“Divan” is one of the most complicated words I know. The American Heritage Dictionary gives the following definitions:
1. A long backless sofa, especially one set with pillows against a wall.
2. a. A counting room, tribunal, or public audience room in Muslim countries.
b. The seat used by an administrator when holding audience.
c. A government bureau or council chamber.
3. A coffeehouse or smoking room.
4. A book of poems, especially one written in Arabic or Persian by a single author.
The OED adds the meaning ‘a room having one side entirely open towards a court, garden, river, or other prospect’ and expands on the fourth sense as follows: “A Persian name for a collection of poems (Persian, Arabic, Hindustani, Turkish); spec. a series of poems by one author, the rimes of which usually run through the whole alphabet. [From the original sense ‘collection of written sheets’, perh. influenced by later uses of the word.]” And speaking of “original sense,” check out the etymology:
A word originally Persian, dēvān, now dīwān, in Arabic pronounced dīwān, diwān; in Turkish divān, whence in many European langs., It. divano, Sp., Pg., F. divan. Originally, in early use, a brochure, or fascicle of written leaves or sheets, hence a collection of poems, also a muster-roll or register (of soldiers, persons, accounts, taxes, etc.); a military pay-book, an account-book; an office of accounts, a custom-house; a tribunal of revenue or of justice; a court; a council of state, senate; a council-chamber, a (cushioned) bench. The East Indian form and use of the word is given under DEWAN. Another European form, older than divan, and app. directly from Arabic, is It. dovana, doana, now dogana, F. douane (in 15th c. douwaine), custom-house: see DOUANE.
For a more discursive collection of definitions, with 19th-century stabs at etymology, see the Hobson-Jobson entry. The mix of senses is so confusing that when I asked the proprietors of an excellent Lebanese restaurant in Astoria called Al Dewan (long defunct, I’m afraid) why it was so named, they muttered and fumfered and couldn’t come up with anything convincing. (To my mind, it was clearly named with the ‘poetry collection’ sense in view, since the window displayed a plaster model of an open book with the name inscribed calligraphically, but when I drew their attention to this, they shrugged—I’m guessing whoever named the place and ordered the plaster book was no longer around, and nobody else knew.)
And where does the Persian word come from, you ask? The AHD says:
Persian dīwān, place of assembly, roster, probably from Old Iranian *dipivahanam, document house : Old Persian dip-, writing, document (from Akkadian tuppu, tablet, letter, from Sumerian dub) + Old Persian vahanam, house; see wes-¹ in the Appendix of Indo-European roots.
I hope that’s correct, because there aren’t many words in English that go back to Sumerian (tunic and chiton are two more; according to AHD they go back via Akkadian kitû, kita’um, ‘flax, linen’ to Sumerian gada, gida).
The reason I’m telling you all this is to give you the background for appreciating the amusing error made in this article (Google cache; the original story has gone 404) by Ana Keshelashvili:
Revaz Baramidze looked in amazement at the crowd of people gathered in Parnassus, Tbilisi’s newest bookstore. The store’s two rooms were so tightly packed that it was difficult to move around, and more people stood outside waiting to get in.
“What do I see, so many young people and everybody came to buy a book. I can’t predict, but it seems to me that we are turning back to reading literature,” said Baramidze, professor of literature at Tbilisi State University.
That cold but sunny winter afternoon, Vakhushti Kotetishvili was seated at a small desk in the downstairs room, signing copies of his newly published collection “East-West Sofa.”…
Now, you also have to know that a famous collection of Goethe’s was called West-östlicher Divan, translated as West-Eastern Divan. I strongly suspect that Mr. Kotetishvili (described here as “an incredibly dignified translator of persian poetry”) gave his book the same title in Georgian. But Ms. Keshelashvili looked up Georgian დივანი (divani) in her Georgian-English dictionary, found “sofa,” and the rest was history.
I’d feel worse about making public fun of Ms. Keshelashvili if she hadn’t publicly identified me as David Foster Wallace in her master’s thesis, “Patterns of Self-Expression and Impression Management in Blogs” (pdf; Google cache here). Check out #104 in APPENDIX A: LIST OF BLOGS ANALYZED.
Addendum. See now dahween and divan at Balashon, which (among other things) explains the origin of Chicken Divan (mentioned in the thread below).
Update (Mar. 2025). The OED has updated its entry, but not very effectively; see Xerîb’s comment below.
“East-West Sofa” — well, that’s a dialect form. We would use the correct form where I come from — “East-West Davenport”.
