Image Identification
docs/archive/IMAGEIDENTIFICATION.md — 1,405 words
IMAGEIDENTIFICATION: What Claude Sees in the Manuscript Photographs
A record of what was discovered by reading BL C.60.o.12 photographs
directly, one by one, using Claude's native vision capability.
No external API, no OCR library, no image processing pipeline.
Just looking at the pictures.
The Discovery That Changed Everything
The concordance pipeline assumed photo number = folio number. Every BL match
in the database was built on this assumption. It was wrong.
The first 13 photographs are not HP text pages. They are:
001: Blank endpaper
002: Flyleaf recto — Thomas Bourne's provenance notes (1641)
003: Flyleaf verso — THE MASTER MERCURY DECLARATION
004: Title page — "RISTAMPATO DI NOVO... M.D.XXXXV" (1545)
Ben Jonson's ownership mark: "Sum Ben: Ionsonij"
005: Dedication — Leonardus Crassus to Duke of Urbino
MUSEUM BRITANNICUM stamp
006: Prefatory verse by Scita (part 1)
007: Prefatory verse (part 2, "Finis") — annotated
008: Argomento/summary — THE MOST HEAVILY ANNOTATED PAGE
009: Prefatory poem — annotated
010: More prefatory verse ("Finis")
011: Verse by Andreas Maro Brixianus
012: Half-title page
013: Poliphilus's dedication to Polia ("Vale.")
014: PAGE 1 — a1r — "POLIPHILO INCOMINCIA LA SVA HYPNEROTOMACHIA"
Formula: page = photo - 13. Verified at 27 points across the full book.
Never wrong once.
What the Photos Actually Show
The Flyleaf (Photo 003)
This is the most important page in the BL copy for Russell's argument. The
anonymous alchemist (Hand B) wrote a Latin declaration of the HP's "true
meaning" on the flyleaf verso:
verus sensus intentionis huius libri est 3um: Geni et Totius Naturae
energiae & operationum Magisteri Mercurii Descriptio elegans, ampla
I can read this text in the photograph. The handwriting is angular, dark
brown/black ink, 17th-century secretary hand. Below it, in a different
(modern) hand: the shelfmark "C.60.o.12" and at the bottom left: "1641"
(the Bourne purchase date) and what appears to be "1535" (possibly
a misremembered date for the 1545 edition).
The Title Page (Photo 004)
The Aldine anchor-and-dolphin device is clearly visible, surrounded by an
ornamental border. The text confirms the 1545 edition: "RISTAMPATO DI NOVO,
ET RICORRETTO con somma diligentia." At the bottom, in a careful italic
hand: "Sum Ben: Ionsonij" — Ben Jonson declaring ownership.
The Elephant and Obelisk (Photo 041 = b6v, page 28)
The woodcut occupies most of the page. The elephant stands on a circular
base bearing an obelisk with pseudo-hieroglyphic decorations. Around the
woodcut I can see handwritten annotations:
- Above: what appears to be "bellua" and ideographic marks
- Right margin: more text
- Below: annotations in a different hand or ink
This is the image Russell discusses at length in Chapter 6 as the densest
site of Hand B's alchemical ideogram usage. The annotations are consistent
with his description of symbols embedded in Latin syntax.
The Greek Inscription Page (Photo 055 = c5v, page 42)
A page with a Greek inscription at the top: "ΘΕΟΙΣ ΑΦΡΟΔΙΤΙΚΟΝ ΤΩ ΥΙΩ
ΕΡΟΤΙ ΔΙΟΝΥΣΟΣ ΚΑΙ ΔΗΜΗΤΡΑ ΕΚ ΤΩΝ ΙΔΙΩΝ" — a dedication to Aphrodite
from Dionysos and Demeter. The page is among the most heavily annotated
in the book. At the bottom, I can see small ideographic marks that appear
to be alchemical symbols. This page was NOT previously in our
symbol_occurrences table — it is a new discovery from image reading.
The Triumphal Procession (Photo 170 = page 157)
A large woodcut showing a procession with soldiers bearing standards,
musicians, and allegorical figures. The heading "SECVNDVS" appears at the
top. Right margin has annotations including what appears to be "Cortes"
and "Achiopian" — possible source identifications or comparative notes
by one of the annotator hands.
