Thiorosan: Create the documentation subpage

2026-08-27T12:19:36Z

Create the documentation subpage

New page

This is the documentation for [[Module:Auphen/YAQ-GA]], the sound data for [[Gaillean language|Gaillean]] (registry code <code>YAQ-GA</code>) on the Yezur wiki. It is pure data and is not invoked directly: it returns a table of categories and rules which [[Module:Auphen/frame]] loads and hands to [[Module:Auphen]] whenever [[Template:Auphen]] is called with the code <code>YAQ-GA</code>. The rule notation is the engine's; its behaviour and known divergences are logged at [[Module talk:Auphen]].

== Fields ==

{| class="wikitable"
! Field !! Contents
|-
| <code>ipacats</code> || The sound categories the ruleset scopes over: <code>[C]</code>, the consonants, and <code>[V]</code>, the vowels. Gaillean defines no others. <code>7</code>, the working symbol that stands in for /t͡sʰ/ through the ruleset, is listed among the consonants; /t͡sʰ/ itself is not.
|-
| <code>ipa_rules</code> || The initial transposition, from Gaillean spelling to sound.
|-
| <code>pronounce</code> || The ruleset run by <code><nowiki>{{auphen|word|YAQ-GA}}</nowiki></code>, which carries the transposed form to an estimated pronunciation.
|-
| <code>sets</code> || Named derivation rulesets, keyed by name. Empty for Gaillean.
|}

The page defines no <code>cats</code> field. [[Module:Auphen/frame]] reads <code>cats</code> in preference to <code>ipacats</code> when it runs a named ruleset, and hands the engine no transposition with it, so a named ruleset works on spelling rather than on sound. Gaillean has none.

== Transposition ==

Fifteen digraphs are read as units:

{| class="wikitable"
! Spelling !! Sound !! Spelling !! Sound
|-
| <code>ae</code> || aɪ || <code>mp</code> || ᵐp
|-
| <code>ao</code> || ɶ || <code>oe</code> || oʊ
|-
| <code>bh</code> || b͡β || <code>ou</code> || ɯə
|-
| <code>ei</code> || e: || <code>ph</code> || p͡ɸ
|-
| <code>eu</code> || y || <code>th</code> || c
|-
| <code>gh</code> || g͡ɣ || <code>zh</code> || ʒ
|-
| <code>ll</code> || ʎ || <code>zw</code> || ʃʷ
|-
| <code>mb</code> || ᵐb ||   ||  
|}

The engine matches the longest spelling first, so a digraph wins over its letters wherever it stands in the list; the data page lists the digraphs ahead of the single letters for reading, not for precedence. Where three letters meet, <code>mbh</code> and <code>mph</code> therefore transpose as <code>mb</code> or <code>mp</code> followed by <code>h</code>, and the ruleset later draws them together into /ᵐb͡β/ and /ᵐp͡ɸ/.

Among the single letters the values that do not follow the spelling are <code>a</code> /a:/, <code>c</code> /t͡sʰ/, <code>e</code> /ɛ/, <code>f</code> /ɸ/, <code>h</code> /x/, <code>o</code> /ɒ/, <code>r</code> /ɹ/ and <code>v</code> /β/. The rest transpose to their own value, save <code>x</code>, which becomes the working symbol <code>é</code>, and <code>q</code>, which stands unchanged; both are settled by the ruleset, <code>q</code> as /kʷ/ in <code>qu</code> before a vowel, /k͡x/ at the end of a word and /kʲ/ elsewhere, and <code>é</code> as /kʃ/, or /kəʃ/ finally. <code>~</code> transposes to nothing; because the transposition runs before the engine divides the word, the divider is gone by the time the division is made and has no effect in Gaillean, where a space separates words instead. Length is written with a colon, both here and in the engine's output.

== Pronunciation ==

The vowels alternate with their position in the word. /a:/ lowers to [ä] throughout and returns to [a:] before a consonant and vowel and at the end of a word. /ɛ/ raises to [e] before a consonant and vowel, though not in <code>[C][C]_[C]</code> or <code>[C]_[C][C]</code>, and before a final vowel; it reduces to [ə] between a vowel and a consonant, in <code>[C]_[C][C][V]</code> and <code>[V][C]_[C][C]</code>, after a cluster at the end of a word, and at the end of a word after /ʎ ʒ c g͡ɣ/, though never after <code>qu</code>. /ɒ/ raises to [o] before a consonant and vowel, lengthens to [o:] at the end of a word, and centres to [ʌ] between a vowel and a consonant and in <code>[C][C]_[C]</code> and <code>[C]_[C][C]</code>. The return of [a:] and both raisings are blocked before /ʎ ʒ c/, the sounds of <code>ll</code>, <code>zh</code> and <code>th</code>, so <code>liath</code> is /ljäc/ where <code>dwalia</code> is /dwa:lia:/. <code>ei</code> becomes [ɛi] except at the end of a word, where the transposed [e:] stands, and <code>ie</code> becomes [iə] unless another vowel precedes it.

Among the consonants, /n/ palatalises to [ɲ] before /a: i ɛ/ and /s/ to [ʃ] before another consonant. /i/ becomes the glide [j] beside a vowel, but not at the end of a word, not after <code>qu</code>, and not where a single consonant precedes it and a word-final vowel follows; this is what carries the [i] of [ɛi] to [j], so <code>ceir</code> is /t͡sʰɛjɹ/. At the end of a word, and before a word-final consonant, /b d g/ harden to [p t k] and /z ʒ ʒʷ/ to [s ʃ ʃʷ]; /β/ becomes [w] finally, except in <code>bh</code>, and [ɸə] before a final consonant.

Several sounds take a supporting schwa at the end of a word: /ʎ/ becomes [lə], /m/ becomes [ɱə], <code>é</code> becomes [kəʃ], and a word consisting of <code>h</code> alone is [xə]. A final /ɛɹ/ after /t p k/ reduces to [əɹ], so <code>awiter</code> is /a:witəɹ/.

Two rules use <code>%</code>, which in an environment stands for the target as matched: <code>[C]//_%</code> drops the first of a doubled consonant, so <code>arreie</code> is /äɹɛjɛ/ against /a:ɹɛjɛ/ for a single <code>r</code>, and <code>[V]/:/%_</code> writes the second of a doubled vowel as length, so <code>duu</code> is /du:/. Late rules then collapse [nɲ] to [ɲ] and [ʃs] to [s] and restore the working symbols <code>7</code> and <code>8</code> to /t͡sʰ/ and /iə/; while those two stand, they belong to neither category and so match neither <code>[C]</code> nor <code>[V]</code>, as <code>é</code> does not. Two further rules act on characters the transposition never produces: <code>1</code> becomes a hyphen and <code>2</code> is deleted, and both fire only where the character is given in the word itself.

== Named rulesets ==

<code>sets</code> is empty, so <code><nowiki>{{auphen|word|YAQ-GA|ruleset}}</nowiki></code> has nothing to call: it reports an unknown ruleset and files the page in [[:Category:Auphen errors]]. Gaillean noun declension is not derived by the engine — [[Template:YAQ-GA-decl]] builds it from the paradigm and irregular-stem tables in [[Module:Gaillean]]. [[Template:Auphen/testcases]] holds the engine's live acceptance tests; Gaillean has no rows there.

Module:Auphen/YAQ-GA/doc - Revision history

Thiorosan: Create the documentation subpage