Phonology

Transcription

In these pages I am using a system of transcription which broadly reflects the priestly pronunciation (the most broadly influential dialect) of the late 9th century (Imperial Reckoning). The native syllabary reflects the pronunciation of about a millenium earlier, and would be unduly confusing.

I have toyed with several transcription schemes; here I am using ‘q-notation’. There is also ‘x-notation’ (identical but for the use of <x> where q-notation has <q>) and ‘short notation’ (which has <x>, <q>, <c> for <sq>, <zq>, <nq>). I personally favour the last, but others have objected to the use of <c> for the velar nasal.

Consonants

The consonants are shown below, together with their IPA and X-SAMPA explanations:

transcription

IPA

X-SAMPA

There is, unsurprisingly, considerable dialectal variation, particularly in the point of articulation of the palatal series, which appears frequently as far forward as palato-alveolar (especially in the case of the fricatives).

Vowels

A simple 5-vowel system; IPA and X-SAMPA are the same:

i		u
e		o
	a

There is no constrastive vowel length, nasalisation, or anything else interesting. Diphthongs are purely a matter of /y/ or /w/ appearing in the coda (note I am using q-notation even here). Phonetically, though, vowels do lengthen in open syllables (i.e. those ending in the vowel).

Syllables and Stress

The syllable structure is CVC, with any consonant able to appear in either position.

The stress accent is not predictable. I mark it in transcription with an acute accent unless it falls on the first syllable of the word (which is the case in most nouns). Doubtless certain words (such as common postpositions) lack stress, and so this omission takes away complete clarity of pronunciation, but I want to avoid diactritics in names as much as possible.

Sound Changes

The fairly recent loss of noninitial unaccented vowels (before which the syllable structure was (C)V; current "unaccented" vowels had secondary stress) has led to the juxtaposition of consonants, some of which underwent changes. These changes are active synchronically at this period, and so need mention. (For the purposes of these rules /w/ does not count as a velar, despite its placement in the chart above.)

unvoiced stops become voiced before voiced stops or nasals
voiced oral stops become unvoiced before unvoiced stops and /h/
rh becomes rr
h following a fricative assimilates in place of articulation (but not in voicing) to that fricative
alveolar oral stops beome palatal before palatal oral stops
palatal oral stops become alveolar before alveoar oral stops and velar before velar oral stops
palatal fricatives become velar before velar fricatives, and velar fricatives before palatal before palatal fricatives
n becomes nj before any palatal
nj becomes n before any alveolar and nq before any velar
l becomes lj before any palatal
lj becomes l before an alveolar or velar (more accurately, l and lj are neutralised in a "dark l" before velars)

In tabular form (where the consonant in the left column is the first of the cluster, and showing the results only where there are changes):

	p	b	m	t	d	s	z	n	r	l	tj	dj	sj	zj	nj	ly	y	k	g	sq	zq	nq	h
p		bb	bm		bd			bn				bdj			bnj				bg			bnq
b	pp			pt							ptj							pk					ph
f																							ff
v																							vf
t		db	dm		dd			dn			tjtj	djdj			dnj				dg			dnq
d	tp			tt							tjtj	djdj						tk					th
s																							ss
z																							zs
n											njtj	njdj	njsj	njzj	njnj	njlj	njy
r																							rr
l											ljtj	ljdj	ljsj	ljzj	ljnj	ljlj	ljy
tj		djb	djm	tt	dd			djn				djdj			djnj			kk	gg			djnq
dj	tjp			tt	dd						tjtj							kk	gg				tjh
sj																				sqsq	sqzq		sjsj
zj																				zqsq	zqzq		zjsj
nj				nt	nd	ns	nz	nn	nr	nl								nqk	nqg	nqsq	nqzq	nqnq
lj				lt	ld	ls	lz	ln	lr	ll								lk	lg	lsq	lzq	lnq
k		gb	gm		gd			gn				gdj			gnj				gg			gnq
g	kp			kt							ktj							kk					kh
sq													sjsj	sjzj									sqsq
zq													zjsj	zjzj									zqsq

Note also that homosyllabic /iy/ becomes /i/ and homosyllabic /uw/ becomes /u/.

I discovered an error in my Middle Holic to Early Modern Holic script while writing this; I hope this has not caused any significant problems. Also, in the examples of Holic so far produced it is possible I have forgotten to apply these sound changes in some instances.