User:Mellohi!/Reconstructions/Proto-Japonic

Transliteration and lemma spellings

The transliteration for Japonic-language terms used on this wiki is based on modified Hepburn romanization.

However, the Ryukyuan languages north of Okinawan contain aspirated consonants, which for the languages spoken on the Amami Islands are transliterated with <f, θ, x> and in Kunigami <ph, th, kh>. Note that all three of these are aspirated stops except for <f>, which is /ɸ/ in Southern Amami-Ōshima.

Yonaguni has glottalized consonants deriving from reductions of syllables, such as tta "tongue" < *sita. Wiktionary uses <pp, tt, kk> to transcribe these consonants.

Several Japonic languages feature some sort of affricate (generally in a voiceless-voiced pair) arising from independent palatalization events involving their respective reflex of Proto-Japonic *i. These affricates are often transliterated with the letter c. However, on this wiki, they will be transcribed with either -ch- and -j- or -ts- and -dz- based on whether they are or resemble /t͡ɕ/ or /d͡ʑ/ (-ch- and -j-) or /t͡s/ and /d͡z/ (-ts- and -dz-).

Several of the languages also feature the bilabial fricative /ɸ/, which will be transcribed as f. A notable exception is in Miyako, where the sound is pronounced /f/ instead.

Miyako /ss̩/, usually transcribed as sї and deriving from a historical *si, is transcribed on the wiki as su.

Vowels

Proto-Japonic had the monophthongs /a/, /e/, /i/, /o/, /u/, and /ə/.

Diphthongs

Several diphthongs are also known for Proto-Japonic. The falling diphthongs /ai̯/, /ui̯/, /oi̯/, and /əi̯/ can be reconstructed from the existence of non-dipthongal forms in Japanese (e.g. *tukuy yielded modern Japanese tsuki and the combining form tsuku-. In addition, many Eastern Old Japanese dialects tended to truncate diphthongs to the first vowel where Western Old Japanese dialects tended to monophthongize them whole. It is believed that these diphthongs arose with the addition of an unbound nominal suffix *-i. Vovin traces this to an earlier final *-y, *-r (cf. *wasay < *wasar), *-n. Martin traces this to a sequence *-Gi where the consonant was subsequently lost.

In Proto-Ryukyuan, /əi̯/ merged reflexes with /e/, while in Japanese, it merged with /i₂/ > /i/, as seen in *kəy (“tree”).

Vowel fronting

Various Japonic languages fronted /e/ and /o/ to /i/ and /u/ respectively in several contexts in different ways.

Proto-Japonic /o/ becomes /u/ in Japanese except word or morpheme-finally. Due to this fronting, a good chunk of initial and middle /o/ in modern Japanese words come from -o₂ instead, which did not front before the Western Old Japanese stage.

However, word-medial /o₁/ are known to exist in some Japanese terms. The list of Japanese terms with word-medial /o₁/ includes:

• 空 (sora, “sky”)
• 夜 (yoru, “night”), explainable as bi-morphemic (cf. 昼 (hiru, “daytime, noon”))
• 虎 (tora, “tiger”)
• 靄 (moya, “mist”)
• 弱い (yowai, “weak”)

In addition, the attributive godan suffix -u apparently corresponds to EOJ /-o/, contrasting with the shuushikei suffix, which is /-u/ in both WOJ and EOJ, indicating a seemingly word-final fronting of an /o/. This specific fronting in the attributive suffix is of disputed origin, many theories of which propose a contraction of some bimoraic form.

Ryukyuan languages however performed their frontings later than Old Japanese did. For example, the Omoro Saushi (12th-17th century) records many unfronted /o/'s in Old Okinawan, such as in no vs. modern nu for the nominative/genitive marker.

Miyako is very useful in determining which /i/ in Japanese derived from a Proto-Japonic /e/ or an /i/, since the reflexes of those two remained distinct there; *i became ï and e fronted to i. However, ï and i merged when absolutely word-initially in several dialects, Ōgami being excepted.

Another major hint to whether an /i/ in a Japanese term came from /e/ or /i/ may be whether a regressive palatalization of *k or *t is triggered in its Okinawan cognate. *i triggered regressive palatalization while *e did not. This rule of thumb also applies to progressive palatalization, with one caveat - a succeeding original *u blocked progressive palatalization in Okinawan.

Arisaka's laws

Arisaka's laws state that the reflex of /ə/ can seldom co-occur with /o/, /a/, or /u/ within the same morpheme in Old Japanese. This makes the laws a useful tool on occasion, especially when handling syllables that begin with labial consonants (where the -o₁ and -o₂ distinctions are lacking in attestation).

Consonants

Final consonants?

Frellesvig (2021) proposes to reconstruct three pre-proto-Japonic coda consonants based on synchronically irregular morphological relics, namely *-r, *-n, and *-y.

-r

• *nar (“sound”)
  • 音 (ne, “sound”) < *nay < *nar
  • 鳴る (naru, “to ring, resound, echo, roar”) < *nar-
  • 泣く (naku, “to make a sound emotionally”), 鳴く (naku, “to make (an animal) sound”) < *nar-k-
  • 鳴らす (narasu, “to ring”) < *nar-(a)s-

Grammar

Verbs

Final forms

The final predication suffix was *-u for both vowel and consonant-stem verbs. However, Vovin has consistently contested the form of this suffix, preferring to add a nasal coda to account for the nasal coda in the modern Ryukyuan counterparts. Meanwhile, the mainstream theory (as supported by Hattori, Pellard, and Serafim) is that the Ryukyuan final suffixes are auxiliary uses of *wor- (see also Japanese 居る (oru, “to be (animate)”)), which Vovin rejects as failing to explain the nasal coda. The *wor- theory can account for these final nasals in the Northern Ryukyuan conclusive forms however by supposing they arose from the -am- verbalizer suffix being attached to the auxiliary. In addition, a nasal-coda final suffix can be found in several Southern Ryukyuan languages with possible signs (due to non-palatalization) of a lack of the *wor-infinitive construction theorized in the Northern Ryukyuan languages, as found in Irabu kafum, Yaeyama kakun, and Yonaguni kʰagun. Pellard instead explains the non-palatalization by postdating changes like ku → f after a supposed deletion of the infinitive suffix.

