# -*- coding: utf-8 -*- """Recount the seventeen muqaṭṭaʿāt cases under all eight counting conventions. Text: the Uthmani text shipped with quran-json (Tanzil lineage). Scope rules, fixed across conventions and chosen because they reproduce the chapter's four convergent counts exactly: * the basmala heads every sura in scope (it is the sura's first line in the Uthmani text) — without it cases 3 and 4 fall 1 and 35 short; * ة is a tāʾ, not a hāʾ; ى is a written yāʾ unless a convention says otherwise; hamza on the line (ء) is not an alif. Two documented variant readings, applied in every convention because they are textual, not orthographic: * Ṣād cluster — the Tashkent codex reads bas(s)ṭatan at 7:69 with sīn: −1 ṣād. * Nūn (68) — the Ibn ʿAbbās reading spells the opening نون: +1 nūn. """ import json, re Q = json.load(open('package/dist/quran.json')) SURA = {c['id']: ''.join(v['text'] for v in c['verses']) for c in Q} BASMALA = 'بِسۡمِ ٱللَّهِ ٱلرَّحۡمَٰنِ ٱلرَّحِيمِ' WASLA, DAGGER, SHADDA, HAMZA_YA, MAQSURA = 'ٱ', 'ٰ', 'ّ', 'ئ', 'ى' ALIF_BASE = 'اأإآ' MARKS = set(range(0x064B, 0x0660)) | {0x0670, 0x0653, 0x0654, 0x0655, 0x0656, 0x0657, 0x065E, 0x0640} | set(range(0x06D6, 0x06EE)) SIMPLE = {'l': 'ل', 'm': 'م', 'r': 'ر', 'sad': 'ص', 'n': 'ن', 'q': 'ق', 'ayn': 'ع', 'sin': 'س', 'h': 'ه', 'k': 'ك', 'ha': 'ح', 'ta': 'ط'} # ta' is needed only for the Ta-Sin-Mim / Ta-Ha control cases # Rule 2 (Appendix D, II): ta' marbuta (ة) is counted as ha'. In the undotted # rasm ة and ه are one glyph, and at pause ة is uttered as ha'. Earlier runs of # this script omitted the rule; only cases that count ha' are affected # (K-H-Y-A-S in sura 19, and the Ta-Ha control). RULE2 = False reproduces the # omission for comparison. RULE2 = True MARBUTA = 'ة' # wasla, dagger, shadda, hamza-on-ya CONV = {'T1': (1, 1, 1, 1), 'T2': (1, 1, 0, 0), 'T3': (1, 0, 1, 1), 'T4': (1, 1, 0, 1), 'T5': (0, 1, 0, 0), 'T6': (0, 0, 0, 0), 'T7': (1, 0, 0, 0), 'T8': (1, 0, 0, 1), 'K': (1, 0, 0, 1), # ya rule matches T8; alif rule extended below # P is exploratory: the alif as a sound rather than a glyph. Every # written ā counts (base alif, madda, dagger alif); the waṣla alif is # elided in continuous recitation and the otiose alif of the wāw group # is silent, so neither counts. Non-alif letters are unaffected. 'P': (0, 1, 0, 0)} PHONETIC = 'P' # K is Khalifa's published family rule (submission.org, "How to count the # Quranic initials"): alif glyphs + wasla + madda + the free-standing hamza, # counted as alif; hamza on a ya chair counts as ya, on a waw chair as waw; # no dagger alif. Verified against the Submitters' 2002 published tables: # lam/mim/ra match exactly in all eleven contested suras; the alif/hamza # column lands within 2-23 per sura (Tanzil vs King Fahd hamza-seat encoding). FAMILY = 'K' OTIOSE = re.compile('\u0648[\u064B-\u0652\u06D6-\u06ED]*\u0627(?![\u0621-\u064A])') def count(text, letter, conv): wasla, dagger, shadda, hamzaya = CONV[conv] if letter == 'alif': hits = set(ALIF_BASE) | ({WASLA} if wasla else set()) elif letter == 'ya': hits = {'ي', MAQSURA} | ({HAMZA_YA} if hamzaya else set()) elif letter == 'h' and RULE2: hits = {'ه', MARBUTA} else: hits = {SIMPLE[letter]} n = 0 for i, ch in enumerate(text): if ch not in hits: continue n += 1 if shadda: # a shadda'd letter is two letters j = i + 1 while j < len(text) and ord(text[j]) in MARKS: if text[j] == SHADDA: n += 1 break j += 1 if letter == 'alif' and dagger: n += text.count(DAGGER) if letter == 'alif' and conv == FAMILY: n += text.count('\u0621') # free-standing hamza counts as alif n += text.count('\u0654') + text.count('\u0655') # seat-mark hamzas if letter == 'alif' and conv == PHONETIC: n -= len(OTIOSE.findall(text)) # the silent alif of ـوا n -= text.count('أ') + text.count('إ') # hamza carriers, not a long ā return n CASES = [ (1, [50], ['q']), (2, [42], ['q']), (3, [42], ['ayn', 'sin', 'q']), (4, list(range(40, 47)), ['ha', 'm']), (5, [7, 19, 38], ['sad']), (6, [68], ['n']), (7, [36], ['ya', 