Reconstruction talk:Proto-Japonic/esoi

Reason for reconstructing initial /e-/?

@Kwékwlos, curious why you chose initial /e-/ when you first created this entry. What factors make /e-/ more likely than /i-/? ‑‑ Eiríkr Útlendi │^{Tala við mig} 18:40, 31 July 2023 (UTC)

:I suppose /e-/ yields /i-/ in Ryukyuan, while initial /i-/ usually syncopates. Kwékwlos (talk) 19:11, 31 July 2023 (UTC) ::Thank you for the explanation. I'm curious about our other entries for JPX-PRO that currently are shown starting with /i-/ over at Category:Proto-Japonic_lemmas. Walking through those, some show the initial /i-/ dropping out in some of the Ryukyuan languages, and some don't. The sample size is quite small, but given what we have, it appears that elision is most likely when the following consonant is /n/ or /j/. For Proto-Japonic *isonku, where the consonant is an /s/, it looks like the initial /i-/ is maintained.

From what you've read and what sources you have to hand, are our entries reconstructing initial /i-/ perhaps mistaken? Or does initial /i-/ elision in the Ryukyuan reflexes only happen in certain phonological contexts, such as a following /n/ or /j/? ‑‑ Eiríkr Útlendi │^{Tala við mig} 20:25, 31 July 2023 (UTC)

:::I honestly don't know for sure. /i/ elision does happen near a /j/ or a nasal, but is usually preserved before other consonants. On the contrary an /e/ never drops, usually shifting to /i/. Nevertheless, there are well-identified processes. If it palatalizes a following consonant in Okinawan, then /i/ should be reconstructed. It's better to safely reconstruct /i/ for such cases and leave the other cases as undetermined. Kwékwlos (talk) 20:56, 31 July 2023 (UTC) ::::@Eirikr, @Kwékwlos: This is reconstructed based on EOJ osi (cannot be from 磯 iswo). We need the Southern Ryukyuan evidence (from Miyako dialects or something). Commented by Chuterix.

@Eirikr, @Chuck Entz: please let me have access to my own talk page at least. 2600:8803:760B:B200:D3C6:6F56:D654:FC27 22:31, 3 August 2023 (UTC)

@Chuterix, the key problem that precipitated your latest block was your vandalism of another user's pages. Such behavior is entirely unacceptable here. I suspect that Chuck included your talk page in the block due to your unfortunate habit of excessive pings, despite multiple editors asking you on different occasions to stop.

---

Separately, very interesting and odd correspondence between WOJ /i-/ and EOJ /o-/. Reconstructing an initial /e-/ at the Proto level seems strange -- none of our other Proto terms starting with /e-/ seem to have any EOJ reflexes starting with /o-/...??

I don't think it makes any sense to equate EOJ osi with WOJ iso₁. EOJ osi would instead correspond more cleanly to WOJ isi, and EOJ osu to WOJ iso. See also KDJ entries at Sakura-Paris for EOJ おしへ (osipe₁; unattested WOJ equivalent would be isipe₁ → 石辺) as found in MYS 14.3359, EOJ おすひ (osupi₁; unattested WOJ equivalent iso₁pe₁ → 磯辺) as found in MYS 14.3385. (I disagree with the KDJ and Daijisen entries that equate both EOJ terms to WOJ iso₁pe₁.)

That said, I wonder if these two EOJ terms might have different derivations.

• おすひ

MYS 3385 talks about ままのおすひ. Apparently the まま is in reference to 真間, a place in Chiba just over the Edo river from Tokyo. As you can see in the linked Google Maps, however, Mama has no beach, no direct waterfront at all, not on the river nor on Tokyo Bay. This makes any 磯辺 interpretation a bit problematic.

→ I know that the Tokyo coastline has changed a lot over the centuries. Might Mama have been waterfront property some 1300 years ago?

Even so, it's still unclear from the poem that this word おすひ must definitely mean "beach, shore".