Bentley (2008) believes that both the original -u and the -wor- final constructions existed side-by-side in Ryukyuan. Once again, Pellard rejects this conclusion.

Dougherty combines both schools of thought by proposing that the Northern Ryukyuan verb forms were derived from the contraction of an infinitive + auxiliary woru construction, with the -u(m) being suffixed on the woru auxiliary.

Infinitive form

The infinitive-forming suffix is *-i. There is another suffix that derives nouns from verbs that also takes on the form -i, but is considered to be a separate suffix due to accentual mismatch.

Derivational verbalizers

Several verbalizing suffixes are known in Japonic to derive further verbs with extra meaning. Throughout Japonic, the following verbalizers are known (Russell 2006):

• Intransitive *-ar-
• Transitive *-as-
• Adjective/noun verbalizer *-am-
• Transitivity flipper *-ai- < *-aCi?

The initial vowels of these morphemes are often obscured by vowel deletion rules.

These verbalizers can be combined in various ways.

Particles

Several particles are transparently reconstructable for Japonic. These particles include:

• Double-duty nominative/genitive particles *-nka and *nə
• Accusative particle *wo
• Topic particle *pa
• Locative/dative particle *ni

Numerals

Many numerals can be reconstructed for Japonic. From 1-10, they are: *pitə, *puta, *mi, *yə, *itu, *mu, *nana, *ya, *kəkənə, and *təwə. The status of higher numbers in Proto-Japonic is unknown, due to lexical replacement arising from Chinese influence. Old Japanese used amari to form higher numbers, but Japanese has 十三 (とさ, tosa, “thirteen”), suggesting that they were mostly lost in the individual daughter languages.

The numeral pairs of 1 & 2, 3 & 6, and 4 & 8 demonstrate a curious ablaut system where new numerals double the value of the original were created by replacing ə and i with a and u, respectively, and may have been innovated from an earlier non-ablauting system.

They can all be suffixed with the "generic" classifier suffix *-tu. However, *pitə ("1") has an irregular -tu form, having *pitetu instead of *pitətu; this is evidenced by Ryukyuan forms and a single Eastern attestation, as Western Japanese dialects have replaced the /e/ with -o₂ analogically.

Days and humans were generally not counted with *-tu. Instead, *-ka was used to count days, and *-ri or *-tari was used to count people. The numerals *pitə and *puta have forms suffixed in *-ri that have persisted in the modern language as 一人 (hitori, “one person”) and 二人 (futari, “two people”), while other *-(ta)ri forms have fallen into disuse, such as 三人 (mitari, “three people”), 四人 (yotari, “four people”), and 五人 (itori, itsutari, “five people”).

Notes for specific reconstructions

*ika

This term (reconstructed in Proto-Ryukyuan by Kerri L. Russell as *ika in 2009) is in fact two (near-?)homophonous terms, one of them meaning "squid" and the other being an interrogative word. This word's Ryukyuan descendants contain progressive palatalizations triggered by a preceding original /i/, exhibited by Okinawan icha and Miyako ikya.

*pikari

This term demonstrates an -i deverbal derivation from a verb *pikaru. In Okinawan, there are two synonymous words for "light", one of them is fichai, the other one is fikari. The former seems to be the natively inherited term, undergoing typical pre-/i/ intervocalic R-deletion in addition to demonstrating a progressive palatalization. The latter term is likely borrowed from a form of middle Japanese.

*pikaru in turn seems to be derived from a root *pika, with a -ru verbalizer suffix attached. The bare root is still used in modern Japanese as an onomatopoeia for sparkling and glittering, in the form of ぴかぴか (pikapika). The modern Japanese onomatopoeia somehow reverted the fricativization of its /p/, possibly due to its imitative, onomatopoeic nature.

*pəsi

This term has an uncertain first vowel. The o-fronting law and modern Japanese hoshi may predict a schwa as the first vowel, but what's stopping it from being an -i derivation of a lost -su derived verb whose root might have had /o/ instead? That's not even getting to the unfronted medial /o/ words listed on this page. It does not help that written Old Japanese had lost the -o₁ (/o/) and -o₂ (schwa) distinction inconveniently after /p/ by the time it was written down.

*Uta

Vovin has proposed that the initial vowel was /o/ due to analyzing a hapax otapapu as an EOJ conjugation of what is now 歌う (utau, “to sing”).

To-do list

Verbs

• *əmə- (“to think”)
  • OJP omopu > 思う (omou) (reformed with Japanese -ap- suffix, see also utapu < *Uta)
  • Okinawan umuyun, Kunigami umin
  • Miyako umuu - Nevsky & Jarosz don't note a typical lemma /z/ suffix on this verb for some reason.
• *yoko- (“to rest”)
  • Japanese 憩う (ikou) < *yoko-ap-
  • PR *yokori
    • Okinawan yukuyun
    • Miyako yukuu

Adjectives

• *əmə (“heavy”)
  • 重い (omoi)
  • Okinawan qnbusan, Northern Amami Oshima ubusari, Kunigami ubusen.
  • Miyako iv