Reconstruction talk:Proto-Japonic/sora

Loanword possibility
I'm not really confident on reconstructing this diphthong at the Proto-Japonic level, so I would like to have the Ryukyuan entries be checked on whether they are early loans (especially considering that Okinawan also commonly uses a different loanword for "sky") or not, because if they aren't, the o1 in the Old Japanese word must come from a diphthong due to Proto-Japonic -o- becoming -u- in all word-medial positions. mellohi! (僕の乖離) 18:45, 6 February 2019 (UTC)

Any dice ? 19:19, 6 February 2019 (UTC)


 * I can't find any entries that begin with "ス" in the Miyako Dialect Dictionary: (Okinawa Center of Language Study). Also, there's the issue of orthography. I don't think we should use kanji for Ryukyuan languages that are not Okinawan. Unfortunately, I'm not familiar with sound etymologies so I can't contribute in this area. KevinUp (talk) 03:47, 7 February 2019 (UTC)


 * Miyako and Okinawan seems to have been added by User:Kwékwlos. I think katakana is much more appropriate for most Ryukyuan languages that are not Okinawan. KevinUp (talk) 04:00, 7 February 2019 (UTC)


 * Etymologically, 空・虚 ( → sora) hadn't meant "sky", but "empty space" (cf., , , and dialectal meaning as "feelings, state of mind" whence ). Theoretical contemporary form *sura appears in some dialects as スラ(sura), means "tell a lie, miss estimate".
 * And then in some Ryukyuan equivalents (Kunigami スラー (suraː), Miyako スゥラ (sura)) means "tip of a plant; treetop" derived from the meaning "above".
 * Ideally, I would like to change the page title into *swàrá, because Japanese pitch accents have correspondence with vowel length pattern of Okinawan. If you omit accents from Proto-Japonic, long vowels appear irregularly in the Ryukyuan. --荒巻モロゾフ (talk) 18:00, 20 February 2019 (UTC)

Problematic reconstruction
As discussed some separately with Kwékwlos over at User_talk:Eirikr/2020, this suwara reconstruction is problematic on various levels.


 * No clear basis for reconstructing a, when this is absent from all of the daughter languages.
 * No clear basis for reconstructing a medial, when this is absent from all of the daughter languages.
 * No clear basis for reconstructing three syllables, when all of the daughter languages have only two.
 * No references.

Until we can better nail down the form, I am very uncomfortable adding this to the etymologies of other terms. ‑‑ Eiríkr Útlendi │Tala við mig 18:50, 15 October 2020 (UTC)
 * This entry is totally wrong. 空/虚 (sora) originally doesn't mean "heaven, sky", but "void, outskirt, falsehood". To describe Okinawan form すーら (sūra), with elongation of the first vowel, accent indication should be mandatory. If we cannot get protoform of the accent, we should not make the entry, as should not assume the word has evolved simply from Proto-Japonic ancestor. How should be the name of this entry is *sǒrá (likewise, Proto-Ryukyuan *soraᶜ, with accent class indication).--荒巻モロゾフ (talk) 04:07, 24 December 2022 (UTC)


 * As I learn more about the relationship between Japanese and the Ryukyuan branches of Japonic, it does seem apparent that pitch / accent is an important aspect in understanding how words are connected. I find myself nodding in agreement at Aramaki's post here.  ‑‑ Eiríkr Útlendi │Tala við mig 01:47, 27 December 2022 (UTC)

