lambelan 0.1.0__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -0,0 +1,38 @@
1
+ Metadata-Version: 2.4
2
+ Name: lambelan
3
+ Version: 0.1.0
4
+ Summary: Transliterate Latin text into the Everplay house script (Lambelan, a Tengwar mode) — encoded strings, gematria, and the entered lexicon.
5
+ Author: Mitchell Davis Burns / EVERPLAY
6
+ License: MIT
7
+ Requires-Python: >=3.9
8
+ Description-Content-Type: text/markdown
9
+
10
+ # lambelan
11
+
12
+ Transliterate Latin text into the Everplay house script — **Lambelan**, the
13
+ house dialect written in its own Tengwar mode — and compute a word's
14
+ gematria (its letters as harmonic partials, summed).
15
+
16
+ ```python
17
+ import lambelan
18
+
19
+ script, total, partials = lambelan.transliterate("Silomielle")
20
+ # script: the word in registry Tengwar encoding (render it with the
21
+ # Everplay House Script Concept font, or any registry Tengwar font)
22
+ # total: 135 — partials: [28, 30, 21, 2, 26, 28]
23
+
24
+ lambelan.translate("The research and the record") # per-word results
25
+ lambelan.lexicon()["Kol"] # the entered lexicon
26
+ ```
27
+
28
+ Rules of the mode (of record, v1): transliteration, not translation — the
29
+ script writes the Latin spelling word for word and every letter is written.
30
+ Vowels are marks (tehtar) riding the letter before them; a long vowel
31
+ (á, ú, û) doubles its mark; in names, `3` is a letter of the script — the
32
+ sigil, value 3, read as *e* (so `DUD3` is written with the sigil in the
33
+ e place).
34
+
35
+ The lexicon of record lives at lexicon.llmtech.net and in the Lambelan
36
+ Lexicon sheet; this package carries the same entered words.
37
+
38
+ © Mitchell Davis Burns / EVERPLAY
@@ -0,0 +1,29 @@
1
+ # lambelan
2
+
3
+ Transliterate Latin text into the Everplay house script — **Lambelan**, the
4
+ house dialect written in its own Tengwar mode — and compute a word's
5
+ gematria (its letters as harmonic partials, summed).
6
+
7
+ ```python
8
+ import lambelan
9
+
10
+ script, total, partials = lambelan.transliterate("Silomielle")
11
+ # script: the word in registry Tengwar encoding (render it with the
12
+ # Everplay House Script Concept font, or any registry Tengwar font)
13
+ # total: 135 — partials: [28, 30, 21, 2, 26, 28]
14
+
15
+ lambelan.translate("The research and the record") # per-word results
16
+ lambelan.lexicon()["Kol"] # the entered lexicon
17
+ ```
18
+
19
+ Rules of the mode (of record, v1): transliteration, not translation — the
20
+ script writes the Latin spelling word for word and every letter is written.
21
+ Vowels are marks (tehtar) riding the letter before them; a long vowel
22
+ (á, ú, û) doubles its mark; in names, `3` is a letter of the script — the
23
+ sigil, value 3, read as *e* (so `DUD3` is written with the sigil in the
24
+ e place).
25
+
26
+ The lexicon of record lives at lexicon.llmtech.net and in the Lambelan
27
+ Lexicon sheet; this package carries the same entered words.
28
+
29
+ © Mitchell Davis Burns / EVERPLAY
@@ -0,0 +1,18 @@
1
+ [build-system]
2
+ requires = ["setuptools>=68"]
3
+ build-backend = "setuptools.build_meta"
4
+
5
+ [project]
6
+ name = "lambelan"
7
+ version = "0.1.0"
8
+ description = "Transliterate Latin text into the Everplay house script (Lambelan, a Tengwar mode) — encoded strings, gematria, and the entered lexicon."
9
+ readme = "README.md"
10
+ requires-python = ">=3.9"
11
+ license = { text = "MIT" }
12
+ authors = [{ name = "Mitchell Davis Burns / EVERPLAY" }]
13
+
14
+ [tool.setuptools]
15
+ package-dir = { "" = "src" }
16
+
17
+ [tool.setuptools.package-data]
18
+ lambelan = ["*.json"]
@@ -0,0 +1,4 @@
1
+ [egg_info]
2
+ tag_build =
3
+ tag_date = 0
4
+
@@ -0,0 +1,41 @@
1
+ """Lambelan — transliterate Latin text into the Everplay house script.
