188.8.131.52:59: YELLOW PIG 12/17
songgiyan uliya aniya
juwa juwe biya darhon inenggi
'yellow pig year, ten two month, seventeen day'
Out of time tonight, so I'll just make a few remarks about Jurchen
'Seventeen' in Jurchen dates is either darhon 'seventeen' (in Jurchen Empire usage) or juwa nadan 'ten seven' (in Ming dynasty usage). darhon 'seventeen' is obviously not cognate to juwa 'ten' or nadan 'seven' and is probably a loan from a para-Mongolic language (a nonstandard Khitan dialect?).
Khitan as preserved in written records
has <TEN SEVEN> in both scripts rather than a special word for
'seventeen'. The Khitan word for 'seven' may have been something like dalo
(Kane 2009: 115) with an l like Written Mongolian dologhan.
The Jurchen word for 'seventeen' may actually be dalhon
if the in the Ming Mandarin transcription of Jurchen 'seventeen'
represents -l- rather than -r-. (Ming Mandarin had no
means to precisely represent a consonant cluster -lh-, and
there was no reason for Jurchen speakers to borrow foreign *l
as r.) dalhon has no Manchu cognate, so I cannot use Manchu to
decide on whether represents -l- or -r-.
Janhunen (2003: 399) reconstructs the potential source of Jurchen darhon as pre-Proto-Mongolic *dal.ku/n or or *dal.u.ku/n 'seven-teen'.
184.108.40.206:59: YELLOW PIG 12/16
songgiyan uliya aniya
juwa juwe biya nilhun inenggi
'yellow pig year, ten two month, sixteen day'
1. 'Sixteen' in Jurchen dates is either nilhun 'sixteen' (in Jurchen Empire usage) or juwa ninggu 'ten six' (in Ming dynasty usage). nilhun 'sixteen' is obviously not cognate to juwa 'ten' or ninggu 'six' and is probably a loan from a para-Mongolic language (a nonstandard Khitan dialect?).
Khitan as preserved in written records has <TEN SIX> in both scripts rather than a special word for 'sixteen'. The Khitan word for 'six' may have been something like nil- (nilhun minus -hun 'teen'), but there is no known Khitan-internal evidence for such a reading: e.g., Khitan small script <SIX> alternating with a graph sequence <n.il>.
The nil- of nilhun has no cognate in Mongolic which has an innovation *jir-gu.xa/n 'two (times) three' (as proposed by Janhunen [2003: 399]). nil may be from the Proto-Serbi-Mongolic word for 'six' (also Janhunen's idea, though he uses the term 'Pre-Proto-Mongolic').
The Manchu cognate of nilhun is niolhun
'sixteenth day of the first month' (not 'sixteen'). Did Jurchen and
Manchu borrow the word from different sources (or the same source in
different periods)? Did Jurchen reduce an original *niol [ɲɔl]
to nil? Or is Manchu -o- an innovation in the Manchu
line due to the influence of the following -u-?
Janhunen (2003: 399) reconstructs the potential source of Jurchen nilhun
as pre-Proto-Mongolic *nil.kü/n
2. Today I learned the Serbo-Croatian word свађа svađa
derives it from Proto-Slavic *sŭvadja. What are its cognates in
other Slavic languages? I can't find anything like сважа in Russian.
3. Tonight I realized that Sino-Japanese
have 'irregular' readings by comparison with other Go-on and Kan-on: e.g.,
But I suspect the readings suu and guu are actually regular and from a layer of Go-on predating vowel raising: i.e., they were borrowed as *-ou which became -uu after raising.
4. Tonight I realized that 源氏物語 Genji
was written in roughly the same period (before 1021) as the creation of
the Tangut script in 1036. I regret that so little Tangut
original literature has survived. Imagine what we could learn about
Tangut culture from a Tangut equivalent of Genji.
