Victor Mair’s Language Log post starts off with Japanese 奴隷 dorei ‘slave,’ of which Mair says “Coming at 奴隷 from the Sinitic side, my instinct is to read 奴隷 as beginning with an n-” (in Mandarin it’s núlì) and continues “So I started to ask around how is it that Japanese has a d- initial for 奴隷 (‘slave’) and Sinitic has an n- initial?” The heart of the post is a long and interesting response by John Whitman:
The alternation btw d- and n- with 奴 reflects the general alternation between kan’on 漢音 and go’on 呉音; both go’on and kan’on exhibit characteristics of Middle Sinitic (MS) in Sino-Japanese. The kan’on 漢音 for 奴 is do, but the go’on is nu, identical to the usually reconstructed MS nu for 奴. In this case, the go’on reading is relatively unusual on the Japanese side, but it occurs e.g. in the reading 奴婢 (nuhi ぬひ), the category of slaves in the Ritsuryō 律令 Nara Period legal system.
The d-~n- alternation is standard when there is an opposition btw kan’on and go’on readings involving original MS /n/, for example 男性 dansei ‘male’ vs 男体 nantai ‘male body’. The alternation between 女性 zyosei < dyosei ‘female’ vs女体 nyotai ‘female body is the same thing.
This reflects a change in Sinitic, not Japanese. Some northern MS dialects in roughly Tang times depalatalized MS /m/, /n/, /ng/. South Coblin has a detailed study of this, looking not just at the phenomenon in Japanese kan’on but at Tibetan and intra-Sinitic Buddhistic readings. A mystery, unresolved as far as I know, is how this made it into Japanese kan’on 漢音 but not Sino-Korean, which are both held to have been borrowed around the same time, roughly mid-late Tang, perhaps a bit later in the SK case. One possibility is that the denasalizing region included Chang’an; the Koreans were savvy and in constant contact enough with China to understand that the denasalized pronunciation was substandard, even if associated with the capital region, while the Japanese clerics who imported the readings were less with it, or more superliteralist.
In modern Japanese, kan’on readings are vastly less marked, and almost always used in neologized kango 漢音. Go’on readings have a strong association with Buddhism. The 呉 wu2 designation refers most likely not to any region of China (such as Southeastern China/Suzhgou~Shanghai region), but to the Korean peninsula. The kun 訓 vernacular reading of 呉, kure, refers to Korea and is probably the same word as Korean 고려 Koryŏ [koryə] or possibly Kuryŏ [kuryə], what you get if you subtract the flattering 高 from 高句麗, as non-Korean texts often do. This reflects the fact that both Sinography and Buddhism were originally imported to Japan from Paekche.
What bothers me is the reference to Coblin’s “detailed study of this”; Mair has linked it to Academia.edu, but I found it at JSTOR, and it is not at all a study of the phenomenon in question but a general discussion of early Northwest Chinese phonology. I presume the section Whitman refers to is 2.1 (pp. 12-13) on nasal initials, and I also presume by “depalatalized” he means “denasalized” (since that’s the only thing that makes sense in the context of /n/ > /d/), but I don’t see anything in the passage that would explain the Japanese development (he talks about [nd] but not [d]). If anyone has thoughts about all this, let’s hear them!
I also presume by “depalatalized” he means “denasalized”
Yeah, it’s just a careless mistake. The LL comments correctly say “denasalisation.”
The South Coblin paper does actually talk about this (section 2.1, p12.)
Eastern Oti-Volta does the opposite, changing *b *d to /m n/ before nasalised vowels, as with e.g. Farefare dãam “beer”, beside Nateni nāàmā.
Universals? What “universals”?
The South Coblin paper does actually talk about this (section 2.1, p12.)
He talks about “Tang denasalisation,” yes, but he doesn’t give any examples that involve [d].
Denasalization is also observed in contemporary Korean, at least in Gyeonggi/Seoul dialect.
Previously mentioned here and briefly discussed here (much later in the same thread).
Among Sinitic varieties, Southern Min experienced denasalization of nasal initials in many cases, which was probably an independent development to what took place in early Northwest Chinese. However, after /n/ turned into /d/, it merged with /l/, so the initial of 奴, 男, and 女 is usually treated as /l/ in most romanizations of Southern Min (even if there seems to be variation in the actual realization in different dialects).
In Modern Korean, voiced stops such as [b] and [d] only occur intervocally or after a sonorant consonant and before a vowel. They do not occur word-initially. Japanese has no such constraints. If the situation was similar at the time that Sino-Korean and kan’on borrowings occurred, it would explain why Japanese mapped the denasalized initials to their voiced stops while Korean didn’t. Because initial voiced stops weren’t available in Korean, plain nasals would still have been the closest match even to any denasalized initials in early Northwest Chinese.
Interesting and plausible — thanks!
“he talks about [nd] but not [d].”
I have a dim memory, which may well be wrong (but I don’t have time to check), that in Old Japanese the “voiced” stop series was actually prenasalized. If so, then the Japanese readings might well just have been very precise adoptions of the Chinese sounds in question.
“Because initial voiced stops weren’t available in Korean, plain nasals would still have been the closest match even to any denasalized initials in early Northwest Chinese.”