• おしへ

MYS 3359 is contextually much more irrefutably talking about a seaside context: おしへにおふる,はまつづら: "the beach [vine? shrub?] growing over the shore (?)". This poem is set in 駿河, I think probably the one here in modern-day Shizuoka prefecture, which is ocean-front and an appropriate location for this poem.

That said, it is not clear that osi here must mean "stone, pebble; rocky beach". Waves are also said to osu, and this osipe₁ might arguably be 押し辺. Moveover, Suruga is at the mouth of the Abe River, and much of the coastline here is sand and silt -- not stone or pebble. I don't think the course of the river has changed much in the historical era, so the coast was probably also sand and silt 1300 years ago.

→ Might this 駿河 mentioned in the poem have been a different location, somewhere that would have had rocky or pebbly shores?

• Phonological issues

I question whether EOJ would alternate between pe₁ and pi₁ to mean this same WOJ 辺 (pe, “place, area, location”). Do we have any other evidence that pe₁ and pi₁ were allophones in EOJ?

Do we have any other likely cases of WOJ i- corresponding to EOJ o-?

I also note that these both appear to be hapaxes, appearing only once each.

@Chuterix, @Kwékwlos, do either of you have any other instances of EOJ osu or osi, where the term is definitively used to mean something similar to WOJ iso or isi ("stone; pebble; rocky beach")?

This paucity of evidence makes it very difficult to use either EOJ osu or EOJ osi here as evidence for much of anything. If these are indeed hapaxes, the possibility exists that these are phantoms arising from copyist errors, or phantoms arising from mistaken interpretations of the meanings. ‑‑ Eiríkr Útlendi │^{Tala við mig} 01:09, 4 August 2023 (UTC)

@Eirikr: Note that if you ask me a question, I will need to block evade again. Would it be fair if I just get blocked from User namespace excluding my own (if possible) and/or talk namespace (excluding personal user talk). I do productive work and don't spam ping in mainspace/reconstruction entries (except the one goofy time I did this in 久留米 (Kurume) where I pinged everyone JA group for ertymology). I will try to be calm. - Chuterix 2600:8803:760B:B200:3FF2:8130:9356:35A1 01:42, 4 August 2023 (UTC)

@Eirikr: Also I'm too lazy to do research on these coincidental meanings. You may want to consult Vovin (2010) if you have the book (Wikimedia editors who put dedicated effort get access to the Wikipedia Library, which includes Brill, which includes Vovin 2010.

Also I found Miyako and Sarahama isu in Jarosz 2015 p. 173. The Okinawan term might be obsolete, as I cannot find it in Okinawa-go Jiten data shu or Sakihara (2006) p. 73, not even in a compound. I'll look into the .xlsx file of the OGJ and see. - Chuterix 2600:8803:760B:B200:3FF2:8130:9356:35A1 01:47, 4 August 2023 (UTC)

Nope, not there in the file. - Chuterix 2600:8803:760B:B200:3FF2:8130:9356:35A1 01:50, 4 August 2023 (UTC)

@Eirikr: According to the KDJ2/Concise edition entry here (spelled variously as 磯, 礒, 石 (磯, 礒, 石), but third spelling unknown), records:

① 石。岩。巖(いわお)。
※万葉（８Ｃ後）一五・三六一九「伊蘇(イソ)の間(ま)ゆたぎつ山川絶えずあらば又もあひ見む秋かたまけて」
② 岩石の多い波打ちぎわ。湖、池、浜、海、川などに用いる。
※古事記（712）中・歌謡「浜つ千鳥浜よは行かず伊蘇(イソ)伝ふ」

Though the "sandy rocks flowing through water streams" is the oldest attestation, it must be a unattested (poetic) semantic shift. The "rock, stone" sense must've became obsolete. Any thoughts Eirikr? 2600:8803:760B:B200:3FF2:8130:9356:35A1 01:58, 4 August 2023 (UTC)

