Pistis Sophia: Best Translation, and How to Read the Book

Three English versions of Pistis Sophia are available to read, and they were made from three different languages. So when someone asks for the Pistis Sophia best translation, the useful answer depends on what they want from the book: the story in readable prose, the Coptic word order, or the scholarly standard. This page lays out what the text is, who put it into English, which edition to open, and how to work through it once you do.

What it is: one manuscript, and not a Nag Hammadi book

Everything anyone knows about this text comes from one object. There is no second copy, no quoting Church Father to check it against, no fragment from another library.

●Documented The manuscript entered the British Museum in the eighteenth century. In G. R. S. Mead's description it "was bought by the British Museum in 1785 from the heirs of Dr. Askew, and is now catalogued as MS. Add. 5114". ☿G. R. S. Mead (trans.) (1921). Pistis Sophia. John M. Watkins, London, 2nd revised edition; Celephais Press e-text, archive.org.Read full source →☿¹ Francis Legge, writing the introduction to the 1924 English version, adds the price: ten guineas, at the sale of the library of the antiquary Dr. Anthony Askew. ☉George Horner (trans.), with an introduction by F. Legge (1924). Pistis Sophia. Society for Promoting Christian Knowledge, London; Project Gutenberg #76266.Read full source →☉²

●Documented That purchase date settles a question the search results keep getting wrong. The codex had been in London for a hundred and sixty years before the thirteen codices were found near Nag Hammadi in 1945. The two finds are both Gnostic and entirely separate, which is why this text is absent from Nag Hammadi anthologies and why readers meet it in a volume of its own. The identification question that follows from the same codex is handled separately in Is Pistis Sophia Mary Magdalene?.

Then there is the title, which is the part most worth knowing before you start.

●Documented The book does not call itself Pistis Sophia. Legge's description of the manuscript explains that above one column of page 115 a later hand wrote that this was the second tome of the Pistis Sophia, "from which the whole MS. has taken the title by which it is known". ☉George Horner (trans.), with an introduction by F. Legge (1924). Pistis Sophia. Society for Promoting Christian Knowledge, London; Project Gutenberg #76266.Read full source →☉² A heading over one column became the name of the whole book. The figure of Pistis Sophia appears in the text, and so does the phrase that became its label, but the two are not the same claim.

◐Debated What the collection should be called, and who wrote it, are both open. Carl Schmidt, whose German version of 1905 is the basis of most later work, held that the text "is not Valentinian, but comes from an obscure sect of what he calls Barbelo-Gnostics, known as Severiani", and that its title should be Τεύχη τοῦ Σωτῆρος, the Books of the Saviour. ☉George Horner (trans.), with an introduction by F. Legge (1924). Pistis Sophia. Society for Promoting Christian Knowledge, London; Project Gutenberg #76266.Read full source →☉² Legge, printing that summary, argued the other way for part of the collection: the third document, he thought, was the work of a late Valentinian, written after the reign of Septimius Severus. Neither reading has carried the field, and a reader should know that the codex is a miscellany rather than one author's book.

Who translated it: Latin first, Coptic last

The chain matters, because it is the reason two English versions of the same page can differ in substance.

Moritz Gotthilf Schwartze copied the codex in London in 1848 and published a Latin version in 1851, edited after his death by J. H. Petermann. An anonymous French rendering followed in 1856, Émile Amélineau's French in 1895, and Schmidt's German in 1905. English came out of that European work rather than out of the Coptic.

●Documented Mead's version, the one most readers meet, is openly secondhand, and he said so first. Explaining why he published at all, he wrote that "it stands to reason that the translation of a translation must needs be but an apology for a first-hand version". ☿G. R. S. Mead (trans.) (1921). Pistis Sophia. John M. Watkins, London, 2nd revised edition; Celephais Press e-text, archive.org.Read full source →☿¹ He translated Schwartze's Latin for the 1896 first edition, checked it against Amélineau, and revised the whole against Schmidt's German for the 1921 second edition. Four languages stand between the Coptic and the sentence you read.

●Documented A version made straight from the Coptic arrived in 1924, when George Horner published his for the S.P.C.K. with Legge's introduction. Horner describes his method without decoration: "The literal translation follows the order of the words as far as possible". ☉George Horner (trans.), with an introduction by F. Legge (1924). Pistis Sophia. Society for Promoting Christian Knowledge, London; Project Gutenberg #76266.Read full source →☉² Mead, writing three years earlier, had hoped an English Coptic scholar would take the work in hand and knew of no such project, which places Horner's as the first English made from the original language.

The modern scholarly edition came much later: the Coptic text edited by Schmidt with an English translation and notes by Violet MacDermot, published in 1978 as volume IX of the series Nag Hammadi Studies. ☽Carl Schmidt (ed.) and Violet MacDermot (trans.) (1978). Pistis Sophia. E. J. Brill, Leiden (Nag Hammadi Studies IX).☽³ Coptic and English sit on facing pages. The series title is worth noticing: the standard scholarly edition of a text that is not from Nag Hammadi was published in a series named after Nag Hammadi, and a catalogue entry read quickly is one plausible route by which the two came to be confused.

