X is weird

[Ketsuban]

Rounin Ryuuji is referring to a speculative sound change where the already weak and often unvoiced vowel in the syllables クシスチツ is dropped entirely. Old Japanese had no syllable-final consonants; Francis-Ratte's reconstructed ancestor of Korean and Japonic includes only /m n ŋ r/ syllable-finally, and all are eliminated by sound changes (leading to the prenasalised consonants and vowel inventory of Old Japanese).

[Travis B.]

Rounin Ryuuji is referring to a speculative sound change where the already weak and often unvoiced vowel in the syllables クシスチツ is dropped entirely.

I learned Japanese, which I have now mostly forgotten, incorrectly because I didn't hear these voiceless vowels - rather I thought, say, tabemasu was /tabemas/ and tabemashita was /tabemaʃta/.

[dɮ the phoneme]

Rounin Ryuuji is referring to a speculative sound change where the already weak and often unvoiced vowel in the syllables クシスチツ is dropped entirely.

I learned Japanese, which I have now mostly forgotten, incorrectly because I didn't hear these voiceless vowels - rather I thought, say, tabemasu was /tabemas/ and tabemashita was /tabemaʃta/.

For some speakers they genuinely are [tabemas] and [tabemaɕta], and honestly (especially in light of the introduction of monstrous numbers of English loanwords), those "devoiced" vowels are not going to stick around underlying for long.

[Kuchigakatai]

Though Japanese didn't originally, at least with Hiragana — ん and む descend from Man'yōgana with the "mu" reading, apparently 武 and 无 respectively — though Katakana ン seems to descend from 尓 ("ni", Middle Chinese */ɲie/) — but of course, the development of the final -n characters is in keeping with the evolution of the language; Katakana seem to have been developed more deliberately, though, and possibly after the phonemicisation of terminal -n (I can't get a good timeline on this, unfortunately; perhaps somebody might fill in the gap here?),

Bjarke Frellesvig in his A History of the Japanese Language (2010) says the Old Japanese syncope of medial unchecked vowels ("音便 onbin") after a nasal that produced coda -n, e.g. yomi1te > 読んで yonde, was "an established part of the language by the time of the appearance of the first mainstream EMJ written sources at the beginning of the tenth century", appearing as they do in Buddhist glosses and in discussions of the language, but "onbin forms are underrepresented in the literary sources and remained so through the entire MJ period. In particular, onbin forms are virtually never found in poetry..." (page 194). He later mentions that a somewhat standardized etymological spelling took hold in the 13th century, which was only abandoned in the 1946 orthographic reform.

Regarding the origin of ん and ン:

The hiragana letter ん for the nasal moraic obstruent /N/ is thought to derive from the man’yōgana 无 (mu), though some scholars believe it to be a further development from (a precursor of) katakana ム (itself < 牟). Table 6.2 above shows two letter shapes for /N/, one like present-day hiragana ん and the other seemingly a precursor of katakana ン. However the katakana letter ン seems to be first attested as such in the eleventh century. It appears to have no recognizable resource in a man’yōgana, but to have been invented specifically for the purpose of representing /N/. The dot in ン may well reflect the nasality diacritic 、which was also a source of the dakuten (6.1.2.2).

The "table 6.2 above" refers to an inventory he quotes, from another scholar (TSUKISHIMA Hiroshi), of early kana taken from annotations on a Buddhist text dating from 883, 地藏十輪經 Dìzàng Shílún Jīng / Jizō-jūrin-gyō. It shows a number of variants per syllable, including shapes that we now associate with katakana or hiragana, in a section that says both developed together from and alongside man'yōgana.



Note that when Frellesvig says ン "seems to be first attested as such" (i.e. representing coda -n) in the 11th c., he's basically contradicting the scholar who made the table he quoted, so that Tsukishima's ン shapes would just be variants of 牟~ム.

Wait, what?

Middle Chinese words ending in -t were borrowed with a coda -t in Middle Japanese. This is known from Christian missionaries' transcriptions, which show e.g. "fotnet" for today's 発熱 hatsunetsu < 發 /pʉɐt/ + 熱 /ɲiɛt/, or "connit" for for today's 今日 konnichi, and which can be glanced at in surviving gemination as in 末端 mattan 'tip, extremity of sth' < 末 /muɑt/ + 端 /tuɑn^平/. Since kana at the time had no way of writing this down, they tended to be written with a default vowel, usually -つ -tsu. The orthographic reform of 1946 introduced the practice of small kana to represent various new sounds absent in the traditional inventory defined in the Iroha poem, and so today's sokuon っ or small tsu came into existence (Frellesvig, page 169).

Basically, it's possible that in the future the new coda -s -sh -t etc. will be written with small versions of す し と etc.