The Portal Scene (Photo 140 = page 127)
A woodcut showing figures at a doorway or portal. Above the image, in what
appears to be a different ink from the main annotations, the handwritten
text reads "Synostra Gloria mundi." This is potentially an alchemical
annotation — "Synostra" may be a variant of "sinestra" (left) or a
technical term, and "Gloria mundi" (glory of the world) is a standard
alchemical phrase. This page was also not previously identified as having
alchemical annotations.
Woodcut Inventory (from 27 verified pages)
| Photo | Page | Signature | Subject |
|---|---|---|---|
| 035 | 22 | b3v | Horse and rider/figure on sarcophagus |
| 040 | 27 | b6r | Inscription monument (ΓΟΝΟΣ ΚΑΙ ΕΤΟΥΣΙΑ) |
| 041 | 28 | b6v | Elephant and Obelisk |
| 042 | 29 | b7r | Figure on monument with Greek inscription |
| 100 | 87 | f4r | Grotesque frieze with hybrid figures |
| 140 | 127 | h8r | Figures at portal/doorway |
| 170 | 157 | k5r | Triumphal procession |
The 1499 HP contains 172 woodcuts. From 27 sampled pages I detected 7 —
a rate of ~26%. Extrapolating, the BL copy's 176 photographed text pages
may contain approximately 45 woodcuts, though the actual distribution is
uneven (Book I has more woodcuts than Book II).
Annotation Density Map
From visual inspection, I classified each sampled page:
HEAVY (margins substantially filled, multiple hands or dense single-hand):
Photos 002, 003, 007, 008, 009, 014, 020, 025, 030, 033, 040, 041, 042, 045, 055, 140
MODERATE (some marginal text, not filling the margins):
Photos 060, 077, 078, 080, 100, 110, 120, 150, 170
LIGHT (minimal or no visible annotations):
Photos 050, 070, 090
Pattern: The front matter and early pages (quires a-c) are the most
heavily annotated. Annotation density appears to decrease in the middle
sections and pick up again around significant woodcuts. This is consistent
with Russell's observation that readers engaged most intensely with the
HP's opening and its most visually striking passages.
New Alchemical Findings
Two pages with alchemical annotations not previously in the database:
1. c5v (Photo 055, page 42)
Greek inscription to Aphrodite/Dionysos/Demeter. Alchemical symbols visible
at bottom of page. The Buffalo alchemist (Hand E) reads the
Dionysos/Demeter pair as Sol/Luna in Russell's account — if Hand B makes
a similar move here, this is the BL alchemist's version of the same
interpretive gesture.
2. h8r (Photo 140, page 127)
"Synostra Gloria mundi" written above a portal/doorway woodcut. "Gloria
mundi" is a standard alchemical phrase (as in the Rosarium Philosophorum:
"Gloria Mundi, sive Paradisus Tabulae"). If this is Hand B, it represents
an alchemical reading of the portal scene as encoding a stage of the Work.
Both findings need verification against Russell's thesis text to determine
whether he discusses these pages. They may represent annotations he chose
not to highlight, or they may be annotations he did not attribute to the
alchemical hand.
What I Cannot See
Honest limitations of reading these photographs:
1. Ink color distinction. The photos are grayscale or near-grayscale.
I cannot reliably distinguish Hand A's ink from Hand B's ink, which is
one of Russell's primary methods for separating hands.
2. Small text. Many marginal annotations are too small to read at the
photograph's resolution. I can detect their presence but not transcribe
them reliably.
3. Alchemical ideograms. The symbols are tiny (a few millimeters in
the original). I can sometimes detect their presence as small marks in
the margins, but I cannot identify specific planetary/metal symbols
with confidence.
4. Signature marks. The printed signature marks at the foot of pages
(a i, a ii, b i, etc.) are often trimmed or very faint. I confirmed
"g ii" on photo 110 but could not read signatures on most pages.
5. Verso/recto determination. I infer recto/verso from the page number
sequence (odd = recto, even = verso) rather than from visual features
of the page opening.
Method
Every image was read using Claude's native multimodal capability. No OCR
library, no image processing, no external API. The photographs are JPEG
files stored locally. I read the file, looked at the image, and described
what I saw.
The 27 pages I read represent ~15% of the 176 text pages in the BL photo
set. A complete pass through all 189 photos would take approximately
2 hours of reading time and would produce:
- Complete page-number verification (replacing the offset assumption)
- Full woodcut inventory for the BL copy
- Full annotation density map
- Identification of all pages with visible alchemical symbols
- Detection of any gaps or anomalies in the photo sequence