Every change to the book since it was first published is listed here, newest first, as it was recorded when it was made. Corrections of fact or arithmetic and changes of interpretation are both recorded; the release notes say which is which for the larger revisions.
Release 1 — 22 October 2026
Frozen at releases/2026-10-22. The day’s revisions came from a book-wide fact-check by several AI reviewers and the author’s decisions on it.
- Corrections of fact. Errors of citation, date, anatomy, astronomy and geography were corrected across the book — among them the source of the seventeen literate Qurashīs (al-Balādhurī), the Bible Code history, Galen’s account of semen, the Fed’s mechanics, the geography of Adhriʿāt, and the 57-as-exact-midpoint claim (now “practical midpoint”).
- Corrections of arithmetic and counting rules. The 1-in-2,000 figure now carries its conditions everywhere it appears (about 1 in 29 counted strictly; rules written after the counts were known); the 1-in-150,000 figure is labelled a sensitivity; “six of six” is stated as four of four without the pending pair; the counting rules for “month” and “Qur’an” are stated.
- Changes of interpretation, made openly. The Pharaoh’s body is now a suggestive case, not a decisive one; Chapter 11 now says the classical substitution reading does conflict with history while the reading it favours does not; the iron verse is a consonance, not a demonstration; Chapter 17’s reading of dābbah is labelled an extension, with the commentators’ inclusion of angels and jinn as its ground.
- Definitions made book-wide. Ummī (Chapter 2), the five layers of the text (Chapter 3), and the shared definition of dābbah (Chapters 16–17).
- Site. Chapter 18 moved to ch-18-mysterious-letters.html (the old address redirects); a homepage introduction and three ways in; this page and the research index; claim boxes at the end of the evidence chapters; house spelling Qur’an, surah, Uthman, riba.
Full Record
10/22/2026
- Homepage entry and three ways in; For Researchers page with runnable code; claim boxes at the end of Ch 4-18; consistency pass
- Fact-check handoff applied book-wide; Ch 18 renamed to ch-18-mysterious-letters; house style
- Book-wide review batches 1-2: self-references, layer dates, cross-chapter contradictions
- Ch3: the five layers defined here (box + table: what each added, when, evidence); recitation = closest witness to Layer 0, held to the rasm (van Putten); skeleton fixed once c. 650; regional copies ~40–60 places, clean stemma (Cook, Sidky), Sidky's 284 dotting disagreements; at least four copies; dots from the start (24 AH inscription), al-Khalīl devised the signs; readings selected and mutually incompatible in detail; Wansbrough per van Putten, Shoemaker 2022; Birmingham Higham et al. 2018 + BnF 328c; Kufan verse count; 15:9 in an ALR sūrah; no 'critics generated' charge. Ch18 refers back (no madda date), Khalifa removed 9:128–129, Tashkent not cited for sīn. App A same Birmingham/Wansbrough edits; App D no 'millions'
- Ch3 fact-check: Habash's 49 = differences among ʿUthmān's regional copies (Cook 2004, Sidky 2020), not 14-century scribal variance; memorizers per Bukhārī 5003/4986; writing and recitation checked each other (4986, 4987, rasm condition); dots/vowels within ~150 years (al-Khalīl); seven readers' dates; 'every province'; ḥāfiẓ/qāriʾ as later terms, al-Ṭabarī on 15:9; tawātur as a ḥadīth/law category; Ḥāfeẓ's 'fourteen transmissions'; no census of ḥuffāẓ; Ṣanʿāʾ upper vs lower; coherence with Ch18 (archetype skeleton; both figures reported)
10/21/2026
- Ch18 + App D: letter stability measured (new II.b.2; Ch18 layers paragraph and cautions); first prediction reported as 3 of 6 stable non-sign openings (Ṭā-Hā, Ṣād-38, Nūn-68 fail); Nūn rung recast as stable letter one short; new closing section 'What the Puzzle Leaves'; oath split recorded in III.a as an after-the-fact hypothesis with its control and dependencies; ṣirāṭ counts corrected (5 in the Ṣād cluster, 13 not 15 in II.a.2); stale III.a and 'reported spelling' wording fixed
- Nūn-68: Ibn ʿAbbās report is a vowelling (nūni/nūna, Ibn Khālawayh ed. Bergsträßer p. 159), not a spelling; 133 is Khalifa's uncited spelled-out نون; case kept pending as a near miss (one letter short ≈ 2 in 19, not evidence); Rule 8 now meets one place (7:69); App D Sections III, II.a, IV tables, sources; Ch18 rung four; widgets