220.127.116.11:59: YELLOW PIG 12/15
songgiyan uliya aniya
juwa juwe biya tobohon inenggi
'yellow pig year, ten two month, fifteen day'
1. 'Fifteen' in Jurchen dates is either tobohon 'fifteen' (in Jurchen Empire usage) or juwa shunja 'ten five' (in Ming dynasty usage). tobohon 'fifteen' is obviously not cognate to juwa 'ten' or shunja 'five' and is probably a loan from a para-Mongolic language (a nonstandard Khitan dialect?).
(Khitan as preserved in written records has <TEN FIVE> in both scripts rather than a special word for 'fifteen'. The Khitan word for 'five' was tau < *tabu, not tobo. tobohon might be from a relative of Khitan in which
*CaCu > *CoCo
as there is no Jurchen-internal reason to borrow *tabu- as tobo-.)
Janhunen (2003: 399) reconstructs the potential source of Jurchen tobohon as pre-Proto-Mongolic (dare I say Proto-Serbi-Mongolic?) *tabu.ku/n 'five-teen'.
tobohon is the only one of the '-teen' loanwords in Jurchen
that has a Manchu cognate: tofohon < *topokon
'fifteen'. The Jurchen b : Manchu f < *p
correspondence is irregular and cannot be accounted for via Jurchen or
Manchu-internal sound change.
Was the word borrowed in slightly different forms from two different
dialects of a para-Mongolic language? One of those dialects might have
shifted *b to *p.
Jurchen 'fifty' is susai which isn't much like shunja
'five' and nothing like Janhunen's (2003: 16]) Proto-Mongolic *tabi/n.
2. Last night I ran out of time to describe a linguistic dream I had yesterday. In the dream I came up with various 'underlying' morphological forms for Sanskrit -an nominals.
There was some semblance to reality in the dream: e.g., -an masculine nominative singular present participles were 'derived' from /-ant-s/, /-s/ being the masculine nominative singular suffix preserved in, say, amr̥tas 'Amritas'.
But other parts were pure fantasy - the part that makes me cringe involved trying to explain the nominative singulars of masculine and neuter -an-stem nouns:
rājā < /rājaH/ 'king'; stem rājan /rājaHn/
nāma < /nāma/ 'name'; stem nāman /nāman/
In that bogus fantasy scenario, masculine nouns had an 'underlying' laryngeal /H/ that only surfaced as vowel length when /-n/ was subtracted to form the nominative singular.
In the real world, the final -a of neuter nāma is
from syllabic *n̥: the *n of the stem without a vowel
between it and *m.
As for the final -ā of masculine rājā, Burrow (1955: 230) says that "for phonetic reasons which are not now clear", the vowel of the suffix -an was lengthened, and "there is a tendency for the final semi-vowel of a suffix [i.e., -n] to be elided." He doesn't seem to know what's going on there; I certainly don't.
Wiktionary derives rājā from Proto-Indo-European *ʕʷrḗǵeʕ
(rewritten in my preferred notation). *eʕ regularly becomes
Sanskrit ā, but ... why is *-ʕ there?
3. Today I learned that Arabic فتنة <ftnh> fitna has a much broader scope of meanings than I had thought:
Lane, in his monumental Arabic-English Lexicon compiled from various traditional Arabic lexicographical sources available in Cairo in the mid-19th-century, reported that "to burn" is the "primary signification" of the verb. The verb then came to be applied to the smelting of gold and silver. It was extended to mean causing one to enter into fire and into a state of punishment or affliction. Thus, one says that something caused one to enter al-fitna, i.e. trial, affliction, etc., or more generally, an affliction whereby some good or evil quality is put to the test. Lane glosses the noun fitna as meaning a trial, a probation, affliction, distress or hardship, and says that "the sum total of its meaning in the language of the Arabs" is an affliction whereby one is tried, proved or tested.
The definitions offered by Lane match those suggested by Badawi and Haleem in their dictionary of Qur'anic usage. They gloss the triliteral root as having the following meanings: "to purify gold and silver by smelting them; to burn; to put to the test, to afflict (in particular as a means of testing someone's endurance); to disrupt the peace of a community; to tempt, to seduce, to allure, to infatuate."