2
+
3
+ The house script is a Tengwar mode (the Lambelan dialect) with its own
4
+ letterforms, carried in the Everplay House Script / House Clock concept
5
+ fonts under the ConScript Unicode Registry's Tengwar encoding.
6
+
7
+ >>> import lambelan
8
+ >>> lambelan.transliterate("Kol")
9
+ (script string, 33, [3, 4, 26]) # encoded string, gematria sum, partials
10
+
11
+ Transliteration, not translation (rule G1): the script writes the Latin
12
+ spelling, word for word, every letter written.
13
+ """
14
+ import json
15
+ from pathlib import Path
16
+ from ._core import tokenize, encode_word
17
+
18
+ __version__ = "0.1.0"
19
+ __all__ = ["tokenize", "transliterate", "translate", "gematria", "lexicon", "__version__"]
20
+
21
+ def transliterate(word):
22
+ """One Latin word -> (encoded script string, gematria sum, partials)."""
23
+ return encode_word(word)
24
+
25
+ def translate(text):
26
+ """A line of Latin text -> list of per-word transliteration dicts."""
27
+ out = []
28
+ for w in text.split():
29
+ script, total, partials = encode_word(w)
30
+ out.append({"word": w, "script": script, "sum": total, "partials": partials})
31
+ return out
32
+
33
+ def gematria(word):
34
+ """A word's gematria: (sum, partials) — its letters as harmonic partials."""
35
+ _, total = tokenize(word.replace("'", "").replace("’", ""))
36
+ toks, _ = tokenize(word.replace("'", "").replace("’", ""))
37
+ return total, [t[2] for t in toks]
38
+
39
+ def lexicon():
40
+ """The entered lexicon: {Latin: {section, gloss, units, partials, total, encoded}}."""
41
+ return json.loads((Path(__file__).parent / "lexicon.json").read_text(encoding="utf-8"))
@@ -0,0 +1,112 @@
1
+ """Lambelan core: the house mode's tokenizer and encoder.
2
+
3
+ Tokenizer of record: the mode's shared coverage tokenizer (mode spec of
4
+ record v1, 2026-11-20), vendored with its assignment inventory.
5
+ Encoder grouping rules: tehtar ride the preceding bearer; a vowel with no
6
+ bearer rides telco (short) or ara (long, single tehta); a long vowel on a
7
+ bearer doubles its tehta; silme/esse under a tehta take the nuquerna
8
+ allograph; a vowel after the 3-sigil stands on its own carrier.
9
+ Output encoding: ConScript Unicode Registry Tengwar block (U+E000-E07F).
10
+ """
11
+ import json, unicodedata
12
+ from pathlib import Path
13
+
14
+ _inv = json.loads((Path(__file__).parent / "assignments.json").read_text(encoding="utf-8"))
15
+ SOUND2NUM = {}
16
+ for _p in _inv["primaries"] + _inv["additionals"]:
17
+ SOUND2NUM[_p["sound"]] = _p["number"]
18
+ NUM2NAME = {p["number"]: p["name"].replace(" ", "_") for p in _inv["primaries"] + _inv["additionals"]}
19
+ TEHTA = {t["vowel"]: t["value"] for t in _inv["tehtar"]}
20
+ LONG_TEHTA = {t["vowel"]: t["long_value"] for t in _inv["tehtar"]}
21
+ MULTI = ["nkw", "ngw", "qu", "ng", "nt", "mp", "nk", "th", "hw", "ch", "sh", "ph", "gh", "dh"]
22
+ SINGLE = {"t": "t", "p": "p", "k": "k", "c": "k", "d": "d", "b": "b", "g": "g",
23
+ "f": "f", "h": "h", "n": "n", "m": "m", "r": "r", "w": "w", "y": "y",
