No verified new plaintext. This pass follows the user's suggestion to investigate ancient languages and shared glyphs. It distinguishes historical script identification, possible plaintext language, and historically motivated key material.
Historical comparison
The closest alphabetic match is Anglo-Saxon/Anglo-Frisian futhorc. The distinctive final letters in the working Cicada alphabet, ᚪ ᚫ ᚣ ᛡ ᛠ, correspond to AC, AESC, YR, IOR, EAR in the Unicode Runic chart. The Old English Rune Poem tradition covers 29 runes. This is an existing historical connection, not a newly discovered Cicada solution. Old English is an early medieval language rather than a classical ancient language.
| Candidate tradition | Relevant comparison | Interpretation |
|---|---|---|
| Anglo-Saxon futhorc | Extended rune inventory, including AC/AESC/YR/IOR/EAR | Closest match to Cicada's alphabet |
| Elder Futhark | Earlier 24-rune system, sharing much of the core inventory | Historical relationship, not a complete 29-symbol decoding table |
| Younger Futhark | 16-rune system with different sound assignments | Shared shapes do not establish an Old Norse plaintext |
| Old Italic | Possible historical influence on runes | Does not provide a unique replacement alphabet or key |
| Latin | Tested as possible plaintext independently of visual similarity | No positive result in the bounded tests |
Unicode explicitly cautions that rune shapes vary, and similar shapes can carry different values in different traditions. It specifically notes the GER/IOR shape complication. This argues against choosing whichever historical reading makes an isolated fragment look meaningful. Cicada's own Gematria Primus chart and solved pages remain stronger evidence for the working symbol mapping.
Sources: - Unicode, historical scripts and Runic discussion: https://www.unicode.org/versions/Unicode16.0.0/core-spec/chapter-8/ (section 8.7). - Unicode Runic names/chart: https://www.unicode.org/charts/PDF/U16A0.pdf - Historical overview and links to editions of the Old English Rune Poem: https://en.wikipedia.org/wiki/Old_English_rune_poem - Scholarly facsimile edition landing page: https://oepoetryfacsimile.org/?document=8008 (the interactive edition text was not extracted or tested in this pass). - Cicada's published alphabet as archived by the solving community: https://www.cicadasolvers.com/wiki/gematria-primus/
Language-pattern test without an English scoring assumption
Reference material: 1. Old English Beowulf, Harrison and Sharp edition, Project Gutenberg ebook 9700. Extracted the poem body only, excluding modern introduction, section headings, appendix, glossary and notes. Hyphenated compounds joined. This is an editorial text, not an ancient runic inscription. The ebook is public domain in the USA. https://www.gutenberg.org/ebooks/9700 2. UD Latin Perseus training file. Read surface words, keeping multiword-token surfaces rather than double-counting their grammatical subdivisions. Contributors and sources are documented at https://universaldependencies.org/treebanks/la_perseus/index.html . Dataset license: CC BY-NC-SA 2.5, https://creativecommons.org/licenses/by-nc-sa/2.5/ . Original downloaded data retained unchanged and credited to Universal Dependencies / Perseus contributors.
Applied an explicit experimental transliteration to the 29-symbol working alphabet: accents removed, thorn/eth to TH, ash to AE, wynn to W, yogh to G; C/Q to K, V to U, Z to S. Tested merged digraphs and unmerged Latin-letter spelling separately. This is a reproducible computational convention, not a reconstruction of authentic historical pronunciation or spelling. The code documents the full mapping. In particular, NG is merged in the merged convention, but ING is not greedily collapsed.
The index of coincidence is the probability that two distinct, randomly selected symbol positions contain the same symbol. Any fixed one-to-one substitution preserves it exactly.
| Material | Index of coincidence |
|---|---|
| Eight substantial unsolved sections, individual range | 0.03408–0.03463 |
| Beowulf, merged reference windows, all tested sizes | 0.04984–0.06063 |
| Latin, merged reference windows, all tested sizes | 0.06692–0.07946 |
| Uniform 29-symbol benchmark | 1/29, approximately 0.03448 |
For each language and each section size, sampled 499 contiguous equal-length reference windows. Repeated with unmerged spelling. Every reference window in both conventions had a higher index than the corresponding ciphertext section. These overlapping windows are illustrative reference samples, not independent language-classification trials, and do not supply a calibrated probability of excluding a language.
All eight substantial ciphertext sections also contain no repeated complete word of four or more runes, as previously observed. That statistic likewise survives a fixed one-to-one symbol relabeling.
Interpretation: these sections do not resemble ordinary Old English or Latin made obscure solely by changing letter shapes. A changing substitution, homophonic encoding, substantial padding, alternate segmentation, or further cipher layer remains possible. No conclusion about all ancient languages follows from these two reference corpora.
Historically motivated key test
Used 29 normalized Anglo-Saxon rune names in Cicada order, based on the name forms recorded in Unicode. Added three explicitly speculative constructions: concatenated names, name letter lengths, and name rune lengths. Tested forward and reversed keys, every cyclic phase, continuous section starts and word resets, and four combinations of ciphertext/key signs modulo 29. The search covers 29,952 settings across eight sections; duplicate equivalent settings are retained. This is not an exhaustive search of historical spelling variants, rune-poem text, or possible ciphers.
Ranked each candidate separately under modern English, Old English, and Latin models. The historical-language models were trained on the first 80% of their reference words. Raw scores across different language models should not be compared as calibrated language probabilities.
Two constructed controls use withheld text encrypted with the FEOH key. Both recovered exactly at the best rank: 683 Old English runes and 799 Latin runes. These validate recovery of this simple key family on those samples, not the unknown cipher or every candidate key. No real-text leader produced convincing connected plaintext. Leading candidates for every section and language are retained for inspection.
What this leaves open
Rune names and their traditional associations could be instructions or key hints even if the plaintext is modern English. A more specific next investigation would tie a particular marginal illustration or printed clue to a particular historical name, then require the resulting rule to work on additional text. This pass has not established such an illustration-to-key link. It also has not tested Old Norse, Greek, Hebrew, Sanskrit, or Sumerian language corpora.
Reproduce using the primary runtime Python and historical_pass.py. Dependencies: NumPy and the saved local files. Results: historical_results.json. No background task is running.