The meanings of fitna as found in Classical Arabic largely carry over into Modern Standard Arabic, as evidenced by the recitation of the same set of meanings in Hans Wehr's Dictionary of Modern Written Arabic. In addition, Wehr glosses the noun fitna as also meaning "charm, charmingness, attractiveness; enchantment, captivation, fascination, enticement, temptation; infatuation, intrigue; sedition, riot, discord, dissension, civil strife."
Buckwalter & Parkinson, in their frequency dictionary of Arabic, list the noun fitna as the 1,560th most frequent word in their corpus of over 30 million words from Modern Standard Arabic and colloquial Arabic dialects. They gloss fitna as meaning "charm, allure, enchantment; unrest; riot, rebellion."
And all these years I thought fitna only meant something like 'disturbance'.
At first I couldn't understand how fitna could mean 'charm'. But then I built on the semantic chain established above:
burn > smelt > afflict > test > tempt > tempting quality = charm
It would be interesting to see the order of attestation of the
various meanings in an Arabic historical dictionary. Then again, the
Qur'an already has a wide range of meanings for √f-t-n.
Is the root attested in pre-Qu'ranic Arabic,
and if it is, what does it mean?
4. I just heard gnocchi
pronounced with [tʃi] instead of [ki] in a fake Italian accent on The
King of Queens. I'm surprised the director didn't ask for a retake.
18.104.22.168:58: YELLOW PIG 12/14
songgiyan uliya aniya
juwa juwe biya durhon inenggi
'yellow pig year, ten two month, fourteen day'
1. 'Fourteen' in Jurchen dates is either durhon 'fourteen' (in Jurchen Empire usage) or juwa duin 'ten four' (in Ming dynasty usage). -hon '-teen' is obviously not cognate to juwa 'ten', though dur- 'four-' does resemble Jurchen duin 'four'. Despite that resemblance, durhon is probably a loan from a para-Mongolic language (a nonstandard Khitan dialect?). (Khitan as preserved in written records has <TEN FOUR> in both scripts rather than a special word for 'fourteen'.)
Janhunen (2003: 399) reconstructs the potential source of Jurchen durhon as pre-Proto-Mongolic (dare I say Proto-Serbi-Mongolic?) *dö.r.kü/n* 'four-teen'. The mismatch between the Jurchen and pre-Proto-Mongolic vowels is unexplained. Might it be evidence for a vowel shift in the language that Jurchen borrowed from? Kiyose (1977: 41) proposes that
earlier *ö became Jurchen e in the first syllable and u in the second syllable
earlier *ü became Jurchen u
That would incorrectly predict that *dörkün would become
Jurchen ˟derhun. Was durhon
borrowed after those vowel shifts?
Jurchen 'forty' is dehi < *deki
which seems to be from a para-Mongolic cognate of Janhunen's (2003: 16)
Proto-Mongolic *döci/n 'id.' Jurchen dehi
has e (as Kiyose would predict from *ö) instead of u
(which Kiyose could not predict) like durhon.
Did Jurchen borrow 'forty' and 'fourteen' from different sources, or
was the inconsistency of vowels already in a single source language?
The correspondence of Jurchen h < *k :
Proto-Mongolic c is irregular and needs explanation.
The root *dö 'four' in Proto-Mongolic 'forty' lacks the -r
of Jurchen durhon 'thirteen' and may preserve the
earliest form of 'four' in Serbi-Mongolic. Janhunen (2003: 17) also
observes the r-less root in Proto-Mongolic *dö.tüxer
'third' replaced by *dörbe.düger (with the extended form
of 'four', *do.r.be).
2. Having just mentioned 'forty', by coincidence today as I was
copying the 契丹小字研究 Qidan xiaozi yanjiu (Research on the Khitan
Small Script) hand copy of the
epitaph for Emperor 興宗 Xingzong (1015-1054) of the Khitan Empire, I
came across the block
in line 25. <FORTY> is presumably a phonogram. Could it stand
for something like *deki: i.e., the presumed para-Mongolic
source of the Jurchen word for 'forty'?