10/20/2026
- Ch4: sujūd figure — forelock and prefrontal cortex on the ground (96:16, 96:19)
- Ch4 fact-check: forelock verse is 96:15–16 (not 68:10–16), passage rebuilt on Sūrat al-ʿAlaq; six H–S–F verses (not sixteen) + the qalb contrast; Koenigs (six VMPC patients), Greene, deception imaging (Spence, Christ, Yang, Raine); Ta'if 619 and al-Ḥārith chronology; Abu Dawud 3875; conversion on both branches; bones/flesh qualified; translation movement dates; Zuhayr line, Saad & Said, Hawting, Ullmann, Damasio refs; widgets
- Ch5 recheck: Rojas opening (no names; matched in La Plata); Galton p. 167; al-Ṭabarī's own gloss quoted in Arabic; Khalifa and Sahih International wording of 75:3–4 and 50:16; 50:16 'unanimous' and jugular reading softened; duplicate (45:26) removed
- Ch5 fact-check: documented opening case (Rojas 1892); tafsir on 75:4 gives both classical readings; history (Grew–Faulds, Galton twins 1892); Lin 1982; clones (Shin 2002, Archer 2003); Hume essay 1777 + Locke; 45:24/45:26; NIST FpVTE figures; typos
- Ch18: Nöldeke gave up the owner-monogram theory by 1892 (GdQ II, 1919, p. 75)
10/19/2026
- Ch18: two manuscript plates — Ibrāhīm at 19:41 (Wetzstein II 1913) and the Ṭā-Hā opening (Tübingen Ma VI 165, Wetzstein II 1913)
- Ch18 + App D: Ṭā-Hā opening — no verse divider after طه in Birmingham 1572a, Tübingen Ma VI 165 or Wetzstein II 1913; later-hand heading is Wetzstein's
10/8/2026
- Ch18 + App D: Yā-Sīn and K-H-Y-ʿA-Ṣ pending; strict figures lead
10/7/2026
- App D: restore van Putten's 'the rasm is not a reading' with its source (Laysa r-rasmu bi-qirāʾah!, 11 Dec 2023)
- Ch18 review edits: counted object defined once; frozen frame = no post-run selection; totals recovered through the layers; ending conditional; earlier-chapters concession narrowed; 17:36 nuance kept
- Ch18: Khalifa's Appendix 1 checked — Tashkent cited for 7:69 only; uncited 'original text' claim for Nūn noted
- Ch18: Khalifa took the challenge and stopped — no complete data set; the point does not depend on him; study further
- Ch18 opens on Qāf 50:1 and Alif Lām Rā 12:1 with commentary before the fourteen-century question
- Ch18: opening section renamed The Puzzle: A Fourteen-Century Mystery
- Ch18: Khalifa as the man who first claimed to have solved the mystery — the criticism, the fitnah fulfilled, the controversy as the challenge
- Ch18: pre-registration defined and restored — five numbered pre-registrations, all twenty-nine openings by form, term carried through the results
- Ch18 rewritten top to bottom: mystery → puzzle anchored → checksum hypothesis → transmission and philology → the letters that never moved → down the ladder → the alif → the labels → chance → close; detailed material moved verbatim to App D Section X
10/6/2026
- Ch18 Step Three reordered: How a Recitation Became a Page (chain, five layers, van Putten) before How Do We Test It
- Ch18: the puzzle as a transmission check — hypothesis restated (72:28, 15:9, 74:31), the stability gradient, the sign on the hardest letter with its fork
- Ch18 + App D: yāʾ tier stated at its strength (287/795 carried), 3 of 14 beside 5 of 14, framing fixes, Step Six sentence cut, housekeeping; App D II.e van Putten in his words
- Ch18 Pre-registration 5: introduce Rashad Khalifa at first mention
- Ch18 hook: the sūrah that sets the puzzle is al-Muddaththir, The Cloaked One
- Ch18: the 74:31 question is anyone's response to a puzzle, not the dismisser's mark
- Ch18 Pre-registration 3: drop the Khalifa aside on kitāb marqūm
- Ch18 close: a puzzle built to run fourteen hundred years — the duration is the design
- Ch18 close: a puzzle is evidence of intent; whose intent is a step evidence cannot take; the last step is trust and belief
- Ch18 close: a serious challenge is solvable, with tools mankind built; why not lay it out plainly — The Robot answers
- Ch18: retitled The Mysterious Letters Present a Puzzle to Mankind; hook, The Case Stated, cube, philology
10/5/2026
- App D VIII.E: the lower text closes al-Tawba after 9:129, by name
- Ch18 Marcel 3: three precisions
- Ch18 Pre-registration 2: 74:30 has a seventh-century witness, stated with its limits
- Ch18 Ta-Ha: the classical positions stated at their real weight, with pages
- Ch18, How Do We Test It?: accepting the challenge requires academic philology
- Ch18 bridge: the freedom is the design, not a leak