Just re-notifying people that the 1000-page Hobson-Jobson can be found at bookfinder.com for about $10 including shipping. I’m waiting for mine — I expect a shoddy, slightly musty Indian reprint, but for one cent a page I won’t complain a bit. (Many public-domain books in English relating to India, Asia, Hinduism, or Buddhism have been reprinted cheaply in India.)
XICc Anglo-English was a real hodge-podge, not only of English and the languages of India, but of Persian, Turkish, Portuguese, Dutch, Chinese, etc. I expect to have a lot of fun.
Not to brag, but I got the Wordsworth reprint for half that.
$.005 / page? Hm.
But wait! Did your $4.15 include shipping?
Well, if you want to get technical I think it was $4.98 + tax (so probably around $5.40 total) at the Strand. But that was back in 1998.
Curses!
I’ll get you yet, Dr. Evil! You just wait. Humiliating me in fron of all those people.
Hat you’re wicked. Such deliciously methodical demolitions. I pity the fool who irks.
I recall a project by Edward Said and Daniel Barenboim called the West-Eastern Divan Orchestra, which I always thought was a very laid-back name.
In kannada, and I suspect other Indian languages, Dewan is also a sort of Minister. like the Dewan of Mysore. But I guess they have different roots.
I’ve long wondered what the hell a divan is. Now I’m still not sure I know, but at least I have sort of a clue what they mean when book characters are sitting on them. . .
The quality of the print and binding of my Munshiram Manoharlal reprint of Hobson-Jobson (1st ed. 1902; my printed one 2000; ISBN 81-215-0109-1) leaves nothing to be desired. When quoting on e.g. language sites, however, I cut and paste from http://dsal.uchicago.edu/dictionaries/hobsonjobson
That still leaves me needing help with the etymology British English “div” (= “idiot”).
There’s also a type of American casserole that is, for some inexplicable reason, called Chicken Divan:
http://www.cooks.com/rec/search/0,1-00,chicken_divan,FF.html
I believe it was popular back in the ’50s and ’60s.
Isn’t divan a sofa or couch in Russian? As for chicken divan, despite being born and raised in the US, I had never heard of it until a couple of years ago. And casserole is pot or saucepan (smaller pot)in French. And en español, es caserola.
John: Well, there are those who would call us both fools for spending any money when the whole thing is available online for free. But they don’t understand the joy of flipping actual pages, do they?
Abdul-Walid: “Irks” is the word. I mean, couldn’t she have outed me as, say, Adam Gopnik? Or Ammiel Alcalay? Why DFW?
dinesh: No, it’s the same word; that meaning comes under the OED’s “dewan” entry:
In India: a. The head financial minister or treasurer of a state under former Muslim governments. b. The prime minister of a native state. c. The chief native officer of certain Government establishments, such as the Mint. d. In Bengal, a native servant in charge of the affairs of a house of business or a large domestic establishment, a steward.
trevor: The OED hasn’t caught up with that one; Cassell says the etymology is unknown.
Laura: I suspect the “divan” in the name of the dish is the Indian ‘prime minister’ meaning; at any rate, it’s clearly supposed to carry a whiff of the high council chambers of the East, where men in robes and turbans eat the very finest delicacies.
Toby: Yes, the dictionaries give a word divan ‘sofa, couch’ and another ‘1. Turkish council of state, 2. collection of lyric poetry’; they are of course identical in origin, but are felt to be different words by contemporary speakers — as are the corresponding senses in English, for that matter. I think it would make more sense to list them separately in English dictionaries as well. (Joining them is particularly pernicious in the case of Merriam-Websters, where the common contemporary meaning is buried among the historical/exotic ones.)
At least, Goethe (Ge De 歌德, what a great name for a poet)’s title has been correctly translated in the sinophone world: Xidong shiji 西東詩集 (“West-Eastern Poem Collection”) and not, say, Xidong shafa (sofa).
Jimmy — shouldn’t it be 東西詩集: “Thing Poem Colection”?
Ha, you cannot imagine how much self-control I had to use to resist writing that one. As a matter of fact, many “copy ‘n’ paste” Chinese sites do misquote the title that way (Dongxi shiji).
Self control is bad, Jimmy.
I just remembered that Kant’s Chinese name also includes the 德 de/ German / virtue graph: some form of Kangde.
Ralph Waldo Emerson is Ai-Mo, “loves solitude” as I remember, but my Chinese teacher told me in no uncertain terms that that name is particular to him, and not a family name I could use. I became Ai “mugwort” (not as bad as it sounds, the Chinese honor mugwort more than we do).
