Wiktionary:About Proto-Japonic

Proto-Japonic is the language of the ancestral Japonic language family, which consists of Japanese, Ryukyuan, Hachijō, and possibly. As Proto-Japonic is a reconstructed language, all entries must be in the Reconstruction mainspace.

Abbreviations

 * PJ - Proto-Japonic
 * PR - Proto-Ryukyuan
 * OJ - Old Japanese
 * EMJ - Early Middle Japanese
 * MDJ - (Modern) Japanese

Phylogeny
The descendants of Proto-Japonic include:

Yonaguni also has some shared innovations with Yaeyama, leading Pellard to believe that Yonaguni should be grouped into Yaeyama (Macro-Yaeyama), while the actual Yaeyama language family is called Nuclear Yaeyama. Thorpe believes Yonaguni should be considered a third branch of Ryukyuan, descending from an early dialect in Okinawa. He also posits Taketomi as an early South Okinawan dialect that was reshaped by neighboring Yaeyaman varieties, and that the Yuwan dialect in Amami Ōshima is derived from a back-migration from Yoron Island.

A reconstruction can be reliable if it's attested abundantly in the Japanese (particularly Old Japanese, and even Eastern Old Japanese) and Ryukyuan branches (e.g. PJ or PJ ).

It's best to put sources for the descendants (e.g. PR ). Either directly use the references manually, or you can use in an entry (has pre-made reference templates).

Reconstructions
See Category:Proto-Japonic lemmas.

Transcription and phonology
The transcription uses modified.

Vowels
Six vowels can be reconstructed for Proto-Japonic:

Diphthongs
Diphthongs that can be reconstructed for Proto-Japonic include *ui, *oi, *əi, *ai, *ia, *iə, ua, and au.

Standalone forms and combining forms
Some Japanese nouns have two stems; an atomic root form, and one in which the final vowel is fronted. The second form appears when the word is used as a standalone noun or as the latter element in a compound. Words showing this alternation include ~  (< PJ ),, , etc. In Japanese linguistics, the form of these nouns ending with the front vowels is called the, and the form with the back or open vowels  is called the.

In Proto-Japonic, this vowel-fronting element is reconstructed by some as *-i, as seen in Hattori and Pellard's works. Vovin also reconstructs *-y, although this appears to be based on loanword evidence (compare  or ), and the prohibition of consonant clusters. The final *-y is deleted in ancient compounds (e.g. and ) and in Eastern Old Japanese (*Oy; compare  (< PJ ) and  (< PJ  via vowel raising)), while monophthongization occurs as a standalone noun.

Alexander Vovin traces this *-y element to an earlier *-r (see PJ ), but this appears to be based solely on. Other word-final consonants have been suggested (such as, based on the 1.2 and 2.5 accent patterns of Heian-period Kyoto), but these are speculative. The Hateruma/Shiraho and Yonaguni final nasals are certainly parallel innovations and are not indicative of an archaism.

Vowel correspondences
Within Old Japanese and Proto-Ryukyuan:

Unknown vowels
If a word cannot be determined an unraised vowel, or a type of vowel is unknown, these cover symbols are used:
 * Unknown back vowel */o/ or */u/: *U
 * Unknown front vowel */e/ or */i/: *I
 * Unknown mid vowels */o/ or */ə/: *O

Consonants
Proto-Japonic has the reconstructed consonant inventory of *, *, *, *, *, *, *, *, *, *, *, *, and * (or y in non-IPA notation). Kupchik (2023) also reconstructs *.

Voiced consonants in Old Japanese and Proto-Ryukyuan are thought to be derived from the pre-nasalized consonant + consonant clusters *, *, *, *.

Reconstructing approximants
It was a custom for "Altaicists" and advocates of a "Koreo-Japonic" hypothesis to reconstruct voiced consonants in proto-Japonic (or proto-Japanese), mainly to make "comparisons" with other languages look better. Even if one claims these consonants based on internal reconstruction, it is mostly based on typology, especially in the case of *b.

However, both of these voiced consonants are only confined to Sakishima and no other dialect in Japonic; see below.

*w
In Sakishima (Southern Ryukyuan dialects, including Miyako, Yaeyama, and Yonaguni), there have been mostly consistent correspondences of mainland Japanese *w- to initial *b-.

Words in Sakishima dialects that have inital w- are either secondary developments (e.g. pR > Irabu-Nakachi ), or recent loanwords (e.g. Irabu-Nakachi, , , ; they are explicitly marked as "new (新)").

Some Northern Ryukyuan dialects have a change *w- > gw- (e.g. Benoki), *wu- > gu- (e.g. Aden in Kikai), or *w- > f- (e.g. Yoron Higashi-ku), which has no explanation should *b- be reconstructed. Even if it's somehow possible, it is certainly less natural than a change w- > gw-.