We looked for a published Bulgarian translation of the full text and found none. A Bulgarian reader reads this book in English.

Which edition to read: choosing a Pistis Sophia best translation

Three editions, three different jobs.

For a first reading, Mead. His prose moves, his chapter numbers are the ones everyone cites, and he numbered the verses of the repentances so they can be set beside the Psalms they echo. The cost of that readability is distance from the original.

When a single phrase carries weight, open Horner beside Mead. Horner's English is awkward on purpose, following Coptic word order, and the awkwardness is information: it shows you where the smooth version made a choice. His pages also carry the Schwartze–Petermann marginal numbers, which is how scholarship refers to the text.

If you are going to quote the book in anything that will be checked, use MacDermot. It is the edition a specialist will expect, and it is the only one of the three with the Coptic in front of it.

How to read it: the repentances, and how many there are

Do not start at page one. The opening dialogues assume a cosmology the text only explains later, and most readers who abandon the book abandon it there. Start at Book I, chapter 32, where Sophia begins to speak from the chaos, and read the sequence through.

●Documented The count is the thing people get wrong, and the text itself is the reason. Sophia says she spoke twelve: "Because of this I said twelve Repentances according to each one of the Aeons". ☉George Horner (trans.), with an introduction by F. Legge (1924). Pistis Sophia. Society for Promoting Christian Knowledge, London; Project Gutenberg #76266.Read full source →☉² Then, in the same passage, "she added also in the thirteenth Repentance, saying, Hear me while I hymn unto thee, O Light of the Lights". ☉George Horner (trans.), with an introduction by F. Legge (1924). Pistis Sophia. Society for Promoting Christian Knowledge, London; Project Gutenberg #76266.Read full source →☉² Twelve for the twelve Aeons she fell through, then one more for the thirteenth, the place she came down from. Thirteen in total, and Horner's index lists all thirteen by page. The sequence is read as a soul map in the confession of Pistis Sophia.

●Documented One more thing to know about the page in front of you: its shape is editorial. Mead states that "the detailed paragraphing does not exist in the original, which runs on for the most part monotonously without break", and that the chapter divisions are Schmidt's. ☿G. R. S. Mead (trans.) (1921). Pistis Sophia. John M. Watkins, London, 2nd revised edition; Celephais Press e-text, archive.org.Read full source →☿¹ Every paragraph break, chapter number and verse number you navigate by was added by a modern editor. The manuscript is an unbroken wall of two-column script. That is not a flaw in the editions; it is a reminder that the structure you are reading by is a map laid over the territory, and the territory is one long voice.

Read the repentances as what the word metanoia actually names, a turning of the mind rather than an apology. Sophia does not get rescued because she asks correctly. She is answered when the asking changes into trust, which is the quality her other name, pistis, carries. That turn is the whole instruction, and it happens inside her, not in the geography of the Pleroma above her.

Tonight: open Mead at chapter 32 and read the first repentance aloud, slowly, once. Then read it again and mark the line where the asking stops and the trust starts. In Sophia's mouth those are two different registers, and they switch mid-prayer. Notice where that switch sits in your own asking, the one you have been repeating this week without hearing yourself. The text gives you thirteen attempts to find it. You get the same number.

For the wider pattern this text belongs to, the sacred feminine guide sets Sophia's fall beside the other traditions that tell it.

Verdict: documented text, younger English than it looks

The book is documented, and unusually well documented for a Gnostic text: a dated purchase, a catalogue number, a surviving manuscript anyone can ask to see, and a translation history recorded by the translators themselves. Nothing about its existence or its transmission rests on tradition.

What is debated is who wrote it, which school it came from, and what it should properly be called. What is widely misunderstood is the rest: the title is a scribe's heading, the English most people read passed through three other languages first, the chapter numbers are a nineteenth-century scholar's, and the book has no connection to Nag Hammadi beyond the name of the series its critical edition appeared in. Knowing those four things changes how you read every page.

For the same desk treatment applied to a claim rather than a text, see Is the Kybalion Ancient?.

FAQ

Which Pistis Sophia translation is best? It depends on the use. Mead's 1921 version is the most readable and the most cited, and it is free online, but it came through Latin, French and German. Horner's 1924 version is literal and made from the Coptic. MacDermot's 1978 Brill edition is the scholarly standard, with the Coptic on facing pages.

Is Pistis Sophia in the Nag Hammadi library? No. It survives in the Askew Codex, now British Library Add. 5114, which the British Museum bought in 1785. The Nag Hammadi codices were found in Egypt in 1945. The confusion is understandable, because the critical edition of Pistis Sophia was published in a Brill series called Nag Hammadi Studies.

How many repentances does Pistis Sophia have? Thirteen. The text has Sophia say she uttered twelve, one for each of the twelve Aeons she fell through, and then add a thirteenth for the thirteenth Aeon, the region she came down from. Both statements sit in the same passage, which is why the count is often reported as twelve.

Where can I read the Pistis Sophia full text for free? Two complete English versions are in the public domain. Mead's 1921 edition is on archive.org as a plain-text and scanned file. Horner's 1924 translation is on Project Gutenberg as ebook 76266, proofread rather than scanned, with Legge's introduction on the manuscript's history included.