@Eirikr I am still asking for your input. Chuterix (talk) 20:53, 6 November 2023 (UTC)

@Eirikr Chuterix (talk) 19:26, 15 March 2024 (UTC)

I've lost track -- why are you pinging me? What do you want me to do? ‑‑ Eiríkr Útlendi │^{Tala við mig} 23:32, 18 March 2024 (UTC)

@Eirikr: I just wanted to let you know iso in OJ also had a meaning 'rock, stone'. No further action is required. Chuterix (talk) 23:48, 18 March 2024 (UTC)

Cheers then.  :)

Separately, and pondering on this more, one key difference between this pair iso and isi is that this does not behave in the usual way for pairs of vowel-fronting noun forms: iso is used, unfronted, as an independent noun even in phonetic spellings in the Kojiki (the 伊蘇 spelling above). If both iso and isi were from *esoi, we would expect to only find the unfronted-vowel iso form as the first element in a compound (as the 被覆形), and we would expect to only find the fronted-vowel isi form as an independent noun or the second element in a compound (as the 露出形). But that is not the case here: both appear independently, and both appear as the first element in compounds. Other pairs of 被覆形・露出形 noun forms categorically cannot do this. I believe this points to iso and isi not being such a pair, and not being cognate, and instead being separate words with independent (although possibly related) origins.

Suggestively, there is also the term ipa, less ambiguously denoting a larger kind of stone or rock thing. I suspect we might have three etyma: possibly even three compounds, where the initial i- referred to "rock, stone" and the differing final morae may have originally clarified the size. ‑‑ Eiríkr Útlendi │^{Tala við mig} 18:16, 19 March 2024 (UTC)

There is also isu-no2-kami1, a hapax legomenon attested in the Nihon Shoki; equivalent to 石上 (iso1-no2-kami1, not phonetically attested in OJ), if there's more food for thought available. https://kotobank.jp/word/%E7%9F%B3%E4%B8%8A-432791#E7.B2.BE.E9.81.B8.E7.89.88.20.E6.97.A5.E6.9C.AC.E5.9B.BD.E8.AA.9E.E5.A4.A7.E8.BE.9E.E5.85.B8 Chuterix (talk) 19:54, 19 March 2024 (UTC)

The MYS apparently has eight instances of いそのかみ, all spelled ideogrammatically as 石上.

• MYS 9.1768
• MYS 9.1787
• MYS 7.1353
• MYS 6.1019
• MYS 4.664
• MYS 3.442
• MYS 12.2997
• MYS 12.3313

I did find one instance of ideogrammatic 石上 with the kana gloss of いはのうへ, in MYS 7.1373. That seems to be the only case of kanji spelling 石上 with a reading other than いそのかみ. But given that 石上 (Iso no Kami) was apparently the name of a place near Nara where the palaces of two emperors were located, and that this was used as an epithet (枕詞) to imply "old", the limited range of readings for this is perhaps not that surprising.

Do you have any information on where the kana glosses come from, and how old they might be? I'm curious if isu might be an EOJ thing, and also curious about the possibility of scribal errors (such as if the su were spelled in man'yōgana and maybe using an ambiguous glyph, or just munged handwriting, etc.). ‑‑ Eiríkr Útlendi │^{Tala við mig} 22:02, 19 March 2024 (UTC)

There are various manuscripts which either have a same or different kana glosses. Some could be correctly or incorrectly identified grammatically. The Nihon Shoki probably uses more phonetically accurate spelling (from what I heard; see the argument at Reconstruction talk:Proto-Japonic/osirə). The graph may be ambiguous. Although there are tendencies to raise pJ *o > u, and there are some cases where the vowel value /o/ or /u/ cannot be recovered from neither comparative nor internal reconstruction, so there could be a Japanese and Okinawan word /u/ that could ultimately go back to *o; e.g. pJ *mosirə (“woven mat”), where sporadic distribution of mo occurs in Tohoku and the Shimane (Izumo) dialect has mo which has a possibility to go back to pJ *mo, but the data should also be taken at face value, because otherwise we have variants of 'eel' with initial o- distributed in many dialects but we have Shuri nnaji and the OJ form was munagi1. For clarification (extracted from NKD2):