In NJ [New Japanese] the words which in late LMJ [Late Middle Japanese] had final /-t/ have an epenthetic vowel, usually /u/ (netsu ‘fever’), but sometimes /i/ (konnichi ‘today’). In earlier sources in Japanese script, these words are written with kana for tu (つ) or ti (ち), but it has recently been shown that a distinction also was made in some LMJ kana sources between -tu and -t, by using variant kana (hentaigana, see EMJ 6.1.2) which were originally used as equivalents for /tu/.

[...] Final /-t/ is almost entirely limited to SJ [Sino-Japanese] vocabulary, but it is also found in a very small number of variant shapes of native words in the Christian sources, whose regular shapes, however, have final vowels: ximot ‘cane, rod’ (used in criminal punishment)’ (~ /simoto/), carafit ‘big box; lit. Chinese chest’ (~ /karafitu/); tetdai ‘help’ (~ /tetudai/). Spelling variants by hentaigana in sources in Japanese script confirm that these words had final /-t/.

It may be worth mentioning many medial geminate consonants today are native, coming from syncope that eliminated vowels in Old Japanese, e.g. womi1na > 女 onna 'woman', torite > 取って totte 'take (gerund)'...

[Rounin Ryuuji]

Though Japanese didn't originally, at least with Hiragana — ん and む descend from Man'yōgana with the "mu" reading, apparently 武 and 无 respectively — though Katakana ン seems to descend from 尓 ("ni", Middle Chinese */ɲie/) — but of course, the development of the final -n characters is in keeping with the evolution of the language; Katakana seem to have been developed more deliberately, though, and possibly after the phonemicisation of terminal -n (I can't get a good timeline on this, unfortunately; perhaps somebody might fill in the gap here?),

Bjarke Frellesvig in his A History of the Japanese Language (2010) says the Old Japanese syncope of medial unchecked vowels ("音便 onbin") after a nasal that produced coda -n, e.g. yomi1te > 読んで yonde, was "an established part of the language by the time of the appearance of the first mainstream EMJ written sources at the beginning of the tenth century", appearing as they do in Buddhist glosses and in discussions of the language, but "onbin forms are underrepresented in the literary sources and remained so through the entire MJ period. In particular, onbin forms are virtually never found in poetry..." (page 194). He later mentions that a somewhat standardized etymological spelling took hold in the 13th century, which was only abandoned in the 1946 orthographic reform.

Perhaps my knowledge on this subject is further out of date than I'd thought.

Regarding the origin of ん and ン:

The hiragana letter ん for the nasal moraic obstruent /N/ is thought to derive from the man’yōgana 无 (mu), though some scholars believe it to be a further development from (a precursor of) katakana ム (itself < 牟). Table 6.2 above shows two letter shapes for /N/, one like present-day hiragana ん and the other seemingly a precursor of katakana ン. However the katakana letter ン seems to be first attested as such in the eleventh century. It appears to have no recognizable resource in a man’yōgana, but to have been invented specifically for the purpose of representing /N/. The dot in ン may well reflect the nasality diacritic 、which was also a source of the dakuten (6.1.2.2).

The "table 6.2 above" refers to an inventory he quotes, from another scholar (TSUKISHIMA Hiroshi), of early kana taken from annotations on a Buddhist text dating from 883, 地藏十輪經 Dìzàng Shílún Jīng / Jizō-jūrin-gyō. It shows a number of variants per syllable, including shapes that we now associate with katakana or hiragana, in a section that says both developed together from and alongside man'yōgana.



Note that when Frellesvig says ン "seems to be first attested as such" (i.e. representing coda -n) in the 11th c., he's basically contradicting the scholar who made the table he quoted, so that Tsukishima's ン shapes would just be variants of 牟~ム.

Wait, what?

Middle Chinese words ending in -t were borrowed with a coda -t in Middle Japanese. This is known from Christian missionaries' transcriptions, which show e.g. "fotnet" for today's 発熱 hatsunetsu < 發 /pʉɐt/ + 熱 /ɲiɛt/, or "connit" for for today's 今日 konnichi, and which can be glanced at in surviving gemination as in 末端 mattan 'tip, extremity of sth' < 末 /muɑt/ + 端 /tuɑn^平/. Since kana at the time had no way of writing this down, they tended to be written with a default vowel, usually -つ -tsu. The orthographic reform of 1946 introduced the practice of small kana to represent various new sounds absent in the traditional inventory defined in the Iroha poem, and so today's sokuon っ or small tsu came into existence (Frellesvig, page 169).

I had never heard that, either.

Basically, it's possible that in the future the new coda -s -sh -t etc. will be written with small versions of す し と etc.

I suppose that is most likely the case, though I think it would be a shame not to revive some of those perfectly good Hentaigana, and the archaic ゐ and ゑ, too.

