Old Sinitic "wheat" and Early Middle Sinitic "camel"
« previous post | next post »
[This is a guest post by Chris Button]
OC uvulars tended to condition rounding (e.g OC q- becoming EMC kw-). In the case of ʁ-, we sometimes get m- (for a modern-day example, note how惟, which also had a ʁ- onset in Old Chinese, gives an m- reflex in Fuzhou Min). The classic example is 卯, where Pulleyblank once postulated ʁ- and Li Fang-kuei notes lack of evidence for a cluster, such as ml- or mr-, in its Tai loan. Unfortunately Li’s Tai evidence tends to either be ignored (e.g. 丑 hr- is often erroneously reconstructed with a nasal hn- based on misleading xiesheng evidence) or overly literally interpreted (e.g. 戌 χ- being treated as something like sm-).
麥 məɨkʲ ← ʁək “wheat”, comparable with Proto-Indo-European rugh- (or rw̩kʱ-) "rye", makes far more sense as a loan for 來 ləj ← rəɣː “come”. If the word for 麥 had an original m-, why on earth would the graph have been borrowed for a word with r- as an onset?
As for駱駝 “camel” (EMC lak.da in which the l- comes from earlier r-), its alternative spelling of 橐駝 (EMC thak.da) in interesting. Clauson suspected a Tocharian origin that he thought was lost. The reconstructed Tocharian form partākto “camel” accords well with Chinese apart from the beginning portion. I wonder if 匹 EMC phjit was reanalyzed as a measure word as part of the loan process so that we’re really dealing with 匹駱駝 or 匹橐駝? In terms of a form prior to EMC, we would probably have had something like pətrakda, but I’d need to think about that a little more (Pulleyblank notes how OC *-al rhymes, such as 駝, are used for *-a from the 3rd century and occasionally earlier.)
Selected readings
- "So spoke Zoroaster: camels and ancient Sinitic reconstructions" (1/13/21)
- "Old Chinese onsets and the calendrical signs" (1/12/23)
- "The Alphabet and the Zodiac" (12/6/22)
- "On the Origins of the Alphabet: The Cycle of Emmer Wheat and Seed/Word Selection within the Proto-Sinaitic/Phoenician/
Hebrew Alphazodiac and the Chinese Lunar Zodiac," by Brian R. Pellar. (free pdf) - "Of shumai and Old Sinitic reconstructions" (7/19/16)
- "Old Ukrainian windmills and Old Sinitic reconstructions" (3/27/22)
- "Words for cereals" (7/27/16)
- "Pulled noodles: Uyghur läghmän and Mandarin lāmiàn" (8/8/14)
Chris Button said,
June 13, 2023 @ 7:08 am
Needless to say, 麥 as məɨkʲ ← ʁək “wheat”, and a comparison with PIE rugh- (or rw̩kʱ-) "rye", makes ideas of any association with Japanese “mugi” even less likely. Although, looking at comments by people like Kamei or Miyake, it doesn’t seem there was much support for that association even with an original m- instead of a rhotic.
Chris Button said,
June 13, 2023 @ 2:10 pm
Another nice piece of evidence for ʁ- shifting to m- (beyond how it accounts for various phonetic/xiesheng series) is the case of 貉.
貉 ɣak ← ɢak (ɢ- being an allophone of ʁ-) is also read 貉 maɨkʲ ← ʁak and then loaned for 禡 maɨʰ ← mraɣːs (or mraks).
I’ve never seen a good account for the readings of 貉 as ɣak and maɨkʲ outside of Pulleyblank’s observation that the onset must have involved some kind of labialization.
Chris Button said,
June 14, 2023 @ 12:03 pm
I should add that the rounding sometimes associated with uvulars was proposed by Pulleyblank for Old Chinese back in the 1970s and 1980s. It seems like Pan Wuyun in the late 1990s and then Jin Lixin in the 2000s then adopted similar ideas. The specific details are different, but the concept is the same.
Chris Button said,
June 14, 2023 @ 8:24 pm
Meanwhile, Baxter & Sagart have two distinct series of uvulars and labio-uvulars to parallel their velars and labiovelars. That strikes me as unnecessary and quite a divergence from the spirit of the original proposals.