3. Why is 'Beijing' spelled ᠪᠡᠭᠡᠵᠢᠩ begejing in the
traditional Mongolian script? It looks like a transcription of a
combination of 北 Middle Chinese *pək 'north' (> modern
Mandarin běi) with 京 modern Mandarin jīng
'capital'. But I doubt the -ge- has anything to do with a
Middle Chinese *-k that was already gone in the north before
Mongolian was first written.
My guess is that bege- is merely an orthographical
convention to write [pəː] by analogy with native words in which spoken
[əː] corresponds to written -ege-: e.g., spoken [təːr] and
written ᠳᠡᠭᠡᠷᠡ deger-e, both from Proto-Mongolic *dexere
'top' (written g is not quite etymological).
4. I just heard the brand name DiGiorno pronounced with an un-Italian [ʒ] instead of [dʒ]. That's an example of how [ʒ] is used by English speakers to signal that a word is foreign (even if that word doesn't actually have it). Another example is in the last topic: Beijing, sometimes pronounced [dʒ] in English even though standard Mandarin j is an affricate [tɕ], not a fricative.
Why is DiGiorno called Delissio in Canada?
5. How did Japanese 大根 <BIG ROOT> daikon come to mean 'ham' in the sense of 'actor known for an exaggerated, over-wrought style'?
How did ham
come to mean that?
I just learned that ham can be an antonym of spam:
'electronic mail that is wanted; mail that is not spam or junk mail'.
22.214.171.124:42: YELLOW PIG 12/13
songgiyan uliya aniya
juwa juwe biya gorhon inenggi
'yellow pig year, ten two month, thirteen day'
1. It just occurred to me that Jurchen doesn't use omsho
'eleven' or jirhon
'twelve' in month names; 'twelfth month' is 'ten two month', though Jin
Jurchen had 'twelve day' which was later replaced by Ming Jurchen 'ten
2. Here I use the more interesting Jin Jurchen day name. 'Thirteen' in Jurchen dates is either gorhon 'thirteen' (in Jurchen Empire usage) or juwa ilan 'ten three' (in Ming dynasty usage). gorhon is obviously not cognate to juwa 'ten' or ilan 'three' and is probably a loan from a para-Mongolic language (a nonstandard Khitan dialect?). (Khitan as preserved in written records has <TEN THREE> in both scripts rather than a special word for 'thirteen'.)
Janhunen (2003: 399) reconstructs the potential source of Jurchen gorhon as pre-Proto-Mongolic (dare I say Proto-Serbi-Mongolic?) *gu.r.ku/n 'three-teen'. The mismatch between the Jurchen and pre-Proto-Mongolic vowels is unexplained. Might it be evidence for a vowel shift in the language that Jurchen borrowed from?
Jurchen 'thirty' is gusin (cf. Janhunen's [2003: 16] Proto-Mongolic *guci/n 'id.') with u instead of o. Did Jurchen borrow 'thirty' and 'thirteen' from different sources, or did a single source language have *u > *o in 'thirteen' but not 'thirty'?
The root gu 'three' in 'thirty' lacks the -r of gorhon
'thirteen' and may preserve the earliest form of 'three' in
Serbi-Mongolic. Janhunen (2003: 17) also observes the r-less
root in Proto-Mongolic *gu.taxar 'third' replaced by *gurba.dugar
(with the extended form of 'three', *gu.r.ba).
3. When I saw the title of Linda Konnerth's "The Proto-Tibeto-Burman *gV- nominalizing prefix" (2016), I immediately thought of Old Chinese *k-. She writes on p. 2,
The velar prefix is notably absent in the Southeastern branch [of Tibeto-Burman] consisting of Yi languages and Burmese. This is, however, not surprising as the Southeastern branch is generally characterized by an isolating typological profile and lack of morphological structure.
