine locating did not exist. The texts existed in circulation, and circulation changed them.

What transmission distortion actually looks like

Here is the shape of the problem made concrete.

A practitioner reads a version of a conjuration in a modern paperback edition and wants to know if the instruction — light a red candle, face east, recite this prayer — is genuinely old practice or a later interpolation. The path from modern paperback to medieval manuscript looks approximately like this:

1. A medieval scribe copies an earlier manuscript, occasionally correcting what seems wrong to him.
2. The manuscript passes through several owners; marginal notes from one copy enter the main text of the next.
3. A Renaissance editor compiles several manuscripts, harmonizing discrepancies by choosing the reading that seems most coherent.
4. A printer in the seventeenth or eighteenth century sets the text in movable type, further normalizing spelling and occasionally “improving” passages.
5. A Victorian occultist translates into English with the theology and aesthetics of the 1880s operating on every word choice.
6. A modern publisher reprints the Victorian translation, sometimes with additional material drawn from other sources without clear distinction.

At step six, the “red candle facing east” instruction may be perfectly medieval. Or it may have been introduced at step two, three, four, or five. Without access to the manuscript tradition and some familiarity with how to read it, there is no way to tell from the text alone. This is not cause for despair. It is cause for description: cite the edition you are using, not the ancient authority it claims.

The practitioner’s reading method

A historical grimoire is not a rulebook handed down from a trustworthy authority. It is a primary source — which is actually more interesting. Here is a working approach.

Establish what you are actually holding. Before you read a single instruction, read the introduction. Find out: when was this edition compiled, from which manuscripts, and by whom? What is the estimated date of the earliest manuscript? What language was it originally written in, and how many translation steps does your edition represent? If the edition has no introduction that answers these questions, treat it as a popular reprint rather than a scholarly edition.

Date the claim, not the text. “The Clavicula Salomonis says to use black candles for banishing” is almost certainly not a statement you can anchor to the fourteenth century. It is a statement you can anchor to whichever manuscript tradition and edition you are working from. Be specific about which.

Distinguish the text’s claims from its period. A sixteenth-century Italian manuscript of a Latin conjuration text tells you what a sixteenth-century Italian reader found useful or credible. It is evidence for magical thinking in sixteenth-century Italy. It is not evidence that the procedure works, that the spirit hierarchy described is real, or that the attributed author endorsed the text.

Seek scholarly editions where they exist. For the Picatrix, Attrell and Porreca’s Penn State edition is available and well-annotated. For the Clavicula Salomonis, Joseph Peterson’s critical edition at esotericarchives.com documents the manuscript stemma and corrects Mathers. Critical scholarly editions now exist for many texts that twenty years ago were only available as Victorian paperback reprints. A practitioner does not need to work from the worst available version.

Treat modern correspondence additions as modern. If a contemporary author tells you that a given grimoire procedure is “traditionally” paired with rose quartz or a specific moon phase not mentioned in the source, that pairing is the contemporary author’s contribution, not the text’s. This is not necessarily wrong — living practice adds to tradition — but it should be flagged as an addition, not cited as historical.

How to cite, concretely

The citation standard for historical grimoire material on this site is:

• Name the edition you are drawing from, with translator or editor where relevant.
• Give the section or chapter from the source text, not just a page number in a paperback reprint.
• If the claim you are making is about the historical record — this text circulated in this period, this ingredient appears across multiple versions — cite a secondary source (Davies, Pingree, Attrell and Porreca’s introduction) rather than the grimoire itself.
• If the edition you are using is a Victorian or popular reprint without documented manuscript basis, say so. “According to Mathers’s 1888 translation” is an honest citation. “According to an ancient Solomonic tradition” is not.

The difference matters to readers who are trying to build on your work, and it matters to the site’s credibility as a reference.

Why this is a practitioner’s problem, not only a scholar’s problem

Practitioners are the people most affected by garbled transmission history, because practitioners are the people making decisions based on what a source says.

If you are constructing a talisman according to Picatrix procedures, the question of which version of the text you are working from is not academic — it determines which materials, timing, and prayers you select. If the thirteenth-century Latin translator omitted the theoretical framework that explained why those materials were chosen, you are working partially blind. Knowing that is useful. It tells you where to look for the gap and what questions to ask of a better edition.

The serious practitioner is already doing a kind of textual criticism every time she chooses one version of a procedure over another, one published edition over a different one. Making that process explicit and documented is the difference between a research-grounded practice and one that accumulates errors with each copy — becoming, in time, another layer of transmission distortion for the next reader to unknot.

A short reading list

These are the editions and secondary works that support this guide directly:

• Owen Davies, Grimoires: A History of Magic Books (Oxford University Press, 2009) — the indispensable survey of how these texts circulated, who read them, and what social function they served. Start here before you start anywhere else.
• Dan Attrell and David Porreca (trans.), Picatrix: A Medieval Treatise on Astral Magic (Penn State University Press, 2019) — the first complete English translation with scholarly apparatus. The introduction is the best single account of what happens to a text in transit across languages and centuries.
• Joseph H. Peterson’s annotated edition of the Clavicula Salomonis (esotericarchives.com) — documents the manuscript tradition and the specific problems with Mathers’s choices; free to read online and updated as new manuscript evidence appears.
• Sophie Page and Catherine Rider (eds.), The Routledge History of Medieval Magic (Routledge, 2019) — the best current scholarly overview of medieval learned magic as a field, with individual chapters on specific texts and their transmission.

Cross-references on this site: (forthcoming — Key of Solomon correspondence breakdown; Picatrix planetary hours; the Ars Notoria and the grammar of angelic petition.)