24
+ "v": "v", "s": "s", "l": "l", "z": "z"}
25
+ DIACRITIC_LONG = {"á": "a", "ú": "u", "û": "u"}
26
+ DIACRITIC_PLAIN = {"ë": "e", "é": "e"}
27
+
28
+ CP = {"tinco": 0xE000, "parma": 0xE001, "calma": 0xE002, "quesse": 0xE003,
29
+ "ando": 0xE004, "umbar": 0xE005, "anga": 0xE006, "ungwe": 0xE007,
30
+ "sule": 0xE008, "formen": 0xE009, "aha": 0xE00A, "hwesta": 0xE00B,
31
+ "anto": 0xE00C, "ampa": 0xE00D, "anca": 0xE00E, "unque": 0xE00F,
32
+ "numen": 0xE010, "malta": 0xE011, "noldo": 0xE012, "nwalme": 0xE013,
33
+ "ore": 0xE014, "vala": 0xE015, "anna": 0xE016, "vilya": 0xE017,
34
+ "silme": 0xE01C, "silme_nuquerna": 0xE01D, "lambe": 0xE01A,
35
+ "esse": 0xE01E, "esse_nuquerna": 0xE01F, "hwesta_sindarinwa": 0xE021,
36
+ "yanta": 0xE022, "ure": 0xE023, "telco": 0xE025, "ara": 0xE026, "sigil": 0xE063}
37
+ TEHTA_CP = {"a": 0xE040, "i": 0xE044, "e": 0xE046, "o": 0xE048, "u": 0xE04A}
38
+
39
+ def tokenize(word):
40
+ """Return (tokens, sum); tokens are (grapheme, kind, value)."""
41
+ w = unicodedata.normalize("NFC", word).lower()
42
+ toks, i, prev_vowel = [], 0, False
43
+ while i < len(w):
44
+ ch = w[i]
45
+ if ch == "3":
46
+ toks.append(("3-sigil", "sigil", 3)); i += 1; prev_vowel = False; continue
47
+ if ch == "5":
48
+ toks.append(("5->s", "letter", SOUND2NUM["s"])); i += 1; prev_vowel = False; continue
49
+ if ch in DIACRITIC_LONG:
50
+ toks.append((ch, "tehta-long", LONG_TEHTA[DIACRITIC_LONG[ch]])); i += 1; prev_vowel = True; continue
51
+ if ch in DIACRITIC_PLAIN:
52
+ toks.append((ch, "tehta", TEHTA[DIACRITIC_PLAIN[ch]])); i += 1; prev_vowel = True; continue
53
+ if ch in TEHTA:
54
+ kind = "glide" if (prev_vowel and ch in "iu") else "tehta"
55
+ toks.append((ch, kind, TEHTA[ch])); i += 1; prev_vowel = True; continue
56
+ matched = False
57
+ for g in MULTI:
58
+ if w.startswith(g, i):
59
+ if g in ("sh", "gh", "dh"):
60
+ for c in g:
61
+ toks.append((c, "letter", SOUND2NUM[SINGLE[c]]))
62
+ elif g == "ph":
63
+ toks.append((g, "letter", SOUND2NUM["f"]))
64
+ else:
65
+ toks.append((g, "letter", SOUND2NUM[g]))
66
+ i += len(g); matched = True; prev_vowel = False; break
67
+ if matched:
68
+ continue
69
+ if ch == "x":
70
+ toks.append(("k", "letter", SOUND2NUM["k"])); toks.append(("s", "letter", SOUND2NUM["s"]))
71
+ i += 1; prev_vowel = False; continue
72
+ if ch == "j":
73
+ toks.append(("d", "letter", SOUND2NUM["d"])); toks.append(("z", "letter", SOUND2NUM["z"]))
74
+ i += 1; prev_vowel = False; continue
75
+ if ch == "q":
76
+ toks.append(("q->k", "letter", SOUND2NUM["k"])); i += 1; prev_vowel = False; continue
77
+ if ch in SINGLE:
78
+ toks.append((ch, "letter", SOUND2NUM[SINGLE[ch]])); i += 1; prev_vowel = False; continue
79
+ raise ValueError(f"uncovered character {ch!r} in {word!r} at {i}")
80
+ return toks, sum(t[2] for t in toks)
81
+
82
+ def _vowel_of(g):
83
+ return DIACRITIC_LONG.get(g) or DIACRITIC_PLAIN.get(g) or g
84
+
85
+ def encode_word(word):
86
+ """Return (script_string, gematria_sum, partials) for one Latin word."""