In a footnote, she adds,
The same arguably holds for Sinitic languages, where no strong evidence of the prefix has turned up so far. However, two anonymous reviewers point out that both Sagart (1999: 98-107) and Baxter and Sagart (2014: 57) discuss a reconstructed *k-prefix for Old Chinese. The problem with this evidence is that multiple functions are discussed and only one such function is to derive nouns from verbs, while other functions include deriving action verbs and stative verbs apparently from already verbal roots. There are only two examples of the proposed nominalizer *k- given by Baxter and Sagart (2014: 57), and while this is promising, we should not put too much weight on the limited evidence for the time being.
Here is how I would reconstruct those two examples:
方 EOC *CIpaŋ > MOC *CIpɨaŋ > EMC *puaŋ > Md fāng '(to be) square'
匡 EOC *kIpaŋ > MOC *kIpɨaŋ > *kʰpɨaŋ > EMC *kʰuaŋ > Md kuāng 'square basket'
方 isn't always a verb, and its *C might be *k, so
perhaps these are simply two different spellings of *kIpaŋ
which underwent different paths of reduction rather than evidence for *k-prefixation.
The shift of *kp- to *kʰp- has a parallel in Khmer
synchronic phonology: /kp/ is pronounced [kʰp]. (But are there any
other examples in Chinese?)
明 EOC *mI-raŋ > MOC *mIrɨaŋ > EMC *mɨeŋ > Md míng 'to be bright'
囧 EOC *k(V)-mI-raŋ-ʔ > MOC *kmIrɨaŋʔ > EMC *kwɨeŋ > Md jiǒng 'bright window'
These words share a root *raŋ 'bright' also in 朗 *kV-raŋ-ʔ
'to be bright'. The functions of *mI-, *kV-, and *-ʔ
are unknown. If the noun 囧 has the same velar prefix as the stative
verb 朗, then that prefix cannot be a nominalizer.
EOC: Early Old Chinese
EMC: Early Middle Chinese
I feel obligated to say that Pyu does not seem to have a k-nominalizer. I have identified five nominalizers in Pyu:
°o /°o/ (/°/ could be zero or a glottal stop)
kviṃṁ /k.vï/ < */kVPï/
Only one has initial k-, but it has a disyllabic source and
may have originally been a noun.
4. How did this happen?
The Chinese name Ālóng 阿龙, sometimes misread Ayi, refers to Nung (Anong).
Ālóng is a standard Mandarin reading, and I presume Ayi
is also supposed to be standard Mandarin, albeit an erroneous reading.
The trouble is that neither 龙 nor its full form 龍 are read yi.
And I can't imagine anyone mistaking 龙 for 衣 which is read yī.
Benedict (1972: 10) proposes that Pyu might have a "rapprochement" with Nung, but I am not sure what he means by that. As he is discussing subgrouping, he may be saying that Pyu might subgroup with Nung.
126.96.36.199:51: YELLOW PIG 12/12
songgiyan uliya aniya
juwa juwe biya jirhon inenggi
'yellow pig year, ten two month, twelve day'
1. 'Twelve' in Jurchen dates is either jirhon 'twelve' (in Jurchen Empire usage) or juwa juwe 'ten two' (in Ming dynasty usage). I've chosen the more interesting of the two. jirhon is obviously not cognate to juwa 'ten' or juwe 'two' and is probably a loan from a para-Mongolic language (a nonstandard Khitan dialect?). (Khitan as preserved in written records has <TEN TWO> in both scripts rather than a special word for 'twelve'.)
Janhunen (2003: 399) reconstructs the potential source of Jurchen jirhon as pre-Proto-Mongolic (dare I say Proto-Serbi-Mongolic?) *jï.r.ku/n 'two-teen'. *r is a shared element of *jï.r 'two', *gu.r 'three', *dör 'four', and *pa.r 'ten' (not cognate to *-ku/n '-teen'!); the presence of *-r in 'ten' prevents me from regarding *r as a lower numeral prefix. That *-r reminds me of the unrelated lower numeral prefix g- in Tibetan gcig 'one', gnyis 'two', and gsum 'three' (but bzhi 'four'!). Are those real affixes, or has a consonant spread to adjacent numerals? ('Ten' isn't adjacent to 'two' through 'four', so maybe pre-proto-Mongolic 'ten' is simply *par with an original root-final *-r that has nothing to do with the *-r that spread through the lower numerals.)