Pulleyblank’s 1970s proposal focuses on uvular codas. Notably it accounts for the lack of a nasal coda (a uvular nasal being topologically and articulatorily unlikely) where other reconstructions just have an unexplained gap. It also accounts for why the Middle Chinese reflexes of the uvular codas attest sporadic but inconsistent rounding—again a major stumbling block for many other reconstructions.
Chris Button said,
June 14, 2023 @ 8:25 pm
*typologically
Chris Button said,
June 14, 2023 @ 9:50 pm
The phonetics via ʋ- may be noted in Pulleyblank's observation that LMC ʋ- gave Fuzhou Min m-, and Pulleyblank's observation of occasional Han-time use of m- to transcribe Sanskrit ʋ-.
Jichang Lulu said,
June 16, 2023 @ 12:42 pm
Do you have references for Kamei and Miyake’s comments on
mugi?
There has been much speculation on the further origin of Toch. B
*partākto. A
recent-ish paper has references to earlier literature. Skt.
pṛdāku ‘…all manner of animals’ (already ‘snake’ in Vedic) is
presumably part of less the solution than the problem — which problem
includes the Greek from which leopard (e.g., Mayrhofer notes the Iranian
etc. parallels but leaves the etymology ‘unclear’).
One remains only too eager to see the forthcoming exposition of your
reconstruction, without which discussion of specific etyma here is
underdertermined.
Chris Button said,
June 16, 2023 @ 11:02 pm
@ Jichang Lulu
Kamei monograph: “Chinese Borrowings in Prehistoric Japanese” (1954).
Miyake article: “Pre-Sino-Korean and Pre-Sino-Japanese: Reexamining an old problem from a modern perspective” (1997).
Regarding “camel”, it’s a very speculative idea, but I haven’t come across anything better. Hopefully someone can uncover some more fodder for further speculation. The proposal for “wheat” is far more compelling in my opinion—the fact that there is a nice PIE comparison is simply a bonus rather than the impetus of the discussion.
All these little tidbits that I share once in a while on LLog are things I happen to stumble across as I compile my dictionary. The OP here is nothing more than a direct copy of an email I sent to Prof. Mair since I knew he would be interested. I then said he could post it here on the off chance anything interesting comes out of the comments. You have, alas, so far provided the only one.
I appreciate that the posts can seem a little random. But I sit outside of academia so have nothing to gain by publishing multiple theoretical treatises based on very carefully worded cherry-picked data. I would rather let the dictionary (when it appears) speak for itself—the proof will be in the pudding.
In terms of my reconstruction, I by and large follow ideas proposed by Edwin Pulleyblank at some point or another, which includes ones he proposed and then abandoned. However I do diverge in certain areas (both through novel approaches and where I disagree with him), and do not mean to imply that Pulleyblank would have supported anything I propose. I do, however, owe him credit for the majority of places where I happen to be “correct”. For Middle Chinese, I simply follow him verbatim although make some typographical tweaks here and there,
The challenge is that Pulleyblank’s Old Chinese proposals are often hard to follow (both in terms of exposition and how they challenge the orthodoxy) just as hard to locate (buried in old issues of academic journals), and do not exist in one comprehensive volume with a list of example reconstructions. Unfortunately that also means that people just looking for reconstructed forms tend to turn to more accessible work, which (in my opinion at least) often struggles to adequately account for the evidence,
David Marjanović said,
July 10, 2023 @ 2:09 pm
First of all, I really like the phonemic transcription */kʱ/; it shows clearly that there's only a single-feature difference to */k/.
Whether this word can be reconstructed to PIE is doubtful, however. It's uncontroversial to set up a Proto-Germanic *ruɣiz (not attested in East Germanic, but there's no rye in the Bible, and it's in Northwest Germanic and Gutnish); it may also be possible to postulate a Proto-Balto-Slavic *rugís, though Szemerényi thought in the 1960s, for reasons I can't look up right now, that that's not actually possible and the word must have been borrowed separately into the separate Balto-Slavic branches; and… that's it. Wiktionary, at least, does not know of any postulated Iranian or Tocharian cognates (to mention the two branches geographically relevant to Sinitic) or indeed any others at all. (Proto-Finnic *ruɣis pretty much has to be a Germanic loan for several reasons; no other loans seem to have been postulated.)
