Module:Auphen/YBS-PJ/doc

This is the documentation for Module:Auphen/YBS-PJ, the sound data for Pjany (registry code YBS-PJ) on the Yezur wiki. It is pure data and defines no functions: Module:Auphen/frame loads it whenever Template:Auphen is called as {{auphen|word|YBS-PJ}} and hands its tables to Module:Auphen, which does the work. The rule notation, and the engine's divergences from it, are documented on Module talk:Auphen.

Fields

Field	Contents
ipacats	The category sets the ruleset works with, written [X]. They hold sounds, not letters, because they are read after the transposition. Pjany supplies ipacats only; Module:Auphen/frame falls back to a cats table where a language declares one for rulesets that operate on spelling.
ipa_rules	The transposition from spelling to sound, applied once before pronounce runs. It is passed to the engine in pronunciation estimation only, so a named ruleset from sets would see the orthography untransposed.
pronounce	The ruleset used for pronunciation estimation.
sets	Named rulesets for grammar and derivation. Empty for Pjany.

Spelling and sound

The word is lower-cased before transposition, and the longest spelling that matches at a given point wins, so digraphs are read ahead of their component letters whatever order they stand in.

Spelling	Sound		Spelling	Sound
a	a		ae	æ:
e	ɛ		ao	ɑ:
i	i		au	ɐʊ
o	ɔ		aú	ɑu
u	ə		éi	eɪ
é	e		ou	ɔə
ó	o		oú	oʊ
ú	u		ui	əɪ
			úi	ɥi

Spelling	Sound		Spelling	Sound
b	b		n	n
d	d		p	p
g	g		q	q
h	h		qh	q͡χ
k	k		r	ɹ
l	l		s	s
m	m		t	t
			z	z

Two spellings carry no sound of their own. An apostrophe ' geminates the consonant before it; a hyphen -, or an en dash –, marks a seam within a compound. Both are carried through the ruleset as digits, 4 and 5, so that rules can see them: the 4 is dropped as soon as the doubling has applied, and the 5 becomes a ~ in the last rules, which Module:Auphen renders as a space. The digits 1, 2 and 3 do the same work for palatalisation and 8 for vowel length; none survive to the output. Length is written : throughout the data and rewritten to ː by the last rule.

An e that stands before another vowel is not read as ɛ where it opens the word or follows a consonant: there the ruleset takes it as a palatal element, described below.

Categories

Members are declared in order, and category-to-category mapping is positional, so the paired sets below line up member by member.

Category	Members	Role
[V]	the vowel inventory, with the backed vowels of [B] appended	Vowel environments.
[C]	the consonant inventory, including the palatals, lenited velars and uvulars that later rules derive	Consonant environments.
[A]	i e ɛ æ: æ a	Front vowels.
[K]	p b k g t d	Plosives.
[R]	r ʀ ɹ	Rhotics.
[Q] → [B]	i ɛ e ə a æ ɐ → ɯ ʌ ɤ o̞ ɑ ɑ ɑ	Backing next to a uvular.
[H] → [J]	s z l n t d p b h k g → ʃ ʒ ʎ ɲ tʃ dʒ ɸ β j x ɣ	Palatalisation.
[W] → [Y]	g k q → ɣ x x	Lenition after a lateral.

[V] holds vowels proper: the offglides ɪ and ʊ and the glide ɥ that the diphthongs bring in are not members, so a rule keyed on [V] does not match them.

The pronunciation ruleset

pronounce applies the following, in this order.

Process	Effect
Seam breathing	A seam standing between two vowels takes an inserted h.
Gemination	A consonant before ' is doubled and the ' dropped. A glottal element is inserted between the two copies late in the ruleset, so t' surfaces as tˀt.
Palatalisation	The palatal element surfaces as j word-initially and as a ʲ diacritic after ɹ or m, with ɹ hardening to r before the diacritic. Before a front vowel a preceding plosive instead takes an inserted j, unless that vowel is itself followed by q or qh. Otherwise the preceding consonant shifts to its palatal counterpart, [H] → [J]. After q the element is dropped and the q held aside as a marker until the backing rule has run, so the following vowel does not back.
Uvular colouring	Beside q or q͡χ, front and central vowels back, [Q] → [B].
Nasal assimilation	n becomes m before a labial, ŋ before a velar and ɴ before q.
Rhotic backing	r and ɹ become ʀ beside q.
Lenition after a lateral	g k q become ɣ x x after l or ʎ.
L-vocalisation	l becomes ʊ before t d n s z and ə before p b m ɸ β, in both cases only when no consonant-plus-vowel follows. A word-final l after a consonant takes a preceding ə.
R-vocalisation	A rhotic becomes ə after a front vowel when a consonant or the word end follows, and after ə when a consonant follows. An ə before a word-final rhotic opens to ɔ.
Final adjustments	hə becomes x except word-finally; ɛə becomes ɛ:; ɥi becomes ɥʲ before a vowel; and two identical vowels merge into one long vowel.

A seam blocks adjacency, so rules that are meant to reach across one list it in their environments explicitly; the rest do not apply over it.

Named rulesets

sets is empty: Pjany has no derivation rulesets yet, and the three-argument form {{auphen|word|YBS-PJ|name}} therefore reports an unknown ruleset and files the page into Category:Auphen errors.

Four worked Pjany estimations are checked on every parse at Template:Auphen/testcases.