A very nice-sounding explanation! I hope it’s true.
Vovin’s grammar of Classical Japanese prose says that the voiced stops were “possibly” prenasalised, including initially.
Prior to the adoption of Chinese loanwords, Old Japanese didn’t have word-initial initial voiced stops, only non-initial, where they do indeed seem to have been prenasalised. Frellesvig’s History of the Japanese Language says “this feature of pronunciation is generally thought not to have been lost until early in the NJ [Modern Japanese] period.” Among other evidence, he cites things like yonde, the gerund of yobu “call”; IIRC, there is evidence from early European transcriptions to support this, too.
Swahili voiced stops are implosive unless they are prenasalised, a very common pattern in Bantu languages. (In other words, explosive voiced stops are always prenasalised.)
“Prior to the adoption of Chinese loanwords, Old Japanese didn’t have word-initial initial voiced stops, only non-initial, where they do indeed seem to have been prenasalised.”
Interesting. In that case, I suppose the adoption of initial [nd] (or [ⁿd], but I’m being typographically lazy) as [nd] in Japanese wouldn’t be quite as trivial as I was suspecting, but support from medial [nd] (etc.) would presumably still make it a good deal easier for Japanese speakers to perceive and replicate the distinction initially than it would be for Koreans (if these systemic factors are really what mattered most, rather than sociolinguistic ones).
This seems like the kind of topic an enterprising student might want to consider for a thesis or something.
Swahili voiced stops are implosive unless they are prenasalised, a very common pattern in Bantu languages.
My very poorly informed impression of European languages pronounced by African immigrants is that their basis of articulation has as a feature something like a lowered larynx. Is this right? And if so, would it be related to that default implosive articulation of stops? Noting also that a great part (a majority?) of European immigrants from Africa come from Western Africa and speak non-Bantu languages.
Yes, it doesn’t seem likely. Lots of Chadic languages, including Hausa, have contrastive voiced implosives, as does Fulfulde, but most recent African immigrants don’t speak languages like those. (I don’t think many victims of the slave trade did, either.)
The Oti-Volta languages don’t have any implosives (not that many African immigrants have Oti-Volta L1s either.) The weirdo Northwest Bantu languages don’t seem to either, judging by the few grammatical descriptions I’ve seen.
Having said that, a lot of older descriptions of African languages are pretty vague on phonetic detail, especially with subphonemic stuff. Most accounts of Swahili don’t mention that the voiced stops are implosive by default (or even that the voiceless stops have an actual aspirated/unaspirated contrast, for that matter. The standard Swahili orthography is not actually phonemic …)
John Stewart had a whole elaborate theory of “proto-Niger-Congo” in which all stops had a contrastive fortis/lenis opposition, which he seems to have conceived as explosive/not-explosive, but even he eventually conceded “the non-explosive unvoiced stops of Proto-Potou-Akanic-Bantu thus appear to have been lost on a vast scale.” Much more likely, they were never there in the first place. Ebrié, one of his particular specialist languages, seems to be the sole “Niger-Congo” language with a synchronic stop system anything like his proposals.
There are a lot of unsolved questions about proto-Volta-Congo stops. Even within Gur, proto-Oti-Volta *v corresponds to Grusi *v, but proto-Oti-Volta *b corresponds partly to Grusi *b, and partly to *v:
Mooré vãoogo “leaf”, Kassem vɔ̄ɔ;
Mooré biiga “child”, Chakali bìé;
Mooré baaga “dog”, Chakali váà.
Looking at other examples, this doesn’t seem to be explicable by environment, and it presumably does reflect a distinction in proto-Central Gur. It does partly correlate with some of Stewart’s examples, but Bantu, disobligingly, has only /b/ in both sets …
Thanks. Otherwise, does my vague idea of African L2 phonetics sound to you anywhere near plausible?
I think there is too much variation among the major West African languages themselves for explanations of that kind to be plausible.
On the overall acoustic impression created by languages, I remember a conversation I once had with Tony Naden, where he was talking about “swallowed” languages – with Kusaal as an example, in fact, contrasting with the language in which he is particularly expert, Mampruli.
I’ve not myself been much exposed to Mampruli, but I was often struck by the typically very clear-and-distinct enunciation of Mooré speakers compared with Kusaal: Italian versus Danish. (Well, versus English, maybe. Kusaal is not that blurry …)
Mooré speakers often render velars practically as uvulars before back vowels, which also gives a rather different impression from Kusaal, where the allophonic variation of velars is a lot more like English.
Maybe what you’re hearing is intonation – perhaps the default with the tones stripped off, or with Germanic stress reinterpreted as tone.
Not intonation, and I was thinking of French more than German. It has lower resonances, which makes the voice sound deeper, though it isn’t.
Certainly the great majority of West African languages are tonal, (though not non-Mel Atlantic, nor the western bits of Songhay.) And stress contrasts in these languages are rare*, so reinterpretation of stress as tone would be natural, and indeed has actually happened in the West African English-lexifier creoles.