＜なまり＞ ウナキ〔大隅〕ウナッ〔鹿児島方言〕ウナヌ・オナキ゚・オナコ゚〔岩手〕ウナンコ〔仙台方言・福島〕オナキ〔福岡〕オナギ〔茨城・富山県・石川・福井・福井大飯・岐阜・南伊勢・志摩・京言葉・淡路・神戸・播磨・大和・紀州・和歌山県・和歌山・鳥取・岡山・周防大島・徳島・讚岐・愛媛周桑・伊予・瀬戸内・佐賀・長崎・島原方言・豊後〕オナン〔富山県〕オラギ〔豊後〕ムナギ〔鳥取〕ヲナキ゚〔青森〕

One could theoretically speculate there was a pre-pJ *(m)onaNki, but then the Ryukyuan terms might have been loans from Japanese after the initial m- had been lost. The modern Japanese dialects are generally considered to be descendants of central Japanese, with some peculiarities that could theoretically indicate an early layer, but there is no evidence.

For 兎 (usagi, “rabbit”), which is attested in EOJ and possibly in a psuedo-Goguryeo toponym, the following dialectal variants occur:

＜なまり＞ ウサキ゚〔岩手・千葉〕ウサイ・ウサゲ・ウシヤギ・ウシヤゲ・オサキ・オシヤゲ〔飛騨〕ウサキ〔鹿児島方言・大隅〕ウサッ〔鹿児島方言〕ウサニ・オサン〔富山県〕ウシャギ〔富山県・石川〕ウシャゲ〔岐阜〕ウチャ〔千葉〕ウッサギ・ウッサグ〔島原方言〕オサギ〔茨城・越後・新潟頸城・富山県・石川・福井大飯・信州読本・飛騨・志摩・伊賀・京言葉・神戸・淡路・大和・紀州・和歌山県・和歌山・鳥取・岡山・広島県・徳島・讚岐・愛媛周桑・伊予・瀬戸内〕オサキ゚〔青森・津軽語彙・岩手〕オサケ゚・オサ_ンキ゚コ〔津軽語彙〕オサビ〔愛媛周桑〕オッシャギ〔石川〕ヲサギ〔秋田鹿角〕

I mean, the former Aramaki Morozov tried to use the palatalized forms for evidence of "pre-glides" in pre-pJ, as s/he posted at Talk:過ぎる:

兎 (usagi), is described as *osaNki in existing proto-Japonic theory, but some dialectal forms suggest older layer; ウシヤギ /usijagi/ (Hida), ウシヤゲ /usijage/ (Hida), オシヤゲ /osijage/ (Hida), ウッサギ /ussagi/ (Shimabara), ウッサグ /ussagu/ (Shimabara), ウシャギ /usjagi/ (Toyama/Ishikawa), オッシャギ /ossjagi/ (Ishikawa), ウチャ /ucja/ (Chiba) [retrieved from Nihon Kokugo Daijiten].

—Aramaki Morozov

For rainbow, it seems like a possibility, but for 'rabbit' is more speculative. And like I keep complaining, he hasn't edited since September 2023, and he is stuck at 4,188 edits (if anyone sees this value change, please ping me). Chuterix (talk) 22:30, 19 March 2024 (UTC)

Syncopation only happens in *en- (c.f. *enu (“dog”)).

I am sorry for evading the block, but I'm impatient and school starts soon. I do not wish to vandalize any pages. - Chuterix 2600:8803:760B:B200:D3C6:6F56:D654:FC27 22:32, 3 August 2023 (UTC)