Should we again move problematic Proto-Japonic/suwara to just Proto-Japonic/sora?
Eiríkr indicated above many problems with reconstruction *suwara. While mellohi! pointed out this fact "Proto-Japonic -o- becoming -u- in all word-medial positions" to propose that "the o1 in the Old Japanese word must come from a diphthong", Pellard (2013:92) wrote "There are several cases of e(1) and o1 in OJ that cannot be explained away as originating in earlier diphthongs, and these thus constitute exceptions to the vowel raising process that affected OJ. Examples of unraised e(1) in OJ include for example pe1ra ‘moldboard, spatula’, ke1pu ‘today’, pe1ta ‘near the shore’, sake1b- ‘shout’, kape1r/s- ‘return’, uke1ra ‘Atractylodes japonica’, ter- ‘shine’, etc. Instances of unraised o1 can be seen in ko1pi2 ‘love’, ko1ga- ‘burn’, to1ma ‘woven rain-cover’, so1ra ‘sky’, to1ra ‘tiger’, yo1wa- ‘weak’, mo1zu ‘shrike’, ko1mo ‘eelgrass’, etc." Erminwin (talk) 17:06, 1 September 2021 (UTC)
 * In the first place, being a original research, this article name shouldn't allow in itself. In my opinion, sòrá might be derived from verb sòr- (/), because from accent and meaning. Derived from a 1-mora verb stem, so it does not shift vowels.--荒巻モロゾフ (talk) 18:00, 3 September 2021 (UTC)
 * soru was not attested until the 14th century, and possibly a (dialectal?) shift from seru, which appeared in the Manyoshu.
 * I should agree this reconstruction should be *sora. Due to shifting being very unusual. Chuterix (talk) 17:24, 22 December 2022 (UTC)
 * A couple points about dating.
 * is attested since 1190. NKD entry.
 * Derived causative / transitive form is attested since the late 900s (further down that page).
 * The spelling is attested since the mid-900s.  NKD entry.
 * The intransitive form is attested since 974 (further down that page).
 * The form is indeed attested in the , in book 14, which is a compilation of poems written in Eastern Old Japanese (EOJ).  EOJ manifests different vowels in specific words, and this looks like it might be one such case.  I don't recall the exact vowel correspondences, but if EOJ -e- is reflected as Western Old Japanese (WOJ) -o- in similar phonological environments, then this EOJ seru could be circumstantial evidence for the existence of a WOJ verb soru.  Then again, the NKD entry notes include the comment that 『「万葉」例は一説に、「競(せ)る」の意ともいう. 』 ("In one theory about the MYS instance, this is intended to be .")
 * So we can date Japanese verb soru to at least the mid-900s. It is certainly possible that the verb existed prior to this and during the Old Japanese stage, given the paucity of ancient texts.
 * My local copy of the NKD includes the etymological note for soru that 『「そ」は「背(そ)」で、後方に曲がる意』 ("The so is from ; meaning to bend backwards"). The noun  is attested already in the  of 712, and there are numerous verbs derived from this, so circumstantially, the existence of a verb soru prior to the mid-900s does not seem too unlikely.  ‑‑ Eiríkr Útlendi │Tala við mig 00:46, 23 December 2022 (UTC)
 * I want to ask, what is the etymology notes section heading title in the physical Daijiten? [語誌]? At least where is it located; the word header before the actual definition?
 * Anyways, there is a problem with this etymology.
 * Sora is so1ra (swora, "sky") in OJP.
 * so- is so2- (so, "back") in OJP.
 * So the phonetics are problematic; there's no real shift from so2 (in PJ as *sə) to so1 (in PJ as *so) (I explained this problem in fukuro etymology, does not connect to -ro2 suffix, -ro element is -ro1).
 * Therefore, *sora might just be a word on it's own and not derived from any other words/languages, and the soru etymology is probably unlikely.
 * (kinda off-topic) I search every word I plan to reconstruct into PJ (in English to find translations) and I find zero (not even close) cognates for most of them. So I feel like PJ is a language that develops on its own, like other Proto-Languages (any resemblances may be Wonderworts or likely accidental). Chuterix (talk) 01:19, 23 December 2022 (UTC)
 * Re: the NKD etym notes for, these are in the headline just before all the definitions. For some reason, the Kotobank data tends to keep all the [語誌], but lose all the headline notes.
 * Re: and, agreed that these are unlikely to be related.
 * Re: relatedness more generally, it is clear from the historical record that there were multiple language families active in the region around what is now Korea. The archaeological record also provides strong evidence that Japonic speakers migrated from the peninsula.  I suspect that the language of Baekje was a cousin of Japanese, with Goguryeo likely a somewhat more distant relative, and Silla more distant still.  The problem with trying to find cognates is that we have only very limited records of these languages -- we must piece together what we can from fragmentary and phonetically incomplete texts transcribed into Middle Chinese.
 * Imagine if the scholars of Proto-Indo-European had only a handful of texts representing a handful of the totality of PIE languages: perhaps only scraps of documents in Old Russian, Old English, Old Farsi, and Old Albanian, with everything else lost to time. That is kind of where we are with Japonic and other areal languages -- even if they are cousins, we are looking at quite a depth of time, and we are missing huge pieces of the puzzle.
 * Personally, I think Japonic and Koreanic are likely to have sprung from the same roots, but with large influxes of other linguistic influences that have replaced substantial portions of the respective lexicons. Chinese is the obvious example for Korean and Japanese both, where Sinic terms have crowded out a lot of native vocabulary.  Without any record of these lost terms, cognates are that much harder to find.
 * Still fun to look, though. :)  ‑‑ Eiríkr Útlendi │Tala við mig 06:04, 23 December 2022 (UTC)
 * @Eirikr Regarding the NKD2 etymological heading before a definition, some of it may still be there. For instance, the NKD2 baka entry at Kotobank (entry here):
 * "〘名〙 (「馬鹿」はあて字. 梵語の moha =慕何（痴）、または mahallaka =摩訶羅（無智）の転で、僧侶が隠語として用いたことによるという)"
 * Meaning:
 * "〘Name〙('馬鹿' is ateji. A shift from Sanskrit moha =boka (foolish), or mahallaka =makara (stupid); the word was used as a slang term by monks)"
 * Does your local copy of NKD contain exactly this etymological information text?
 * Anyways, not much to say for the rest of your information given to me. Chuterix (talk) 13:13, 23 December 2022 (UTC)
 * yes, I typed in the etym note for exactly as written, excluding only the outermost parentheses.
 * Interesting about the 馬鹿 entry -- makes me very curious what the Kotobank devs did. Prior to the latest site upgrade, NKD entries were easier to find (now they are indexed solely by the exact headword string, but in the past, partial matches and matches with okurigana used to work), and NKD entries included more etym information.  I have gradually discovered more and more portions of my local copy that are missing from the Kotobank version, when I'm pretty sure the Kotobank version used to include those same portions.
 * Ah, well. ¯\_(ツ)_/¯  ‑‑ Eiríkr Útlendi │Tala við mig 19:43, 23 December 2022 (UTC)
 * Note: This etymological information text = the baka etymology information, not the soru etymology information. Chuterix (talk) 19:49, 23 December 2022 (UTC)
 * Ah, sorry — for, my NKD says:
 * "（「馬鹿」は当て字. 梵moha＝慕何(痴)、または梵mahallaka＝摩訶羅(無智)の転で、僧侶が隠語として用いたことによる. また、「破家」の転義とも）"
 * This basically matches yours, but for the bit on the end about 『また、「破家」の転義とも』.
 * HTH, ‑‑ Eiríkr Útlendi │Tala við mig 21:56, 23 December 2022 (UTC)
 * HTH, ‑‑ Eiríkr Útlendi │Tala við mig 21:56, 23 December 2022 (UTC)