I didn’t get along with that teacher at all, significantly damaging my already-slender career prospects.
Speaking of divans, there is a Russian “folk” song called “Sidyel’ Vanya” which plays on the name Vanya and the word “divan” (if you will forgive my rotten Russian transliteration — I do not know what the standard is — and inability to enter Cyrillic):
Sidyel’ Vanya na divanye (Vanya sat on the divan),
Stakan romu nalival (And poured a glass of rum);
Ne nalivshi polstakana (He had not yet filled the glass),
Sam za Katyenkoi poslal (when he called his Katyenka to him).
I believe the tune was adapted into a string quartet by one of the 19th century Russian composers. I was quite startled to hear it at a wedding reception some years ago. I’m not sure just how “folk” it is as the setting I know is highly choralized. You can hear it as sung by my old choir (I miss Chicago) at the Golosa.org web site. The group is the Russian folk choir of the University of Chicago — check out other tracks for our Semeiskie style improvised polyphony.
John: Are you familiar with this comprehensive page on mugwort and related plants? I regret to inform you that Ukrainian chornobyl’ [чорнобиль], and Russian chernobyl’ [чернобыль] mean ‘mugwort.’
xiaolongnu: Great to see you! That composer was Tchaikovsky; he used the tune for the slow movement of his first string quartet (you can hear a 30-second snippet here (mp3)). The text in Cyrillic is:
Сидел Ваня на диване,
Стакан рому наливал.
Не наливши полстакана,
Сам за Катенькой послал.
Yes, John, Kant’s name is Kang De 康德 and, as you rightly noted, that is the de in De[yizhi]guo 德[意志]國 (Deutschland / Germany). Unsurprisingly, the same character is frequently used to transcribe foreign names (Freud = Fuluoyide 弗洛伊德; Derrida = Delida 德理達, Alain Delon = Yalan Delun 亞蘭德倫, etc.).
I guess your famous homonym’s mo is the one in desert (shamo 沙漠, poetic root to French “chameau” = camel). Why not choose the mo for “ink” 墨 (but maybe you don’t want to be associated with mohist ideology)? Seriously, I’ll have to check what a “mugwort” is.
That is a great multilingual page. So it is that ai 艾, which relates you to modern writer Ai Qing 艾情, aka Artemisius Affectus (I had mistakenly assumed that ‘mugwort’ referred to the ‘mo’ part). Not bad at all, if you ask me.
Jimmy, Ai Qing is also the father of two important contemporary Chinese artists, Ai Weiwei 艾未未 (a founding member of the Stars group) and Ai Xuan 艾轩 (a realist painter specializing in Tibetan subjects). I just lectured on these guys in class today.
Thanks, xiaolongnü. I have probably heard of them, but frankly, I didn’t remember.
In Polish, a diwan (w=v) is a carpet.
Huh — that’s weird. (But according to my dictionary it’s dywan, with a y.) In Czech and South Slavic it has the normal range of meanings. I wonder how it wandered to the floor in Poland.
Thanks for the Russian folksong or whatever. I was glad to be able to appreciate the Vanya and na divanye alliteration. How cool is Russian, like Spanish, where the subject and predicate can be inverted.
And divan = carpet in Polish!!! Just last night I subbed in a beginning adult ESL class full of Salvadorans. There was one lone Polish guy in it. It was a treat to hear his English. And yes, he’s learing Spanish.
Yes mr. hat, you’re right, dywan. I was (not)preparing for classes when I wrote that and in a hurry/frazzled.
I have no idea how dywan ended up as carpet, but it might (wild guessing here) have something to do with designs, that is if your classic divan had woven designs on it …
William Cowper, of course, wrote a long poem called The Task, which was all about a sofa. I wonder if it has ever been translated into Persian.
Was it inspired by Crébillon fils’s Le sopha (where the narrator is a man magically changed into a sofa, who reports the words said and actions perpetrated on and around him by unwary “Arabian” princesses and their galants)?
Jeez, that sounds gross. He’d be talking about butts all the time.
I suppose Crebillon managed to evade that part. Forget I said anything.