There are no correspondences of Old Japanese u- to any Sakishima dialect bu-, as wu is basically absent from Japanese altogether. To account for this missing correspondence, it is more plausible to reconstruct *w-.

Sino-Japanese loanwords have b- in them (e.g. Irabu-Nakachi, Yonaguni , ), but then, it may have simply been a process of adjusting all initial w- in the loanwords to b-.

Compounds that involve *-w- often get deleted, e.g. "couple" has Irabu-Nakachi, Minna , Yonaguni ( + ); also Irabu-Nakachi  <  +. There is no explanation for the Minna form to be derived from *meboto(ra). However, one can also argue that the medial *-w- had lenited.

One evidence that Whitman and Vovin add to support *b- internally is using (Hirara) Miyako data that has v(u)-, which can be observed in words that have *-bu-, e.g. Miyako. However, Pellard (2023, 09) gives a note that some data from Miyako, including the Hirara dialect, could have been potentially been respelled to make it look closer to its Japanese corresponding form.

*y
It is only Yonaguni that has most of Japanese and Ryukyuan y- correspond to d-, e.g., , ,.

Typologically, Yonaguni has undergone many other fortitions: e.g. *z- > d-, *ch- > t.

There are no correspondences of Old Japanese i- to Yonaguni di-. yi- can only be reconstructed from morphological alternations, e.g. . To account for this missing correspondence, it is more plausible to reconstruct *y-.

Many compounds have a y that have been formed from compounds: e.g. compounds with "house" include:, , , , , , , ; one compound with "night" is.

Sino-Japanese loanwords also have d- for y-: e.g., , but it is possible that Yonaguni adjusted y- to d- in the loanwords.

The most decisive evidence in favor of y- is that the placename Yonaguni is recorded in a Korean record as 閏伊是麼 (zyun.i.si.ma), which implies a fortition *y- > *z- > d-. The modern placename is.

Pitch accent
The following accent patterns can be reconstructed for Proto-Japonic (an abbreviation indicates dubiousness, but after an accent, implies the accent in a particle if the accent is different):

proposes that the falling pitch of 2.5 nouns were derived from a final *-m, based on both internal and external evidence (see also ). For internal examples, has a verb, and  has a verb. For external instances, see and. He also assumes the low register goes back to an initial voiced consonant, claiming that this is how the low register corresponding to Proto-Ryukyuan tone class B has developed from such. However, there is no internal evidence for this.

It has been argued that a split in classes B and C of class 2.4 and 2.5 nouns were caused by either of the following: It is thus that most linguists, such as Thomas Pellard and Yosuke Igarashi, propose that the split goes back to Proto-Japonic.
 * The split was caused by a condition in the vowel.
 * There are conflicting correspondences.
 * That class C nouns were derived from vowel length, as suggested by Hattori.
 * However, many Ryukyuan dialects often have accent on the penultimate mora, and even in some heavy syllables (e.g. "today" (*keu), "body; self" (*dou")), there is a penultimate accent. In addition, words that actually have etymologically long vowels are typically preserved in some compounds, while long vowels from class C are typically shortened in compounds.
 * That the split is due to dialectal contact, as suggested by de Boer (2010).

Key

 * H - High pitch.
 * L - Low pitch.
 * R - Rising pitch.
 * F - Falling pitch.

Unraised medial vowels
Normally, Proto-Japonic /e/ and /o/ raises to /i/ and /o/ in attested daughter languages. An example is and, or  and.

There are word medial e1 and o1 in Old Japanese that cannot be explained away as originating from diphthongs. Examples include:


 * e1
 * (< PJ ; no Ryukyuan cognates, which is unsurprising as the plant is not native there)
 * o1
 * (< PJ )
 * (< PJ )
 * (< PJ )
 * (< PJ )
 * o1
 * (< PJ )
 * (< PJ )
 * (< PJ )
 * (< PJ )
 * (< PJ )
 * (< PJ )

These should be reconstructed with a medial *-e- and medial *-o- in Proto-Japonic, unless excepted (e.g. internal etymologies such as PJ, from + ).

Speculatively, most of the unraised *o seems to be adjacent to *a, suggesting that the adjacency blocked vowel raising in Mainland Japanese.

Verbs
For the time being, verbs should be reconstructed as root + terminal suffix. There has been arguments that the terminal suffix must have contracted an ancient root form, but there is no proper consensus about this on Wiktionary.

In the Ryukyuan languages, some verbs appear to derive from the + *worV-, an auxiliary suffix cognate with.

An example:


 * Proto-Japonic
 * Proto-Ryukyuan
 * Kikai
 * Okinawan

Adjectives
Use only the adjective root, not a suffix added to it. For example, descends from, without a **-sV suffix. Exceptions are : e.g.,.