- Ch18 bridge: the sign-label plank, and the comparison stated before any count
- Ch18: the counting as Pre-registration 3; the bridge rewritten on nineteen as prime
- Ch18: sixteen chapters -> seventeen, both occurrences
- Ch18 restructure: the argument first, the method to Appendix D
- Dates: index 'Current as of 10/4/2026'; chips and page headers for the five edited pages; chapter-updates.json
- Ch18 citations: pinpoint pages and exact wording
- Ch18 reread: nine corrections
- App D VIII.E: the manuscript case against the 9:128-129 excision
- Literature audit: correct four errors, narrow two priority claims, answer the 74:31 gloss
9/7/2026
- App D VII: the Gold Koran is a ninth-century Kufic luxury copy, not an early witness — every codex the appendix uses for the alif is older; the recount takes its orthography at twelve positions. Ch18 Step Six: what rides on the alif inventory — exact agreement with Khalifa wherever the letter is settled, failure only on the letter that is not; the two stories that would distinguish, and the claim the chapter does not judge
- Ch19 closing litany: 'The letters. The signs that divide.' — the last 'The proof' naming Chapter 18
- Ch19: its four descriptions of Chapter 18's evidence aligned — 'the proof the previous chapter presented' -> the puzzle; 'the verse that specifies the proof' -> names the number; 'mathematical proof' -> a count set as a puzzle; 'present across the full corpus of initialed surahs and extraordinarily unlikely' -> the frozen frame at its measured size, with the control's verdict on the historical form
- App D: history is evidence, labelled — the conditional sister-test product restored with its label
- App D VIII + V.E + Ch18: multiplication only inside a fixed, exhaustively scored frame
- Ch19: Iblis as doubt personified — vows to spread it (7:16), hands it to Adam as 'God is withholding' (7:20), the whisperer (114:4); his invitation the counterfeit of 17:36. And the creature's condition: all knowledge a belief on evidence; absolute knowledge with God, reached as far as He permits (2:255); the gap from there is trust
- Ch19 Iblis passage: the refusal is irrational only on the surface — a verdict on the Judge is a denial of the Judge (kafirin, 2:34); no creature can verify that God is God; doubt is set aside by trust, and obedience — willful trust — is the anchor. 'Overwhelming evidence', never 'perfect knowledge'. Ch18's three Iblis passages aligned: evidence complete, trust withheld; a proof produces facts, not trust
- Ch19 + App D: the previous chapter is The Signs That Divide — four stale 'Proof' references replaced; Ch19's back-reference restated as the puzzle that yields certainty to the examiner and cannot compel
- Ch18, The Paradox of Choice: why set a puzzle at all — meant to be solved, not solved for you; the solution is the solver's own (Iblis: knowledge is not assent); and, from the controls, the solution is a method — visible where the text makes the choices, noise where the reader makes them
- Intro + Ch17: the Ha-Mim subgroup cut described accurately (40-42 = 1,121, 43-46 = 1,026; one of six cuts, ~1 in 3 given the total; not used) and digit sums, the Bonferroni x1,800 and the 'one in 500,000' removed — all four passages described a Chapter 18 that no longer exists
- Widgets: Nūn-68 is pending in every figure (statbox 7/7 + 1 pending; row chips; declarations card 132/133 §); Ṣād row labelled as recited; alignment panel replaced with the chapter's letter-group / formula-only ledger tables (1 in 5,000 / 1 in 5) instead of the retired 1-in-110; Monte Carlo note and count-hint ranked under the frozen frame and V.E
- Ch18 + Intro: the invitation — the text asked 'what does God mean by this?' and placed its signs; the puzzle is handed to the reader willing to examine history, philology and the arithmetic
- Introduction: brought fully into line with Chapter 18's account of analyst freedom
- Ch18 + App D + Intro: a control on the investigator, not the text
- App D V.E + Ch18: the pattern-discovery controls, and the pre-registration finding
- App D Rule 8 + Ch18: why the 21:88 precedent does not settle Nun-68
- Ch18 + App D: canonical ledger vocabulary and no cross-axis products