I haven’t read much of The Task, but it certainly isn’t based on Le sopha, which I have read but don’t remember too much about. Crébillon was a favourite novelist of the Marquise de Merteuil in Les Liaisons Dangereuses as well as Becky Sharp in Vanity Fair, so he sounded promising, since I like a lot of eighteenth century fiction, but he didn’t seem all that striking to me when I read a couple of his things. I remember it was very “French”, with more cerebration than action (“just exactly what subtle nuance of feeling am I experiencing at this moment?”). There were also quite a few “post-modern” comments on the art of story-telling, which were quite funny. Crébillon also wrote a book called L’Ecumoire, about a Japanese prince whose genitals are transformed into a skimmer. God knows what that’s like.
!
Perhaps the Polish dywan=carpet comes from the fact that most Middle Eastern cultures don’t use chairs, instead opting to sit on the ground–suitably covered with carpets or rugs?
But Poland is by no stretch of the imagination Middle Eastern.
Maybe divans were covered with carpets or something carpet like?
“Poland is by no stretch of the imagination Middle Eastern”
There is however a long history of relatively friendly (for the region) contacts with middle east cultures and there was a turkic minority in Poland until sometime in the late 19th century (their descendents are still around, they just don’t speak a turkic language anymore).
True, and if the Turks used divan to mean ‘carpet,’ it would make sense for the Poles to. But they don’t, and the Poles don’t sit on carpeted floors.
As I understand, the Uhlans write Polish with Arabic script.
Out of curiosity, I checked Anthimos Papadopoulos’s old Pontic dictionary.
I didn’t find anything close to divan / div’ani (it does not mean there isn’t; his use of Greek for the transcription of Pontic words is pretty problematic), but i sof’a (note the gender switch: it is o sof’as in standard Greek) is first defined as “a surelevated part of the room” where people gather and sit together (I always imagine something similar to the Chinese kang, but I may be wrong), undoubtedly covered with carpets.
The second definition is “a wooden bed”.
Inspired by you, I checked Papathanasopoulou’s Rumeliot dictionary and found:
ntivani to = malako krevati, exostes.
‘Soft bed’ makes sense, but ‘balcony’??
I’m surprised as well. Then again, I know close to nothing about Roumeliotic. Meanwhile, I checked with an older Pontic, and it seems that “divan-” was indeed not attested in the patrida.
(Surprised about the definition as “exostis”, I mean.)
what do you think, if i give the name “DIVAN” for my baby, do you agree?
Fun to be a part of such a discussion, although I find myself not in the best position. Glad to hear that somebody’s reading our student newspaper published in Georgian Journalism School. I’m not sure that you were able to find it on the net, or whether it was at all on the net, but in the print version of the paper we printed apologies for the mistake regarding the title of the book. Unfortunately my editor didn’t appear to be a Goethe or Persian language fan and didn’t trust my English either, so that’s how collection of poems turned into sofa. My mistake was not to include in the article author’s comment regarding the title, stating that the book has nothing to do with Goethe, but it means what it is all about – in other words, it is a collection of poems translated from western and eastern authors.
As for Wallace, isn’t it an American author? I haven’t read, but I believe this shouldn’t be you:) Well, accept my apologies, but as you probably know, there are several people involved in coding the material for the research. Even though I’m not sure I remember coding your blog by myself, I don’t want to blame any of my coders for that mistake and won’t go back to find out who really took Wallace for the blog author’s name, but I accept responsibility and apologise for that. Sorry for late comment, I just came across your blog accidentally today, so…
Anyway, I’m glad to hear that my english (which is my third language) is becoming an issue for discussions on your blog – that must mean something:)
Hi, Ana! Your English is great, and I’m glad it turned out to be an editorial mistake rather than yours (and I apologize on behalf of my profession, which is editing). Yes, Wallace is an American author, and quite a decent one, but I didn’t like a long article he wrote on grammar and attacked it, so I was pretending to be insulted that he was listed as author of my blog. I was actually highly amused. And I’m glad you found your way to my blog; if I get back to studying Georgian, I may ask you for help!
Thank you:) Will take into consideration and get a copy of Wallace’s article on grammar:) I guess there’s too much to edit on the net and it must keep you pretty busy, but if you ever get back to studying Georgian, I hope I can be of some help.
In European French ‘divan’ is a word that means “couch” while Quebecois French prefers ‘sofa’. e.g.
Le chat dort sur le divan.
Le chat dort sur le sofa.
“The cat is sleeping on the couch”.
A synonym for ‘divan’ in European French seems to be ‘canapé.’
Brian: The cat is sleeping on the couch: couch does not mean the same everywhere in the English-speaking word.
A synonym for ‘divan’ in European French seems to be ‘canapé.’
