Date: 20 September 2026.
Result: no distinctive new quotation, source-order rule or justified cipher key. Thomas supplies a plausible inward/outward comparison. Crowley remains the stronger existing cluster of textual parallels. The comparison does not identify Cicada's religion, authors or intent. No unsolved-page decryption was attempted in this session.
Scope and editions
The automated comparison covers all 94 labeled entries, 727 alphabetic tokens, in the working solved-LP translation. This is a community transcription aid, not independent verification against every original rune. It is compared with these explicitly bounded sources:
| Source | Coverage | Extracted passages | Tokens |
|---|---|---|---|
| Thomas, attributed to Thomas O. Lambdin | Sayings 1–114 except 70 | 113 | 5,184 |
| Thomas, saying 70 attributed to Elaine Pagels | Kept separate | 1 | 36 |
| Dhammapada, Acharya Buddharakkhita | All 423 numbered verses; combined verses remain together | 413 | 11,139 |
| Corpus Hermeticum, G. R. S. Mead | Tractate I only, §§1–32 | 32 | 3,093 |
| Crowley, The Book of the Law | Chapters I–III; later Comment excluded | 220 | 5,878 |
The Thomas source explicitly substitutes Pagels for saying 70. Its Greek fragments, editorial commentary and duplicate webpage renderings are excluded. This is not treated as one uniform translation. The Nag Hammadi publication history places the English collection in 1977, but the exact print revision behind this webpage has not been authenticated.
The Dhammapada attribution identifies the 1985 edition and a 1996 transcription. Mead's title page dates the volume to 1906 and its digital preparation to 2009. These are older translations inspected through current copies, not authenticated snapshots of what the puzzle authors saw before 2014. Crowley is modern religious literature, not an ancient work; his earlier puzzle connection is documented separately in Session 15.
Exact wording test
Search every consecutive sequence of three through eight alphabetic tokens within each LP entry and each source passage. Uppercase text, remove punctuation and split contractions/hyphenated words into alphabetic tokens. Preserve spelling, including LP's KWESTION. No stemming, synonyms, translation switching or crossing passage boundaries. Counts below are distinct shared token sequences within each source, not independent observations or significance estimates.
| Source | Three-token sequences | Four-token sequences | Five through eight |
|---|---|---|---|
| Thomas, Lambdin-attributed passages | 20 | 1 | 0 |
| Thomas 70, Pagels-attributed passage | 2 | 1 | 0 |
| Dhammapada | 13 | 1 | 0 |
| Hermeticum I | 4 | 0 | 0 |
| Book of the Law | 6 | 0 | 0 |
Every shared sequence was inspected. The longest are ordinary constructions:
| Shared wording | LP entry | Source locator |
|---|---|---|
| “he went to the” | 0.2.0.2 | Thomas 64 |
| “what you do not” | 0.2.0.8 | Thomas 70, separate translation |
| “there is nothing to” | 0.3.0.14 | Dhammapada 317 |
The second example illustrates the method's limits: stripping punctuation joins wording across a clause boundary in LP. Four tokens alone are not a meaningful quotation. The absence of any five-token match also rules out longer exact matches under this tokenization and passage segmentation. It does not exclude paraphrase, another translation or words spanning our entry boundaries.
The three-token Crowley overlap about changing the book occurs at I:36. Its surrounding subject makes it more relevant than generic grammatical overlaps, but it is short and already belongs to the preservation theme identified in Session 15. An artificial copied four-word passage was recovered at the correct offsets, checking the matching procedure; this is not a control for detecting paraphrase.
Passage comparison and mismatches
These are qualitative readings, not statistical attribution scores.
Thomas. Sayings 3 and 22 combine interior/exterior imagery with self-knowledge or unity, making them reasonable counterparts to LP 0.1.0.6. Saying 13 has an identity dialogue, but its participants and progression differ from LP's student losing successive identity labels. Saying 110 connects wealth with renunciation, a possible comparison for LP's wealth-and-detachment instructions. Saying 1 promises avoidance of experiencing death, whereas LP 0.0.0.5 asks the reader to experience it. Metaphorical readings could reconcile that difference, so it neither proves nor disproves borrowing. No distinctive shared wording selects Thomas. Text
Dhammapada. Verses 160 and 165 emphasize self-mastery and responsibility for purification. Verses 275–279 combine an end to suffering, personal effort and non-self. These resemble parts of LP's pilgrimage and self-command language, but LP's claims about individual divinity and personal law cannot simply be equated with these teachings. The shared themes do not uniquely choose this text. Chapter 12, chapter 20
Hermeticum I. Sections 19–21 connect knowing oneself with divine origin; §§25–26 describe relinquishment during ascent; §30 joins closed eyes with inward awakening. LP likewise uses inward discovery and shedding constraints. However, this tractate's cosmic ascent and explicit divine framework are not an exact counterpart to LP's emergence parable. Other Hermetic tractates remain outside this audit. Mead's translation
Crowley. I:36 and I:54 concern preservation of lettering; I:56 treats words as sacred; II:54–55 connect letter value, alphabet order and replacement symbols. Together these remain unusually relevant to LP's opening warning and rune/number system. Yet they specify no unique 29-rune order, prime mapping, passage selector or decryption operation. II:3 also uses circumference, illustrating why a shared motif does not uniquely choose Blake or Crowley. Chapters I–III
Does the order help?
There is no uniquely defined mapping to test as a numerical sequence. We can still check selected, explicit assignments for simple order preservation:
| Candidate | Selected LP-order correspondences | What happens |
|---|---|---|
| Thomas | Inward/outward → identity dialogue → wealth/renunciation | Choosing sayings 22 → 13 → 110 gives a reversal. Choosing 3 instead of 22 removes it. The ambiguity itself prevents a unique coordinate sequence. |
| Dhammapada | Ending the self → self-command → programming the mind | Mapping these broadly to 279 → 160 → 1–2 runs backward. Inserting later LP inward-truth instructions requires new discretionary assignments. This is a thematic selection, not a demonstrated reverse traversal. |
| Hermeticum I | Inner divinity in LP's final parable | The same motif appears at several locations in the tractate, so it does not determine one source position. |
| Crowley | Preserve lettering → sacred words → individual sacred worth | I:36/54 → I:56 → I:3 is neither consistently forward nor consistently backward. Alternative assignments remain possible. |
These examples do not exhaust all alignments and are not a randomized sequence test. They show why the present correspondences cannot supply unambiguous book-cipher coordinates. Most of LP2 remains unsolved, so the order of the solved excerpts is not the full book's narrative order.
Decision and next constraint
Keep Thomas as a possible thematic influence and Crowley as an existing, stronger literary comparison. Do not promote either to a key on this evidence. There is no new passage reference, distinctive quotation or exact instruction that selects an operation. Trying arbitrary religious names, verse numbers or numerology would introduce choices the evidence has not earned.
Return the next cryptanalytic effort to the measured adjacent-repeat shortage. One useful bounded question is whether section-level differences in that shortage survive conditioning on each section's rune frequencies. That can distinguish a simple composition effect from an ordering constraint without guessing the plaintext's philosophy. It is a proposed next experiment, not a result of this session.
Audit files and limits
- Comparison code
- Counts, match locators and per-passage fingerprints
- Source URLs, extraction ranges and input hashes
The code checks complete numbered coverage for each selected source and excludes repeated webpage renderings. Full source books are not redistributed. A complete rerun requires local source captures in the documented formats; the published hashes and locators support checking the extracted passages but do not replace those inputs. English wording is translation-dependent. This is an exact-text scan plus a selected conceptual review, not an exhaustive semantic search or a test of ancient original-language ciphers.