- Ch18: the reciters' axis stated in the chapter body — the Qunbul objection raised and answered in four places, each referring to App D II.a.2; Hafs named as the fixed reference reading with the reason; Q21:88 cited as the external precedent; stale 'the primary convention cannot yet be applied' clause on the alif cases corrected
- App D: recited-consonant audit (prompted by van Putten on Q21:88). New II.a.2 — the reciters' axis: Qunbul recites sin for sad in sirat at 15 words and moves four counts; Qaf-50, Qaf-42 and Ha-Mim hold on the page and in every canonical recitation. Corrects II.a's unqualified 'no canonical reading moves any of the eight counts'. Rule 8 now names Hafs as its reference reading and cites Q21:88 as the external precedent, with substitution / omission / defective-long-vowel told apart. Sad-cluster readers named; small-ya figures disambiguated; audit script published
- App D IV.0: data dictionary and the canonical unit table (22 units, both frames, rarity/alignment eligibility, rule-known-before-count); Ch18 scoreboard names three quantities incl. discovery-adjusted surprise (unknown); likely conclusion labelled the chapter's judgment, not a computation
9/6/2026
- Ch18: 'metric' removed from claims about the text; 74:31 paragraph opens on reception vs arithmetic; preservation superlative softened; ending no longer treats scepticism as fulfilment
- Ch18: 74:30 never called a sign; number/targets/classification/operation vocabulary; the four links labelled explicit/explicit/inference/hypothesis; reception-division vs arithmetic-division kept apart; self-sealing bridge sentence replaced; alif fork tightened
- Ch18: the pre-registration section split into textually fixed vs analyst-supplied; 'nineteen is an ayah' removed; 17:36 no longer 'makes mandatory'; 'any reasonable null model' -> 'the null models tested here' (chapter and App D)
- 17:36 displayed literally in the Introduction and Chapter 18; the singular address (laka) argued in the ladder; the strong reading kept as the chapter's conclusion, not as the verse's translation
- Ch18/App D: text fixes the ingredients and the outcome classification, the chapter supplies the operation; 17:36 shown literally; nineteen described as fitnah not sign; App D III.c the twenty-nine-opening test with rules frozen before the run (14 cases, 5 hold, no sign case holds; 1 in 36 / 1 in 16 / 1 in 1,850); supplementary scripts published
- Ch18: textually constrained before measurement, with the puzzle answer (building the bridge is accepting the challenge); letter-group rule stated as articulated after the results, 1-in-5 pinned beside 1-in-5,000; pattern-discovery control adopted as future work; falsification box; 17:36 translation-interpretation-application ladder
- Ch18: puzzle premise and subtitle; six-step restructure; closing transition to The Robot; Grok patches (vocabulary, cuts, Sad set aside, two sign tables, fork, five-trial ledger); seven hold one pending with the three-line scoreboard
9/1/2026
- Code 19 explorer: Ta-Sin-Mim and Ta-Sin numbered 18 and 19
- Ch18 Part Two and The Pattern of the Division tightened; per-case detail deferred to Appendix D
- Ch18 Part One rewritten: history that supports counting, the eight in tiers, brief entries; details to Appendix D
- Code 19 explorer: primary convention = interpreted rasm, named as the closest approximation to what was recited; standard edition as the comparison; sensitivity grid on request; Khalifa's totals reported as published with a disclaimer; K dropped from the selector
- Ch18 conclusions: lead with the likely conclusion (the text pre-registered both the secure and the insecure letters; probabilities cited), coincidence as the alternative
- Code 19 explorer: split into two tables — signs (none at the claimed total) and oaths/declarations (all eight); Ta-Ha and Khalifa's set under the outside-the-seventeen toggle
- Ch18/App D: Ta-Ha placed in Form D provisionally — letters or a single word; note in both classifications
- Ch18: forms classify openings, the sign grouping classifies letters — stated in both places; 'already' means elsewhere in the text
- Ch18/App D: letter-group sign rule (letters already pre-registered as a sign in another chapter); alif cases reported as results under the primary convention; sign table at two scopes; Ta-Ha as the one non-sign failure
- Rename Chapter 18 to The Signs That Divide (new slug, redirect from old URL, sitemap)