If we take PBS *rugís for granted, it seems easy to set up a PIE **rw̩kʱís. But that may not be possible. There are no known clear cases of word-initial *r in PIE. And if we just slap a laryngeal on it (**Hrw̩kʱís), that should, as far as I understand, have given the first syllable "acute intonation" in Balto-Slavic (a tone or register). It isn't there.
PIE wasn't big on words for cereals anyway. To find a word for something as specific as rye would be quite unexpected. Probably rye wasn't even deliberately cultivated that early.
And on top of all that, we'd need to come up with a scenario for how PIE (or later) *r got borrowed as *ʁ and not as *r in early Sinitic (or earlier).
I think the most likely scenario is a local Scandinavian substrate loan into Germanic that was passed on to Balto-Slavic (once or several times) and Finnic but no further, and the Sinitic "wheat" word is unrelated.
David Marjanović said,
July 10, 2023 @ 2:15 pm
I arguably forgot to mention that PIE *h₃ was most likely [ʁ], so the Sinitic *ʁ you postulate could in principle represent that, and not the *r, of a PIE **h₃rw̩kʱís. But that leaves all the other problems in place: no acute in Balto-Slavic, no potential IE cognates outside Germanic and Balto-Slavic, and the history and geography of rye cultivation.
Chris Button said,
July 10, 2023 @ 9:14 pm
@ David Marjanović
Thank you for the post.
Sounds like we have a similar take on some of these hotly debated PIE issues …
Couldn't agree more. In fact, I just posted about that here: https://languagelog.ldc.upenn.edu/nll/?p=59422#comment-1606888
What you (or wiktionary) write above seems to agree with what the literature broadly says, but it misses some crucial details …
Mallory & Adams (2006) claim the word exists as far as the Pamir region:
" 'Rye' is found mostly in the North-West … but also in the Iranian Pamir languages (e.g. Shughni roɣ̌z 'ear of rye')."
Kroonen et. al. (2022) seem to challenge Mallory & Adams' Pamir association, but still support the Iranian connection elsewhere while noting a Celtic link too:
"Importantly, forms resembling the European word appear in the Permic languages … these have been adduced to substantiate a Proto-Iranian form … No such form is attested anywhere in Iranian, however. Shu. (Bajui) rōɣ̌ʒ and Rosh. růz ‘ear (of grain)’ … rather continue *rārza- … Given the presence of other Pre-Proto-Slavic loans in Ossetic … it is perhaps more plausible that the Slavic word was adopted by Iron Age steppe Iranian and from there permeated into Permic. Without a certain Iranian continuant, the word at any rate receives a (North) European distribution. In view of this areal range, it may be a late (dialectal) lexical innovation or—if the irregular Celtic form [Welsh rhyg "rye"] is to be relied on—a non-Indo-European Wanderwort."
Whether rye or wheat shouldn't really matter–same family
I'm not familiar with the challenge to the standard practice of reconstructing an *r- onset. Could you tell me more?
David Marjanović said,
July 11, 2023 @ 3:28 pm
Ooh, a Pre-Slavic loan in Vaguely Scythian makes perfect sense. There are plenty of loans in the other direction, so there was contact; and semantically this is a perfect candidate for a loan into Iranian instead of into Slavic.
I wonder if the Welsh form is straight out of Old English.
Sure; reconstructing *r- ceased to be standard a few decades ago. The latest and most compact look at this issue seems to be chapter 5 of this MA thesis (…although it's a very Leidenite thesis).
Typologically, this restriction is common (even more common than mentioned in the thesis). Basically all of northern Eurasia shares it or used to, and it's also found elsewhere.
David Marjanović said,
July 12, 2023 @ 1:43 pm
Sure, but there isn't a reconstructible word for anything as specific as "wheat" either, or "barley" or "millet" for that matter. I recommend the first few pages of this paper. (The paper's main thesis is less convincing, but that's a separate topic.)