 * Revisiting the headline question: "Should we again move problematic Proto-Japonic/suwara to just Proto-Japonic/sora?"
 * I would support moving this away from apparently-unjustifiable reconstructed spelling . I am uncertain which spelling would be most appropriate:  or  or  or .  I agree with Aramaki's points about including pitch information where possible, but I don't know how best to do that.  ‑‑ Eiríkr Útlendi │Tala við mig 03:41, 9 January 2023 (UTC)

Possible etymology via surface analysis
OJ so1ra < PJ *sora < Pre-PJ *saUra (reconstructions by me)

A surface analysis suggests that this might possibly be derived via fusion of *sa- (prefixing element?) + *Ura ("interior? hollow (place)?"; also the bound form of OJP ure2 ("growth (of plants); tip") < PJ *Uray). This would merge to o1 in Old Japanese. Chuterix (talk) 04:23, 12 February 2023 (UTC)


 * I think you might have that backwards -- ⟨o₁⟩ is reconstructed by some linguists as /wo/, which could conceivably be a fusion of /u/ + /a/, but not /a/ + /u/. Historically, we have evidence of /a/ + /u/ → /ɔ/, which is closer to the /ə/ that some authors use as the reconstruction of ⟨o₂⟩.  ‑‑ Eiríkr Útlendi │Tala við mig 22:40, 13 February 2023 (UTC)


 * Original research is not good. We shouldn't do "reconstructions by me" or should make it in the extent where have been already solved by decent researchers.😭--荒巻モロゾフ (talk) 02:51, 23 May 2023 (UTC)
 * only known reconstructed OJ Cora is PJ *tora ("tiger") (Vovin 2021); so either PJ *sora, *suara, or *saura, according to the diphthong correspondences in Pellard (2008).
 * PJ could go from *yuawa or *yauwa, not just *yowa. closest compound is  +  but these semantics are strange.
 * The Ryukyu-go onsei database is expected to reopen March 2024.
 * As for OR, what about etymology? Chuterix (talk) 04:02, 23 May 2023 (UTC)
 * We have other dictionaries listing to "place", toko "place", and tokoro "place", which appears to show a clear pattern of successive suffixation. We have other dictionaries listing locative suffix -ko.  We have textual evidence of -ro as a suffixing morpheme that appears in numerous terms, and the extrapolations about -ro at  are extrapolations based on known terms.
 * As OR goes, any such OR regarding is based in confirmed and known background information.
 * The ideas above about sora and yowa, are much more conjectural, and rely upon reconstructions that are neither widely accepted nor widely known, and then from that shaky base, the ideas reach further out into even shakier suppositions. Some of these suppositions lead to clearly mistaken conclusions, as separately discussed over at Reconstruction_talk:Proto-Japonic/ko((r~C)a).
 * Conjecture is great food for thought and conversation. For derivations, speculation should generally be avoided.  If it's the best we have, we should lay out the pieces and indicate how we got there -- surface analysis, known sound shifts, known morphemes, etc. etc.  That said, I do not think we should be creating entries based on speculation. ‑‑ Eiríkr Útlendi │Tala við mig 18:28, 23 May 2023 (UTC)