- Ch18/App D: manuscript-and-treatise test (App D II.d); Hafs recites 7:69 with sin — Sad cluster 152 as recited, 153 as written; tiered probabilities; Nun-68 pending
8/31/2026
- Ch18 + App D premise rewrite: primary convention = the Uthmanic archetype as the tradition identifies it (van Putten stemmatics/Hijazi phonology, Bursi dots), T8 + Rule 2; multi-convention grid demoted to sensitivity analysis; 8 of 8 non-alif cases hold (Ya-Sin, KHYAS direct), 9 alif cases undetermined at the source; V.A six trials ~1 in 55M with four-trial floor ~1 in 150k; V.D [[5,0],[3,9]] ~1 in 110; invariance test extended to 13 suras and ten letters (three marginal flips named, 68:42 recorded); App D sections reordered (II.a, II.b, II.c, III.a-c); widgets updated
- Rule 2 (ta marbuta as ha) applied in the multi-convention grid: K-H-Y-A-S reproduces its claimed 798 under T4/T8/K and is reported rasm-verified alongside Ya-Sin; interactive table gains a Rule 2 toggle and an outside-the-seventeen block (Ta-Ha 279, TSM 1192, TS 121, Khalifa overlap set 1767); declarations widget labels corrected; ha letter-battery figure 426
- Ch18 Ta-Ha paragraph corrected: Khalifa did publish Ta-Ha's letters inside the overlapping 1,767 claim (reproduces only with ta marbuta as ha; 1,705 under chapter rule); ha across 19-20 is 426 in his tables, 364 in ours; Sura 20 is Form D not Form C; reframed under the one-directional principle
- Appendix D III.a: Form D reclassification concessions removed; window admission confined to the alternative exclusive cuts (sign-formula sits within verses 1-2 in all eight cases)
- Ha-Mim openings classified on wording alone: 40/45/46 Form D (tanzil al-kitab), 41 Form D (tanzil), 42 declaration by wahy, 43/44 dual; arithmetic no longer used to justify any form
- Remove all earlier-edition/version references, Appendix D included
- Ch18 How Do We Test It: van Putten's June 2021 preservation thread added (rasm stability, unknowable spellings at 8:65, Ya-Sin consensus, preservation vs truth) with reference; 'access to the source' softened to match Layer 0 as undefined
- gitignore _to_delete
- Ch18 Five Layers: Layer 0 defined as undefined - approximated from the surviving layers, not directly observable; numbering from zero explained
- Ch18: four opening forms recast as four pre-registrations; the ayat pre-registration is the one 74:31 ties (via fitnah, cf. 3:7) to the number and to division; fixed misstatement that 74:31 uses the word ayat; 43/44 dual Form B+D; earlier-version references removed from chapter; form overlap stated (16 of 29 openings multi-form)
- Chapter 18 reframed: signs divide, what is not named a sign is free to hold. Sign-language restricted to the eight tilka-ayat openings (Form B extension withdrawn); oath and descent lists corrected (Ta-Ha removed from oaths, 43/44/68 added); the prediction stated as one-directional; 12/43 minimal pair added; all six convergent cases confirmed to carry no sign-formula, eight of eight sign-framed openings contested.
- Resolve the Form B / Form D contradiction for suras 43 and 44: classified Form D on a three-verse window via the declarations at 43:3 and 44:3, with the two-verse alternative stated. Adds the window dependency and the Form B split to Appendix D III.a as a third admission.
- Appendix D III.c: the six oath openings compared; no feature distinguishes Ya-Sin, so its failure is a counting decision not a textual one. Frames the hamza-ya question as category uniformity, reports rather than adopts.
- Appendix D III.b: correct the opening classification by formula (0 of 8 sign-framed openings converge); state the alif confound at full strength; add the Ta-Sin-Mim / Ta-Sin / Ta-Ha controls, which are sign-framed and alif-free and fail under all ten conventions
- Appendix D II.b: how much of each count is independent of the dotting layer (HM 97.9%, all six 94.9%). III.a: descent and sign cuts tested and reported null; the convergent category named accurately as oath openings; the five-verse window stated as a live degree of freedom.
- Layer 2 corrected: i'jam is ancient and contemporary with the rasm, not a late invention; only vowel marking is later. Cites Bursi; adds a facsimile plate showing both ink systems.
- Appendix D: rasm+oral / rasm+dots as the best approximation to Layer 0; invariance test; promote the Tashkent sin, suspend Nun-68 pending the orthography question. Withdraw the attack page from the site (now a local draft).