Didn’t we discuss le divan just a few days ago? I remember defining the French word: a bed without a headboard and footboard. Un canapé is not the same as un divan since it has at least a backrest (sometimes replaced by cushions), and usually also armrests. It is actually what is called in Canada (in French and English) “sofa”. The French equivalent of a “sofa bed” is un canapé-lit, a convertible piece of furniture. You don’t normally need to say un divan-lit since a divan, from its shape and design, can serve as a bed as it is, all you need to add are bed linens and pillows. Un divan can also do double duty as a sofa if you add a few cushions and bolsters.
Interestingly, if you do a Google image search of ‘divan’, you’ll find a lot of sofas depicted.
The OED lists only one English word of Sumerian origin (it doesn’t bring the chiton/tunic doublet past its immediate sources): petasi ‘Sumerian priest/king’. What’s odd about that is that the reading is now understood to be enki, for whatever reason spelled PE.TA.SI. I searched AHD4 for more Sumerian words, but none.
Maybe purim, from pur ‘lot’, from Persian, thence from Akkadian pūru ‘lot’, homophonous with (and maybe deriving from) pūru ‘bowl’, a Sumerian loanword.
The OED lists only one English word of Sumerian origin (it doesn’t bring the chiton/tunic doublet past its immediate sources): petasi ‘Sumerian priest/king’.
What edition are you using? I just did an advanced search of the online 3rd ed. (which would have found it even in etymologies, let alone as a lemma) and got “No results found for ‘petasi’.”
Sorry, it should be patesi.
Thanks. An odd and unfortunate business:
Couldn’t we just sweep this misbegotten word under the rug and pretend it never happened?
There is also “mina” – ancient unit of weight (and also a monetary unit) equal to 1/60 of talent – a bit less than one pound.
It’s first mentioned in Sumerian texts as “mana”.
There is also “mina” – ancient unit of weight (and also a monetary unit) equal to 1/60 of talent – a bit less than one pound.
It’s first mentioned in Sumerian texts as “mana”.
Not to get too Nostratic about it, but moon, month, et al, attested in many Germanic languages, also indicate a kind of measure. As does Sanskrit mimīte, he measures, Hebrew מנה mana, he counted (and a host of derivatives), plus cognates in other Semitic languages.
Aren’t the first efforts at writing generally taken to be signs for numbers, quantities, etc., as needed for trade?
petasi, patesi, petasi, patesi, let’s call the whole thing off!
From the Zompist spelling reform page:
Pronunciations include “GINK-o” (AmE, BrE) and “GINK-go” (BrE).
That’s where the paper comes in that I found yesterday. Which thread was that again…
This paper, linked to from this Wikipedia article.
Cross-linking the Sumerian post.
That still leaves me needing help with the etymology British English “div” (= “idiot”).
The OED Third Edition (new entry, June 2022) has:
And divvy n² is:
I think that’s the first emoji I’ve seen in the OED. And this is interesting:
Etymology: Apparently < Romani diviō, diviou wild, mad (apparently < a Slavonic language; compare Slovenian divji, Serbian and Croatian divlji, Bulgarian div, Czech divý, all in the sense ‘wild’), probably with remodelling of the second syllable after -y suffix¹.
Emoji began appearing in OED quotations in the June 2019 update. In the blog post No tears of joy (yet): emoji make their OED debut, they describe some of the quotations and why they were chosen. I particularly like this citation for inverse:
Individual emoji are searchable; according to their search function, that’s the first quotation they’ve added that contains
.
John Cowan: The OED lists only one English word of Sumerian origin
Only one of *direct* Sumerian origin. Others that it gives as ultimately from (or speculated to be from) Sumerian:
abracadabra (revised 2009)
ass, n.1 (revised 2018)
get, n.3 (Jewish bill of divorce; revised 2016)
Nisan (Jewish month; revised 2003)
zizania and †zizany (a weed or aquatic grass; revised 2021)
I searched AHD4 for more Sumerian words, but none.
You missed mudra, shofar, and Tammuz; AHD5 added saros (eclipse cycle). These haven’t been revised yet in the OED, except mudra, where they don’t take it beyond Sanskrit.
There’s a very recent borrowing directly from Sumerian: kunga, a domestic equine, in paleogenomic studies claimed to be “the earliest known human-engineered hybrid animal, predating the earliest mule by about 1500 years.”
Wonderful, thanks for all that Sumeriana!