In NJ [New Japanese] the words which in late LMJ [Late Middle Japanese] had final /-t/ have an epenthetic vowel, usually /u/ (netsu ‘fever’), but sometimes /i/ (konnichi ‘today’). In earlier sources in Japanese script, these words are written with kana for tu (つ) or ti (ち), but it has recently been shown that a distinction also was made in some LMJ kana sources between -tu and -t, by using variant kana (hentaigana, see EMJ 6.1.2) which were originally used as equivalents for /tu/.

That is quite a departure from how I understood those to have developed. The earlier (?) idea seemed to be that -u was the original epinthetic vowel, but that -i replaced it by analogy with unbound forms often ending in -i, or else had that ancient nominaliser -i appended to them.

[...] Final /-t/ is almost entirely limited to SJ [Sino-Japanese] vocabulary, but it is also found in a very small number of variant shapes of native words in the Christian sources, whose regular shapes, however, have final vowels: ximot ‘cane, rod’ (used in criminal punishment)’ (~ /simoto/), carafit ‘big box; lit. Chinese chest’ (~ /karafitu/); tetdai ‘help’ (~ /tetudai/). Spelling variants by hentaigana in sources in Japanese script confirm that these words had final /-t/.

It may be worth mentioning many medial geminate consonants today are native, coming from syncope that eliminated vowels in Old Japanese, e.g. womi1na > 女 onna 'woman', torite > 取って totte 'take (gerund)'...

Did he mention the variety of Japanese whence those samples were taken? I wonder if they weren't from something ancestral to Miyako.

[Nortaneous]

Did he mention the variety of Japanese whence those samples were taken? I wonder if they weren't from something ancestral to Miyako.

Seems unlikely to me - not only for geographic reasons, but also because the Japanese-Ryukyuan split is (IIRC) totally basal within Japonic, meaning that there never was a variety of Japanese ancestral to Miyako.

[Rounin Ryuuji]

Did he mention the variety of Japanese whence those samples were taken? I wonder if they weren't from something ancestral to Miyako.

Seems unlikely to me - not only for geographic reasons, but also because the Japanese-Ryukyuan split is (IIRC) totally basal within Japonic, meaning that there never was a variety of Japanese ancestral to Miyako.

I doubt the Europeans of that era would've been aware of it, and probably would've called everything from the area "Japanese"; my wording was clumsy, however, I admit.

[Kuchigakatai]

Did he mention the variety of Japanese whence those samples were taken? I wonder if they weren't from something ancestral to Miyako.

Unfortunately, he doesn't say what Christian texts fotnet and connit come from. Since they're from Christian texts, they have to be from the 17th century or later though.

But while we're at it, in page 300 you can find this fun sample from João Rodrigues's Vocabolario da lingoa de Iapam (ca. 1604):

Von arujino notamauaqu: QUI SEQITUR ME, NON AMBULAT IN TENEBRIS, SED HABEBIT LUMEN VITÆ. Ioan. 8. Vareuo xitŏ monoua yamigiuo yucazu: tada jumiŏ no ficariuo motçu bexi to nari.
'The Lord says: "Whoever follows me will never walk in darkness but will have the light of life."'

Or, in Frellevig's adaptation of modern romanization (the kanji/kana in parentheses is mine):
on-aruzi(御主) no(の) notamawaku(曰わく):
ware(我) o(を) sitɔɔ(從う) mono(者) wa(は) yamidi(闇路) o(を) yukazu(行かず,不行):
tada(但) zyumyɔɔ(寿命) no(の) fikari(光) o(を) motu(持つ) besi(可し) to(と) nari(也)

About von -> on, he says word-initial /o oo ɔɔ/ had a prothetic initial *[w-], hence also touo for */to.o/ *[towo] 'ten' (十), vô for */oo/ *[woo] (大), vŏ for */ɔɔ/ *[wɔɔ] 'king, emperor' (王, p. 324). Besides many fun words with f- like mŏfaya /mɔɔɸaja/ 'already' (i.e. 最早 mohaya) (p. 312).

[Vijay]

Sampson says that Greek had letters χ ψ because those were the only clusters that could end a syllable.

By χ, did you mean ξ?

[zompist]

Sampson says that Greek had letters χ ψ because those were the only clusters that could end a syllable.

By χ, did you mean ξ?

Yep, skewered by a schoolboy snafu.

[Travis B.]

Sampson says that Greek had letters χ ψ because those were the only clusters that could end a syllable.

By χ, did you mean ξ?

Yep, skewered by a schoolboy snafu.

Mind you, χ was /ks/ in the Western Greek alphabet, hence why x is /ks/ in a variety of languages using the Latin alphabet.

[Vijay]

Thanks! I'd never even heard of the Western Greek alphabet.