87
+ toks, total = tokenize(word.replace("'", "").replace("’", ""))
88
+ groups, prev = [], None
89
+ for grapheme, kind, value in toks:
90
+ if kind == "letter":
91
+ groups.append(["bearer", NUM2NAME[value], []])
92
+ elif kind == "sigil":
93
+ groups.append(["glyph", CP["sigil"], []])
94
+ elif kind == "glide":
95
+ groups.append(["glyph", CP["yanta" if _vowel_of(grapheme) == "i" else "ure"], []])
96
+ elif kind in ("tehta", "tehta-long"):
97
+ t = TEHTA_CP[_vowel_of(grapheme)]
98
+ if kind == "tehta-long":
99
+ if prev == "letter": groups[-1][2] += [t, t]
100
+ else: groups.append(["carrier", "ara", [t]])
101
+ else:
102
+ if prev == "letter": groups[-1][2].append(t)
103
+ else: groups.append(["carrier", "telco", [t]])
104
+ prev = kind
105
+ out = []
106
+ for kind, name, tehtar in groups:
107
+ if kind == "glyph":
108
+ out.append(chr(name))
109
+ else:
110
+ base = name + "_nuquerna" if tehtar and name in ("silme", "esse") else name
111
+ out.append(chr(CP[base]) + "".join(chr(t) for t in tehtar))
112
+ return "".join(out), total, [t[2] for t in toks]
@@ -0,0 +1,359 @@
1
+ {
2
+ "spec": "Everplay house mode - assignment inventory (hs.mode-spec parts a, b, d)",
3
+ "version": "0.3 (2026-11-20 units; hs.mode-spec parts a, b, c, d, e landed; spec of record filed whole remains)",
4
+ "legal": "Mitchell Davis Burns / EVERPLAY",
5
+ "numbering_rule": "Primary number = (grade-1)*4 + series index (I=1..IV=4); row-major by grade, Appendix E order. Re-derivable from the grid by anyone holding the mode (draft s.6).",
6
+ "primaries": [
7
+ {
8
+ "number": 1,
9
+ "name": "tinco",
10
+ "series": "I",
11
+ "grade": 1,
12
+ "sound": "t",
13
+ "note": ""
14
+ },
15
+ {
16
+ "number": 2,
17
+ "name": "parma",
18
+ "series": "II",
19
+ "grade": 1,
20
+ "sound": "p",
21
+ "note": ""
22
+ },
23
+ {
24
+ "number": 3,
25
+ "name": "calma",
26
+ "series": "III",
27
+ "grade": 1,
28
+ "sound": "k",
29
+ "note": "Latin c (hard) and k both write here"
30
+ },
31
+ {
32
+ "number": 4,
33
+ "name": "quesse",
34
+ "series": "IV",
35
+ "grade": 1,
36
+ "sound": "qu",
37
+ "note": "kw / qu, as in Quettalan"
38
+ },
39
+ {
40
+ "number": 5,
41
+ "name": "ando",
42
+ "series": "I",
43
+ "grade": 2,
44
+ "sound": "d",
45
+ "note": "voiced stop, first-class (draft s.1)"
46
+ },
47
+ {
48
+ "number": 6,
49
+ "name": "umbar",
50
+ "series": "II",
51
+ "grade": 2,
52
+ "sound": "b",
53
+ "note": "voiced stop, first-class (draft s.1)"
54
+ },
55
+ {
56
+ "number": 7,
57
+ "name": "anga",
58
+ "series": "III",
59
+ "grade": 2,
60
+ "sound": "g",
61
+ "note": "voiced stop, first-class (draft s.1)"
62
+ },
63
+ {
64
+ "number": 8,
65
+ "name": "ungwe",
66
+ "series": "IV",
67
+ "grade": 2,
68
+ "sound": "gw",
69
+ "note": "general-text labiovelar voiced stop"
70
+ },
71
+ {
72
+ "number": 9,
73
+ "name": "sule",
74
+ "series": "I",
75
+ "grade": 3,
76
+ "sound": "th",
77
+ "note": "Sindarin value (draft s.1: Sindarin th); Quenya s-value NOT used - s is silme, no.25"
78
+ },
79
+ {
80
+ "number": 10,
81
+ "name": "formen",
82
+ "series": "II",
83
+ "grade": 3,
84
+ "sound": "f",
85
+ "note": ""
86
+ },
87
+ {
88
+ "number": 11,
89
+ "name": "aha",
90