- Appendix D: pronunciation counts are Hafs counts — naql, ishmam and the reading traditions; add Warsh recount as attack-page claim M6
- Transliteration: apply van Putten's Classical Arabic rules — 3ms suffix length, qat/wasl hamza after the article, mina + assimilated article, ana vs ana
- Citation audit: correct Gogtay, Hepper, Dobson-Teller, Kucken-Newell and Lebel claims across Ch4, Ch5, figures and attack page; refine Appendix D canonicity principle after van Putten's SST post
- Untrack _to_delete scratch files and ignore the folder
- Attack page: 21 falsifiable claims, refutation conditions, counting script, and a weakest-points section
8/24/2026
- Ch5: the fingertip plate — banan visible without instruments
- Ch5: the visibility claim the reader can check — expanded and highlighted
8/23/2026
- Ch5: the harder comparison — iris vs fingertip, four properties, honest scoping
- Ch4: replace brain-sequence figure with cited interactive timeline; age-40 / 46:15 passage
- Ch4: why the translations say clot — lexicon vs tafsir history, with Lane/Tabari/Zuhayr refs
- Ch4: leech-function interactive, anticoagulant battery, Moore credit on the comparison photo
- Ch4: dual meaning of alaq — clinging tissue becomes the placenta, feeds on maternal blood
- Ch6: substrates interactive — where the fluid comes from
- Title the ultrasound plate: Alaq - A Thing That Clings
- Chapter 1: the alaq ultrasound plate, metadata stripped
- Ch19: Iblis's irrationality as proof of will; doubt of God, not of His existence; iman as trust
- Wording: God, who claims to be the divine author of the Qur'an
- Alif-forms panel with van Putten quote; spellings-residue panel
- Declarations contrast figure, 74:31 layer, and the Fisher bridge panel
- Bold the localization sentence; add The Fault Line Is One Letter interactive
- Chapter 18: the am-I-examining-the-source paragraph
- The alif capstone: the one unstable letter, and the dataset that would decide it
- Unanimous letter-name recitation: the oral layer on the muqatta'at targets
- Van Putten scholarship: rasm primacy, alif orthography resource, Rule 8 candor note
8/21/2026
- State the 74:30-31 linkage directly at the end of finding (a)
- Iblis paragraph: rejection is volitional, hence the dividing sign
- Fitna as functional specification: the layered sign argument
- Correct the falsifiability claim: perfect evidence would not falsify the trial clause
- Appendix D IV.b: invariant-letter battery; the lam total
- Fix RTL over-wrap of mixed paragraphs; say -> specify
- Add Ta-Ha folio figure to Chapter 18; restore the four dropped figures
- Ta-Ha vocative caveat with early-codex manuscript witness
- Add the Ya-Sin signpost reading to Chapter 18
- Golden star on updated chapters
- Add Clear Proof and Dividing Proof section to Chapter 18
8/20/2026
- 74:30-31 accuracy: pronoun referent, misquote, metric-as-choice
- Ta-Ha transparency: unclaimed openings computed and recorded as non-divisible
- Highlight Ya-Sin as rasm-verified; add why-not-rasm-standard passage
- Fix Test 1 transparency point: subtracting 19 preserves divisibility
- Audit Ch18+App D; verify and publish Khalifa's counting method (convention K)
- Divisible column: check or X
- Letter counts inline in their own column
- Per-letter counts; Appendix D IV.a locates the divergence in the alif
- Recount all eight counting conventions from the Uthmani text
- Make the counting-convention selector report only what was recounted
- Give each disconnected-letters verse its own card and Arabic
- Fix reference formatting, restyle inline verse cards, remove master lens toggle, add per-box lens prompt
- Fix stray refs, Bible links, punctuation
- Arabic pilot: RTL edition, language toggle, Arabic-aware search, Chapter 1 translated; fix skip-link that broke RTL layout
8/19/2026
- Raise light-mode contrast above the accessibility floor and soften body text; same palette, stepped
- Rewrite chapter 2's closing line and one assertion in chapter 8; fix the 'n either' typo
- Consistency pass: replace flat assertions of illiteracy with the access and orality framing across all chapters
- Relocate the paradox of the pen to the culture's transformation; revise the book's conclusion; add ummi-debate sources; make every verse reference clickable; add update stamps and living-document note
- Retitle chapter 2: The Prophet in a Culture Without Books
- Chapter 2: state the ummi dispute honestly; rest the argument on access rather than literacy
- Point canonical URLs, social previews and sitemap at the live hostname
- Chapter 19 and Appendices A-C, chapter 2 illustrations, riba charts, larger type
- Read — web edition