'sin']), (8, [19], ['k', 'h', 'ya', 'ayn', 'sad']), (9, [2], ['alif', 'l', 'm']), (10, [3], ['alif', 'l', 'm']), (11, [29], ['alif', 'l', 'm']), (12, [30], ['alif', 'l', 'm']), (13, [31], ['alif', 'l', 'm']), (14, [32], ['alif', 'l', 'm']), (15, [7], ['alif', 'l', 'm', 'sad']), (16, [13], ['alif', 'l', 'm', 'r']), (17, [10, 11, 12, 14, 15], ['alif', 'l', 'r']), ] # Control cases (Appendix D III.b): sign-framed openings that contain NO alif. # These are not among the seventeen the Code 19 literature claimed; they are the # only place where 'opening type predicts convergence' and 'alif-freedom predicts # convergence' make different predictions. CONTROLS = [ ('Ta-Sin-Mim 26+28', [26, 28], ['ta', 'sin', 'm']), ('Ta-Sin 27', [27], ['ta', 'sin']), ('TSM+TS 26,27,28', [26, 27, 28], ['ta', 'sin', 'm']), ('Ta-Ha 20', [20], ['ta', 'h']), ] # Khalifa's overlapping claim (Appendix 1, Table 6): ha' over 19+20, ta' over # 20,26,27,28, sin over 26,27,28, mim over 26,28 -> published 1,767 = 19 x 93. # Not one of the seventeen; a constructed set, reported for transparency. OVERLAP = [([19, 20], ['h']), ([20, 26, 27, 28], ['ta']), ([26, 27, 28], ['sin']), ([26, 28], ['m'])] VARIANT = {5: -1, 6: +1} PUBLISHED = {1: 57, 2: 57, 3: 209, 4: 2147, 5: 152, 6: 133, 7: 285, 8: 798, 9: 9901, 10: 5652, 11: 1669, 12: 1253, 13: 823, 14: 577, 15: 5313, 16: 1487, 17: 9483} table = {} DETAIL = {} print(f"{'#':>2} " + ''.join(f'{k:>8s}' for k in CONV) + f"{'published':>11s}") for n, suras, letters in CASES: text = ''.join(BASMALA + SURA[s] for s in suras) row = {c: sum(count(text, L, c) for L in letters) + VARIANT.get(n, 0) for c in CONV} detail = {c: {L: count(text, L, c) for L in letters} for c in CONV} for c in CONV: if VARIANT.get(n): k = 'sad' if n == 5 else 'n' detail[c][k] += VARIANT[n] table[n] = row DETAIL[n] = detail flag = '' if row['T8'] == PUBLISHED[n] else ' <-- differs' print(f'{n:>2} ' + ''.join(f'{row[c]:>8d}' for c in CONV) + f'{PUBLISHED[n]:>11d}{flag}') json.dump({str(k): v for k, v in table.items()}, open('counts19.json', 'w')) json.dump({str(k): v for k, v in DETAIL.items()}, open('counts19-detail.json', 'w')) print('\ndivisible by 19:') for c in CONV: hits = [n for n in table if table[n][c] % 19 == 0] print(f' {c}: {len(hits):2d}/17 cases {hits}') # ---- control cases: do sign-framed, alif-free openings converge? ---- print() print('CONTROLS (Appendix D III.b) - sign-framed openings with no alif') print(f"{'case':20s}" + ''.join(f'{k:>7s}' for k in CONV)) for nm, suras, letters in CONTROLS: text = ''.join(BASMALA + SURA[s] for s in suras) row = {c: sum(count(text, L, c) for L in letters) for c in CONV} print(f'{nm:20s}' + ''.join(f'{row[c]:>7d}' for c in CONV)) print(f"{' mod 19':20s}" + ''.join(f'{row[c] % 19:>7d}' for c in CONV)) # ---- Khalifa's overlapping set, and the effect of Rule 2 ---- print() print("KHALIFA'S OVERLAPPING SET (H 19+20, T 20/26/27/28, S 26-28, M 26+28); published 1,767") row = {c: sum(count(''.join(BASMALA + SURA[x] for x in suras), L, c) for suras, letters in OVERLAP for L in letters) for c in CONV} print(f"{'overlap set':20s}" + ''.join(f'{row[c]:>7d}' for c in CONV)) print(f"{' mod 19':20s}" + ''.join(f'{row[c] % 19:>7d}' for c in CONV)) RULE2 = False print() print('WITH RULE 2 SWITCHED OFF (ta marbuta not counted as ha) - for comparison only') for nm, suras, letters in [('K-H-Y-A-S 19', [19], ['k', 'h', 'ya', 'ayn', 'sad']), ('Ta-Ha 20', [20], ['ta', 'h'])]: text = ''.join(BASMALA + SURA[x] for x in suras) r2 = {c: sum(count(text, L, c) for L in letters) for c in CONV} print(f'{nm:20s}' + ''.join(f'{r2[c]:>7d}' for c in CONV)) print(f"{' mod 19':20s}" + ''.join(f'{r2[c] % 19:>7d}' for c in CONV)) r2 = {c: sum(count(''.join(BASMALA + SURA[x] for x in suras), L, c) for suras, letters in OVERLAP for L in letters) for c in CONV} print(f"{'overlap set':20s}" + ''.join(f'{r2[c]:>7d}' for c in CONV)) print(f"{' mod 19':20s}" + ''.join(f'{r2[c] % 19:>7d}' for c in CONV)) RULE2 = True