+ "series": "III",
91
+ "grade": 3,
92
+ "sound": "h",
93
+ "note": "the h of hud, heru, hosta-; all positions in v1"
94
+ },
95
+ {
96
+ "number": 12,
97
+ "name": "hwesta",
98
+ "series": "IV",
99
+ "grade": 3,
100
+ "sound": "hw",
101
+ "note": "labiovelar fricative, general text"
102
+ },
103
+ {
104
+ "number": 13,
105
+ "name": "anto",
106
+ "series": "I",
107
+ "grade": 4,
108
+ "sound": "nt",
109
+ "note": "prenasal stop as one letter, e.g. -cant may write its nt here"
110
+ },
111
+ {
112
+ "number": 14,
113
+ "name": "ampa",
114
+ "series": "II",
115
+ "grade": 4,
116
+ "sound": "mp",
117
+ "note": "prenasal stop as one letter"
118
+ },
119
+ {
120
+ "number": 15,
121
+ "name": "anca",
122
+ "series": "III",
123
+ "grade": 4,
124
+ "sound": "nk",
125
+ "note": "prenasal stop as one letter"
126
+ },
127
+ {
128
+ "number": 16,
129
+ "name": "unque",
130
+ "series": "IV",
131
+ "grade": 4,
132
+ "sound": "nkw",
133
+ "note": "prenasal stop as one letter"
134
+ },
135
+ {
136
+ "number": 17,
137
+ "name": "numen",
138
+ "series": "I",
139
+ "grade": 5,
140
+ "sound": "n",
141
+ "note": ""
142
+ },
143
+ {
144
+ "number": 18,
145
+ "name": "malta",
146
+ "series": "II",
147
+ "grade": 5,
148
+ "sound": "m",
149
+ "note": ""
150
+ },
151
+ {
152
+ "number": 19,
153
+ "name": "noldo",
154
+ "series": "III",
155
+ "grade": 5,
156
+ "sound": "ng",
157
+ "note": "the ng cluster as one letter (older ngoldo)"
158
+ },
159
+ {
160
+ "number": 20,
161
+ "name": "nwalme",
162
+ "series": "IV",
163
+ "grade": 5,
164
+ "sound": "ngw",
165
+ "note": "labiovelar nasal, general text"
166
+ },
167
+ {
168
+ "number": 21,
169
+ "name": "ore",
170
+ "series": "I",
171
+ "grade": 6,
172
+ "sound": "r",
173
+ "note": "house simplification: ore writes r in ALL positions; the Quenya romen/ore positional split is NOT adopted in v1"
174
+ },
175
+ {
176
+ "number": 22,
177
+ "name": "vala",
178
+ "series": "II",
179
+ "grade": 6,
180
+ "sound": "w",
181
+ "note": "labial glide, its early value (cf. wilya); the w of Wadesi"
182
+ },
183
+ {
184
+ "number": 23,
185
+ "name": "anna",
186
+ "series": "III",
187
+ "grade": 6,
188
+ "sound": "y",
189
+ "note": "consonantal y in all positions in the house corpus"
190
+ },
191
+ {
192
+ "number": 24,
193
+ "name": "vilya",
194
+ "series": "IV",
195
+ "grade": 6,
196
+ "sound": "v",
197
+ "note": "later Quenya value; the v of -var, Uvanost, Lanyavar"
198
+ }
199
+ ],
200
+ "additionals": [
201
+ {
202
+ "number": 25,
203
+ "name": "silme",
204
+ "sound": "s",
205
+ "note": "the house s (Silomielle, Cestila); silme nuquerna is its allograph above a tehta, same sound, same value 25"
206
+ },
207
+ {
208
+ "number": 26,
209
+ "name": "lambe",
210
+ "sound": "l",
211
+ "note": "ll is lambe written twice, values add (Silomielle, Quettalan)"
212
+ },
213
+ {
214
+ "number": 27,
215
+ "name": "esse",
216
+ "sound": "z",
217
+ "note": "voiced s; the letter's own predecessor aze carried z - house takes that value. ss is silme written twice, NOT esse"
218
+ },
219
+ {
220
+ "number": 28,
221
+ "name": "hwesta sindarinwa",
222
+ "sound": "ch",
223
+ "note": "general English text and loanwords (research, checker); not used inside Lambelan words"
