Classical Tsushiman

Classical Tsushiman is the second stage of the Tsushiman language and the first one with written records. The language was spoken on Tsushima during the Sanqian Period following language contact between Proto-Tsushiman, spoken by indigenous Austronesian peoples who had settled the island prior during the Taoguan Period, and Old Chinese as spoken by Wang Xiang and his retinue as detailed in the Voyage of the Three Thousand.

Evolution
The phonological processes that Classical Tsushiman underwent are a matter of debate. Due to the established nature of Proto-Tsushiman, there is generally a consensus on the Classical Tsushiman reflexes of Proto-Tsushiman's phonemes. On the other hand, Old Chinese has had many reconstructions over the years, and Classical Tsushiman's reflexes have to be constantly updated and changed. As of 2021, the most widely accepted reconstruction is Miklós András Darko's 2016 reconstruction, which will be consistently used here.

From Old Chinese
As mentioned previously, the Classical Tsushiman reflexes of Old Chinese phonemes are not as clear. Darko modified previous theories to fit in with William H. Baxter and Laurent Sagart's 2014 reconstruction of Old Chinese, which significantly differed from that of Zhengzhang Shangfang from 1992.

Darko maintains from previous theories that first, voiceless nasals, as well as *l̥ and *r̥, merged with their voiced forms. Simultaneously, front vowels following pharyngealized consonants moved back.

Following this sound change, *r - which is said to represent the approximant /ɹ/ - merged with /j/.

The most drastic and important sound changes then occured. Pharyngealized consonants with their non-pharyngealized forms (excluding *sˤ, which instead became /ʃ/). These phonemes then gained four reflexes: one in a "normal" environment, one when preceded by a nasal preinitial consonant, one when preceded by a plosive or non-nasal sonorant, and one when preceded by a "loosely attached" or schwa-separated preinitial consonant. In general, he noted the following changes:


 * Aspirated plosives become fricatives regardless of environment.
 * *pʰ became /f/
 * *tʰ became /s/ via *θ
 * *kʰ became /h/ via *x.
 * *kʷʰ became /w/
 * Aspirated affricates (i.e. *tsʰ) spirantized to their fricatives, while *dz always merged with /z/.
 * All plain onset consonants devoiced.
 * All onset consonants with a preinitial nasal voiced. The preinitial nasal then got deleted.
 * Nasal onset consonants with a preinitial plosive or non-nasal sonorant epenthesized with the following vowel. The nasal then either assimilated to the place of articulation of the coda consonant (/m/ for bilabials, /n/ for alveolars including -s, /ŋ/ for velars) then deleted the coda consonant, or remained the same if there was none (*ʔ was considered as "none").
 * Example: Old Chinese 肉 (*k.nuk) "flesh" became Classical Tsushiman 肉 (*kuŋ)
 * Non-nasal onset consonants with a preinitial plosive or non-nasal sonorant remained the same. The preinitial consonant then got deleted.
 * All onset consonants with a loosely-attached preinitial consonant voiced. The preinitial consonant, along with the schwa, then got deleted.

This is consistent with Old Tsushiman phonology, which was heavily influenced by Old Japanese phonology, which had prenasalized voiced plosives.

Phonology
The phonology of Classical Tsushiman is surprisingly well-documented through the copying of the Voyage of the Three Thousand throughout the Sanqian Period. The original text was written in Old Chinese, while later copies were in Classical Tsushiman using rimes to approximate the sound; these approximations are consistent across their estimated date of writing. The issue arises on exactly what sound is being approximated, as mentioned above.