REJECTED
Amino-acid reading
The character distribution includes incompatible ambiguity and rare-residue symbols.
FORENSIC RESEARCH ARCHIVE
Page 23: the matrix
The reproducible Turkish plaintext, the animal hint, and the narrower boundary that still remains.
Page 23 has a readable payload
The 154-letter matrix is cryptographically authenticated and arithmetically reproducible. Prefixes of length 2 2 2 3 2 3 3 2 2 2 3 give BUNU YAZAN TOSUN OKUYANA KOSUN. It ranks first among all 2,048 two-or-three-character prefix assignments and is the only exact five-word result under the retained dictionary-aware model. The same plaintext survives both signed matrix orientations.
The author’s animal hint
On 27 July 2025, the account wrote that Page 23 contains “an animal name … but not the fox.” TOSUN in the recovered sentence means a young bull or steer, making it the direct semantic answer. The literal forward substring OKUZ (“ox”) reinforces the bovine construction. SNAKE remains visible across syllable boundaries in the signed mirror, but unlike TOSUN it is not part of the recovered Turkish sentence.
This resolves the semantic content with high confidence. It does not supply an author-signed statement that the sentence is every layer or the final reward.
Read the final-step audit → · Read the bounded substring audit →
The decoy structure
The printed fraction machinery generates the matrix, but its denominators reduce to repunit-style presentation devices. Eight of eleven rows follow a rigid pseudo-Turkic consonant/vowel template, leaving little entropy for a conventional substitution. The row-seven OKUZ placement plausibly explains its exceptional opening pattern.
Residual boundary
A source-image inventory confirms that the legible 1,500 by 2,167 original book page is preserved locally, alongside a smaller 2021 public repost. Row 9 still has multiple arithmetically legal segmentations, but the author-published row supplies the letters used by the recovered sentence. What remains open is an author-selected downstream operation or a possible second payload in the filler or element-choice layer.
Read the row-9 derivation audit → · Read the source-image inventory →
What it does not do
REJECTED
Classical cipher families
Tested route, columnar, Vigenere, Beaufort, affine, and Hill approaches did not return a calibrated plaintext.
NOT ESTABLISHED
A second payload
No exact input, key, operation, and independently repeatable output has been supplied.
Classical-cipher battery: no candidate beats the matrix
A pre-registered, retained battery tested route reads, columnar and keyed-columnar transposition, Vigenere, Beaufort, affine transforms, and sampled invertible 2-by-2 Hill keys. Its pass line was fixed before the search. Of 63 scoreable candidates, zero passed; the best result was the unchanged matrix itself. This is useful negative evidence for those stated families, not a proof that Page 23 has no author-selected or non-classical operation.
Read the classical-battery audit → · Machine-readable record →