224
+ }
225
+ ],
226
+ "tehtar": [
227
+ {
228
+ "vowel": "a",
229
+ "value": 1,
230
+ "long_value": 2
231
+ },
232
+ {
233
+ "vowel": "e",
234
+ "value": 2,
235
+ "long_value": 4
236
+ },
237
+ {
238
+ "vowel": "i",
239
+ "value": 3,
240
+ "long_value": 6
241
+ },
242
+ {
243
+ "vowel": "o",
244
+ "value": 4,
245
+ "long_value": 8
246
+ },
247
+ {
248
+ "vowel": "u",
249
+ "value": 5,
250
+ "long_value": 10
251
+ }
252
+ ],
253
+ "sigil": {
254
+ "name": "the 3-sigil",
255
+ "reads": "e",
256
+ "value": 3,
257
+ "rule": "Shelley's ruling 2026-11-20: a full letterform of the house set (not a tehta, not the Latin digit traced); takes the e position in a written word; gematria counts its value 3. Its value coincides with calma's grid number 3 - value is not a unique ID; the sigil's 3 is fixed by the ruling, calma's by grid position."
258
+ },
259
+ "carriers": [
260
+ {
261
+ "name": "short carrier (telco)",
262
+ "rule": "a standalone vowel with no preceding tengwa (word-initial vowels, vowel after vowel); carries no value of its own"
263
+ },
264
+ {
265
+ "name": "long carrier (ara)",
266
+ "rule": "a standalone LONG vowel; the carrier itself marks the length (see the long-vowel convention, spec part (c))"
267
+ }
268
+ ],
269
+ "diphthong_glides": [
270
+ {
271
+ "name": "yanta",
272
+ "sound": "-i glide",
273
+ "rule": "second element of ai/ei/oi diphthongs; carries the i-tehta value 3"
274
+ },
275
+ {
276
+ "name": "ure",
277
+ "sound": "-u glide",
278
+ "rule": "second element of au/eu/ou diphthongs (Laurinde); carries the u-tehta value 5"
279
+ }
280
+ ],
281
+ "digraph_rules": [
282
+ {
283
+ "grapheme": "sh",
284
+ "rule": "silme + aha, two letters, values add (Seashaker, Hakodesh)"
285
+ },
286
+ {
287
+ "grapheme": "ph",
288
+ "rule": "formen alone, value 10 (horsegraph)"
289
+ },
290
+ {
291
+ "grapheme": "gh",
292
+ "rule": "anga + aha, two letters, values add (fighter names)"
293
+ },
294
+ {
295
+ "grapheme": "dh",
296
+ "rule": "ando + aha, two letters, values add (general text only; no house-corpus example)"
297
+ },
298
+ {
299
+ "grapheme": "x",
300
+ "rule": "calma + silme, two letters, values add (PIX3L)"
301
+ }
302
+ ],
303
+ "flagged_not_settled": "The 5 in M3MB3R5HIP (lexicon: '5 for S') is NOT ruled. v1 writes that s with silme (25). A 5-symbol parallel to the 3-sigil is Shelley's amendment to make, not this spec's inference.",
304
+ "long_vowel_convention": {
305
+ "rule": "Length sits on the bearer: one rule, position decides, no per-word choice.",
306
+ "riding": "A long vowel riding a tengwa (medial or final, after a consonant) is written with its tehta DOUBLED on that tengwa. No under-stroke, no separate length sign, no second bearer. Final follows medial.",
307
+ "standalone": "A long vowel standing alone (word-initial or after another vowel) rides the long carrier (ara) with a SINGLE tehta; the carrier's long form is the length mark. A short standalone vowel rides the short carrier (telco), single tehta.",
308
+ "value": "Long value = tehta value x2 in both forms (a 2, e 4, i 6, o 8, u 10); the doubled tehta is one sound and one partial-sequence entry at the doubled value; carriers carry no value.",