2. Why is the binomial name of the Japanese badger (穴熊 anaguma 'hole-bear') Meles anakuma with -k-?
In theory, ana 'hole' plus kuma 'bear' could
be either anakuma or anaguma. The latter
has 連濁 rendaku 'consecutive voicing': the voicing of a
voiceless-initial word as the second element of a compound.
Lyman's Law rules out rendaku if a voiceless-initial second element contains a medial voiced obstruent. I view Lyman's Law as a constraint against 'overnasality' because modern Japanese voiced obstruents are from Old Japanese prenasalized obstruents:
/kamu-kaNze/ > ˟/kamu-NkaNze/ [kamuŋganze]
˟/-NkaNze/ 'wind' would have had 'too much nasality': two /NC/. The
actual form is /kamu-kaNze/ with only one /NC/. The modern form still
observes Lyman's law: it is kamikaze¹,
If Lyman's law does not apply to a compound, it's not possible to
predict when rendaku occurs. Which brings me back to whree I
started: in theory, ana 'hole' plus kuma 'bear'
could be either anakuma or anaguma.
Does anakuma without rendaku exist in
dialects? Is there evidence for it existing by 1844 when Coenraad
Jakob Temminck named it in Siebold's
Fauna japonica? If not, is the
specific name from a misreading of 穴熊 <HOLE BEAR>?
¹'God' has two forms in Japanese:
the root form kamu- (Leon Serafim proposed an earlier *kamo-)
the free form kami < *kamu-i (< Serafim's *kamo-i)
The root form cannot stand by itself, but it appears in compounds like Old Japanese kamukaze 'divine wind'.
The modern Japanese form kamikaze replaces bound kamu-
with free kami.
3. I just learned that Siebold [ziːbɔlt] had a daughter who
surname 失本 Shiimoto, a Japanization of Siebold (nowadays Japanized
as シーボルトShiiboruto). I suppose m was considered close
enough to b. And of course there was no way to replicate the
cluster [lt] in Japanese. The big mystery to me is Shii-. I've
never seen that reading for 失 <LOSE> before, and there's no shii
that means 'lose' in Japanese. Shii- is usually written as 椎
4. I never would have guessed Cantonese would have a Yiddish
loanword: 薯嘜 syu4
mak1 'schmuck'. Its etymology
188.8.131.52:05: YELLOW PIG 12/11
songgiyan uliya aniya
juwa juwe biya omsho inenggi
'yellow pig year, ten two month, eleven day'
1. 'Eleven' in Jurchen dates is omsho 'eleven' (in Jurchen
Empire usage) or juwa emu 'ten one' (in Ming dynasty usage).
I've chosen the more interesting of the two. omsho is obviously
not cognate to juwa 'ten' or emu
'one' and is probably a loan from a para-Mongolic language (a
nonstandard Khitan dialect?). (Khitan as preserved in written records
has <TEN ONE> in both scripts rather than a special word for
Janhunen (2003: 399) reconstructs the potential source of Jurchen omsho
as pre-Proto-Mongolic (dare I say Proto-Serbi-Mongolic?) *omcon
'eleven' and suggests it may be "connected" to a nominal root *onca
'special, additional' (> Written Mongolian onca
'special'). Note, however, that *omcon has *-m-
and *onca has *-n-.
2. "Writing kanji on the air is even better practice than writing on paper" - but how about on the glass in the shower? I don't write kanji there, but I do write other scripts I'm studying. Being flustered when writing Tangut, Jurchen, and Khitan on the sands of Waikiki Beach motivated me to practice writing all three every day. Ironically, I just realized I missed yesterday's practice. Off to make up for it now ...
3. I was going to write about how I thought Jurchen
<STAR> osiha 'star'
might be a recycled logogram for <OX> (cf. Chinese牛 <OX>), but I already did so back in November. Duh. I can't remember what I wrote just weeks ago.
I'm going to stop here because I spent too much time writing addenda for the last two posts.