* Stress contrasts are possible in principle in Kusaal, but there are very few environments where a minimal pair could even potentially appear, and there seem to be none at all in practice. But cf e.g. dabin “beer dregs” versus vabin “lie down prone.”
I have no idea how closely the percentage distribution of L1’s of recentish immigrants to Europe from sub-Saharan Africa matches that of the involuntary migrants to the New World from ditto during the days of the Atlantic slave trade. And some individual languages may of course have evolved notably in the intervening centuries. I think the usual story about creoles is that they were often quite sensitive to, and reflective of, the L1 of whatever group due to contingent historical circumstance was the “first mover” in their creation, with speakers of other L1’s who adopted the particular creole later to some extent being stuck with and adopting to what was already there. But I don’t know that in modern Europe there would be a “here’s how African immigrants speak the local majority language” variety with that same dynamic.
File under “oh for god’s sake”: a follow-up at the Log, a guest post by Chau Wu, says “My study of West-to-East lexical loans suggests that the origin of this word is Ancient Greek δοȗλos (doȗlos, m.) and δοȗλα (doȗla, f.), which mean ‘slave’.” This is, of course, catnip to Mair.
Can’t even get the Greek right (δοȗλα for the actual Greek word δούλη.) But Mair-style fantasy linguistics floats serenely above fact, let alone basic accuracy.
I see that Mair has been laying into a Chinese YouTuber who had the temerity to say that she thought the traditional Chinese script was beautiful. Such views cannot be permitted.
I saw Goody Proctor admiring Chinese characters!
Chau Wu is the kook who found Old English elements in Taiwanese:
https://sino-platonic.org/complete/spp262_taiwanese_western_lexicons.pdf
Mair is an embarrassment to linguistics.
A Greek origin is, well, “oh, for god’s sake” seems about right. But it does make me wonder when 奴隸 was actually first attested in Chinese. I know that 奴 on its own could mean slave, but I’m not immediately sure when the longer 奴隸 came into use.
Chau Wu is the kook who found Old English elements in Taiwanese
Wow. I’m reminded of Valerii A. Chudinov, who has found Slavic runes not only on Byzantine icons of the fifth to tenth centuries but on Greek vases from the second to sixth centuries BC—not to mention the prehistoric cave paintings of France.
I saw Goody Proctor admiring Chinese characters!
at least we all have each other, here at the devil’s sacrament.
Comment made, we’ll see if it gets approved (links to a Wikipedia article and a paper on JSTOR, so automatically sent to moderation, I suppose). I deleted all expressions of exasperation except boldface, for example “reinventing the square wheel”.
It’s live.
Thanks @DM, and some follow-up discussion seems to have survived so far.
I’m perplexed. (And I daren’t ask dumb questions over there.)
Is it that 奴隷 and its modern pronunciation can be explained by purely within-Chinese sound changes? So it’s an ancient word(?) I think wikip [and katarina?] is saying the glyph 奴 dates back to Oracle Bone script.
If it’s a borrowing (either from Greek or from some common ancestor), why/how would Chinese take a new word for what was an already ancient practice? [which is in effect what the first comment from wgj is asking. I see no direct answer to it.]
So Chau has found merely a soundalike? How alike would they actually sound as at the time of the alleged borrowing.
“So Chau has found merely a soundalike? How alike would they actually sound as at the time of the alleged borrowing.”
Yes, to the first.
For the second, according to Wiktionary, the Old Chinese pronunciation of the first element was *nˤa in the widely used Baxter-Sagart reconstruction. The second element is, less helpfully, reconstructed as *[r]ˤə[p]-s — the brackets mean that the segments in question might be *r and *p, or might be certain other sounds. I’m actually not quite sure why the initial is regarded as uncertain, as my understanding (based on Hill 2019) is that Middle Chinese l- rather mechanically reflects OC *r-. But I’ve only had a chance to quickly double-check Hill, and could easily be missing something.
A comment from that thread:
The paper is the one I linked to above. P150 tells us
This is not an isolated aberration. The whole thing is like that. Have a look for yourselves. It’s “Hebrew is Greek” level stuff.
I’m not sure that attempting to engage in reasoned argument with Wu or Mair is a productive activity. You’re just falling into the “someone is wrong on the internet” trap. Language Log is now just a dead zone when it comes to comparative stuff. Shed a tear for what once was, and pass on.
You’re just falling into the “someone is wrong on the internet” trap.
He knows that and has said, I believe, that he can’t help it. But I agree that it’s a waste of time.
I’m not going to take the time to read the “papers” or anything. But I really don’t want to let this kind of thing stand uncommented in public. Lots of people don’t know that the science of historical-comparative linguistics exists, and we’ll definitely keep complaining about that fact if we never speak up.
Well, I admire your fortitude.
(Also your restraint.)
As do I.
Since all threads are one, Xerib’s dive into the etymology of Greek doulos was just brought up by a resurfacing thread.
(… which ktschwarz no doubt dug up for that purpose before reviving its hat theme.)
(Also your restraint.)
Restraint to the point of obliqueness is necessary for any comment to survive. (In earlier threads, my more direct observations on Chau’s ‘findings’ all got nixed.)
For me who is not ‘inside baseball’, the obliqueness yields only perplexity.