Chris Button said,
July 12, 2023 @ 10:34 pm
Thanks for sharing. I'll look into it. It seems, however, that the matter is hardly settled. Regardless, it doesn't seem to affect my original proposal in the OP about an external relationship for Old Chinese here.
I have seen this argument come up before. I'd have to look into it some more, but my suspicion is that this is more to do with the nebulous concept of "rhotic" than any actual restriction–the unifying concept being acoustic rather than articulatory (as Ladefoged & Maddieson discuss in their tome)
I'm no expert at all on such things. But I am highly skeptical.
Just to quote Mallory & Adams (1997:8) on the topic: "… we can see that there is no case whatsoever for assuming that the ancestors of all the Indo-European stocks did not know cereal agriculture."
Chris Button said,
July 12, 2023 @ 10:38 pm
Couple more comments/questions on PIE "laryngeals" here too:
https://languagelog.ldc.upenn.edu/nll/?p=59422#comment-1607003
David Marjanović said,
July 13, 2023 @ 10:53 am
Not at all.
They certainly knew cereal agriculture. But they didn't do enough of it to have more than the most rudimentary vocabulary for it. That is pretty clear.
Chris Button said,
July 13, 2023 @ 11:26 am
Well, like I said, it's acoustic factors not articulatory factors that need to be taken into account.
Chris Button said,
July 13, 2023 @ 11:40 am
Except that Mallory & Adam's marshal quite a lot of evidence to the contrary and claim "no case whatsoever" for the counterargument. I'm out of my area of experise here, but this sounds a like the r- onset issue. Debateable.
In any case, like i said on the other thread, I'm just looking for an external connection, and am not surprised that nothing convincing with m- has turned up because I don't believe the OC onset was originally m-.
Chris Button said,
July 13, 2023 @ 11:53 am
It would also be interesting to hear from Mallory & Adam's on the possible Pamir association (where Kroonen et al suggest an alternative) since that's Tian Shan territory.
David Marjanović said,
July 17, 2023 @ 3:45 pm
Don't write about a paper from 1997 in the present tense. The paper I linked to is twenty years younger.
But that doesn't even seem to matter. You quoted Mallory & Adams as saying there's "no case whatsoever for assuming that the ancestors of all the Indo-European stocks did not know cereal agriculture." I agree with this, and so did Garnier et al. (2017)! What Garnier et al. (briefly) argued for is that PIE had a very small vocabulary for agriculture, suggesting a dearth of first-hand experience with it.
Here is a similar look at the same issue. It dates from 2014 and approvingly cites a work by Mallory from 2013…
David Marjanović said,
July 17, 2023 @ 5:09 pm
…oh, and, here is a thorough review of the agricultural terms in early Indo-Iranian, with basically the same conclusion again. As a side effect it argues strongly against connecting the Pamir "rye" words with the Germanic and Balto-Slavic ones: the sound changes don't work.
Chris Button said,
July 17, 2023 @ 10:23 pm
You're betraying your background in the "hard" sciences. Two decades old is positively young!
That seems to agree with Kroonen et. al.'s analysis without the alternative Iranian connection that they do note.
I still maintain that it is interesting that a very similar sounding word to the OC reconstruction I propose is in the vicinity and works phonologically. Perhaps they are unrelated. I'd love to hear some alternative proposals. I do, however, think that the focus on it starting with an m- in OC has been a stumbling block for anything presented in the past. As Pulleyblank (1996) put it: "At present I can see no similarity between the Chinese names for these grains [wheat and barley] and any Indo-European words."
Chris Button said,
July 19, 2023 @ 8:54 pm
Continued here before this thread closes:
https://languagelog.ldc.upenn.edu/nll/?p=59629#comment-1607223
David Marjanović said,
July 21, 2023 @ 6:30 pm
Well, sure. If I understand you correctly, you're proposing that [ʁ] changed into [m] (very weird), but only in some words (how?), and that [r] was borrowed as [ʁ] even though [r] was available and the distinction was phonemic… that seems like a long chain of inference. I suppose it'll be in your book, so there's not much point in commenting on it, but until the book comes out, of course it'll remain a stumbling block.