309
+ "not_length": [
310
+ "e-diaeresis (ë) is short e, single tehta, value 2; diaeresis is Latin-page information only",
311
+ "acute and circumflex are the same script fact: a-acute = long a; u-acute and u-circumflex both = long u",
312
+ "an unmarked vowel is short; length is never inferred",
313
+ "diphthong glides (yanta -i, ure -u) are not long vowels and are never doubled under this rule",
314
+ "the 3-sigil takes no length: full letterform, fixed value 3, cannot be doubled"
315
+ ],
316
+ "proof": "check_long_vowels.py / long-vowel-check-2026-11-20.txt - 12 diacritic words, 8 long vowels, PASS"
317
+ },
318
+ "general_text_rules": {
319
+ "doc": "mode-spec-general-text-2026-11-20.md (part e)",
320
+ "basis": "G1/G2: general text is transliterated by Latin spelling, not translated and not respelled phonetically; every Latin letter is written, none dropped as silent; longest grapheme unit first, inside a word only.",
321
+ "one_letter_units": {
322
+ "qu": 4,
323
+ "gw": 8,
324
+ "hw": 12,
325
+ "nt": 13,
326
+ "mp": 14,
327
+ "nk": 15,
328
+ "nkw": 16,
329
+ "ng": 19,
330
+ "ngw": 20,
331
+ "th": 9,
332
+ "ch": 28,
333
+ "ph": "formen alone, 10"
334
+ },
335
+ "two_letter_digraphs": {
336
+ "sh": "silme + aha",
337
+ "gh": "anga + aha",
338
+ "dh": "ando + aha",
339
+ "x": "calma + silme",
340
+ "j": "ando + esse (5 + 27) - part e's one addition, no new glyph"
341
+ },
342
+ "single_letter_calls": {
343
+ "c": "calma 3 in all positions (soft c not distinguished in v1)",
344
+ "g": "anga 7 in all positions outside G3 units (soft g not distinguished)",
345
+ "q_alone": "q not followed by u = calma 3 alone",
346
+ "y": "anna 23 all positions",
347
+ "w": "vala 22 all positions",
348
+ "r": "ore 21 all positions",
349
+ "h": "aha 11 all positions outside G3 units",
350
+ "s": "silme 25; ss = silme twice, never esse",
351
+ "z": "esse 27"
352
+ },
353
+ "vowels": "G4/G5: tehta rides preceding tengwa, else short carrier (telco); second i/u after a vowel is a glide (yanta 3 / ure 5); other sequence vowels are their own tehtar; doubled consonant = letter twice, values add (G8).",
354
+ "length": "G6: general-text vowels are short unless the Latin transcription carries a mark of record (a-acute / u-acute / u-circumflex long per part c; e-diaeresis short); length never inferred from English spelling.",
355
+ "page_marks": "G9: no case distinction; apostrophe inside a word drops (no letter, no value); hyphen is a word boundary; spaces, punctuation and ordinary prose numerals are Latin-page marks with no value - distinct from the 3-sigil in 3-for-E house names; the 5 in M3MB3R5HIP remains flagged, writes silme 25.",
356
+ "gematria": "G10: s.6 computation unchanged for prose; one-letter units count once, two-letter digraphs count both, carriers/page marks count 0. Signing sums of record: 'The research and the record' 201; 'The website and the checker on it' 190; 'The website's words' 174; pangram sum 367.",
357
+ "proof": "check_general_text.py / general-text-check-2026-11-20.txt - signing corpus + pangram (a-z) + rule cases PASS, no exception handling, no new glyph"
358
+ }
359
+ }