PGM is short for Papyri Graecae Magicae, the modern edition sequence for manuscripts now held in different collections. Its Roman numerals identify editorial units; museum and library numbers identify surviving rolls and codices. PGM VII corresponds to British Library Papyrus 121, while PGM I and II point to Berlin catalogue numbers. A single museum may hold several PGM witnesses, and parts of a witness may survive in more than one collection.
Four labels, four physical situations
| Editorial label | Physical witness | What the holding record adds |
|---|---|---|
| PGM I | P. 5025 A+B (Warsaw / Berlin papyrus depot) | Greek roll; five recto columns; fourth-to-fifth-century date; uncertain Theban provenance. |
| PGM II | Berlin P. 5026 | Greek roll dated to the fourth century; the catalogue notes its relationship to British Library Papyrus 47. |
| PGM IV | Paris, BnF Supplément grec 574 | Large codex with Greek and Old Coptic material; a research project describes 72 pages and 3,274 lines. |
| PGM VII | British Library Papyrus 121 | Fourth-century roll written on both sides, with 19 recto and 13 verso columns and later additions. |
The table is an orientation, not a complete inventory. Berlin’s PGM index ↗ includes additional labels and subdivisions. PGM IV is a codex rather than a roll. The Würzburg account calls it the Great Magical Papyrus of Paris and describes texts in two linguistic traditions: Greek instructions and Old Coptic-Egyptian invocations in some recipes. Calling every line “Greek-only” would erase that evidence. The physical format also matters: turning a codex page is a different action from unfolding an ancient roll.
PGM VII and Papyrus 121
PGM VII is the editorial number for a group of passages on British Library Papyrus 121. The library also records the older publication shorthand P.Lond. I 121. Its catalogue describes the roll, assigns a fourth-century date, and inventories writing by column and side. A PGM line range belongs to the modern edition, whereas the papyrus number belongs to the collection holding the ancient object.
For Berlin, search the inventory number P. 5025 or P. 5026. The database gives acquisition, material, form, language, palaeographic date, publication history, and digital image links under its own terms of use. P. 5025 was acquired from the Anastasi collection in the nineteenth century; that modern acquisition is separate from its uncertain ancient provenance. P. 5026 is associated with another British Library fragment in the catalogue, so even a neat modern number may hide a more complex physical history.
Worked comparison: a roll recipe and a codex vision
One attraction recipe on the verso of Papyrus 121 (PGM VII) is organized as a short chain: a common object is prepared, names and signs are put on or around it, a formula is spoken, and a time condition governs the promised result. The British Library’s examples show the same compact grammar with a shell, a cup, or lamp wicks. Object, inscription, speech, and timing can be read almost as four fields in a recipe. The roll’s column divisions keep one small operation beside many unrelated aims, from protection to dream requests.
The bilingual direct-vision at PGM IV.5–51 occupies a different scale. It opens with seven days of purity, fixes the main action to a lunar date and sunrise, moves through a Nile-bank setting and an initiation sequence, and only then reaches its invocations and promised encounter. The instructions are in Greek while the invocations use Old Coptic-Egyptian. Here language switching marks a functional boundary inside one operation: Greek directs what is done, while Egyptian-language speech addresses the powers expected to appear.
The comparison changes what the editorial numbers tell us. PGM VII preserves brief, modular recipes on a reused roll; PGM IV can sustain a long, staged procedure across a codex page. Both combine material action and speech, but they pace authority differently. The Papyrus 121 recipe makes force seem immediate through a compressed sequence. The Paris text delays the requested answer until purity, place, time, bodily preparation, and bilingual invocation have accumulated. Format and sequence are part of each text’s argument, not neutral containers for a list of “spells.”
Three layers of the text
The manuscript preserves ancient writing on a physical surface. An edition organizes, transcribes, and numbers damaged or dispersed passages. A translation renders an edited reading in another language. Different editions can disagree about a damaged word or passage boundary without implying that the ancient roll itself changed. The London Papyrus 121, Berlin rolls, and Paris codex each pass through these distinct layers.
Sources and further reading
- British Library, Papyrus 121 ↗ — explicit equivalence with PGM VII
- Berlin Papyrus Database, P. 5025 ↗ — PGM I witness and physical record
- Berlin Papyrus Database, P. 5026 ↗ — PGM II witness and British Library connection
- University of Würzburg, PGM IV ↗ — Paris codex and bilingual passages
- Berlin Papyrus Database, PGM index ↗ — editorial number navigation