There’s a very recent borrowing directly from Sumerian: kunga, a domestic equine
Another fine addition to my list of five-letter English words that aren’t in Wordle. (Now up to 27 entries, ranging from abjad to yoink, though for some of them it’s debatable whether they legitimately qualify as English words.)
There are at least two other words entered in the OED that are ultimately of Sumerian origin, but the OED1 and OED2 etymologies do not dig down far enough to uncover their Mesopotamian sources. One is haikal.
Even if this word is actually used in later Coptic writings for the central part of the sanctuary, it seems that it is simply a borrowing of Arabic هيكل haykal ‘temple, altar’, ultimately from Sumerian 𒂍𒃲 e₂-gal ‘temple’ (literally, ‘big house’). The Wiktionary has serviceable entries for the etymology, with lists of descendants for Arabic here and Akkadian here. (I wonder what the echt Coptic equivalent of haikal would be.)
Another is duka, proximately from Swahili:
The Wiktionary entry for دكان dukkān is serviceable enough here, taking it back to Akkadian takkannu, dakkannu and Sumerian daggan (‘small chamber or cell’? also ‘doorway or door part’?), with a list of descendants of this well-travelled word. There is a philological study of the semantics of Akkadian takkannu beginning on page 101 in Paul-Alain Beaulieu, ‘New Light on Secret Knowledge in Late Babylonian Culture’, Zeitschrift für Assyriologie und Vorderasiatische Archäologie 82 (1992), available here. (As far as I have been able to discover, Sumerian daggan is not attested before the Lagash and Ur III periods. Sumerology is not my bailiwick (I have just bare minimum required for Hittitology and Semitic philology), but most sources, like the Chicago Assyrian Dictionary, indicate that Akkadian takkannu is a loanword from Sumerian.)
I wonder why دكان dukkān eventually shows up in Russian as духан with х. Since the word has a palatalized k in the Turkic languages (e.g. Azeri dükan, Crimean Tatar tükân, Turkish dükkân, etc.), I wouldn’t expect any spirantization within Turkic itself. I wonder if there has been crossing at some point in transmission with Persian خان khān ‘caravanserai, inn’ (Ottoman خان khān, Azeri xan, Greek χάνι, etc.)? (Or crossing with Persian دخان dukhān, Ottoman dükhān ‘smoke (esp. tobacco smoke), tobacco’, etc. (all from Arabic دخان duḫān, duḫḫān ‘smoke’), little taverns in the Caucasus being smoky places… Or perhaps there was influence from Russian дух as ‘smell’, with derivative душный…)
Great stuff, thanks!
In 80s during the Afghan war Russian soldiers called their enemies духи (spirits/ghosts/spectres from official “dushman”)
Then in 90s they called Chechens чехи (sounds identical to Czechs).
I wonder if it is normal for Persian official rhetoric to use “enemy” as a slur for the enemy. Or does the word has extra connotations that rather abstract European “enemy” does not have? Or it were Russians who mistook it for some sort of a name? It was a convenient word: it recembled душегуб (“villain”, lit. soul-killer) and душитель …
Yes, thanks very much, Xerîb! Haikal was entered in the OED’s 1933 Supplement and carried through in the 1976 Supplement, but investigating etymology beyond the immediate source of a loanword wasn’t in the remit of those supplements. This will probably get a full etymology when it’s revised, considering that the path from more recent Semitic languages back through Akkadian to Sumerian was already well enough known to be given in Merriam-Webster’s Third Unabridged in 1961 (it remains unchanged in the online version).
Duka, on the other hand, is freshly revised in the East-African-English-focused update of June 2022, so it’s too bad they didn’t take the opportunity to go beyond Arabic. It’s especially interesting that it’s been loaned via Arabic and Turkish into so many other languages. That’s the kind of historical context that ideally I want to see in the OED: East African English hasn’t just picked up a local word, it’s hooked into a large trading network with a long history.
Another word in the OED with an etymology that could be pushed back to Sumerian is cor :
As here, Ezekiel 45:14 :
Hebrew כֹּר kōr, from Akkadian kurru, from Sumerian gur 𒄥.
Greek κόρος also at Luke 16:7, but the KJV doesn’t render this with cor :
According to the ePSD, duggan is attested at least at early as Early Dynastic Lagaš. Thanks to the magic of the internet, there’s even a high-quality image of the first attestation of duggan readily available from the CDLI.
Niek Veldhuis discusses the use of duggan bags in the context of Ur III Gu’abba (somewhere in the centuries on either side of 2000 BC). Several clay tags have been found that summarize the contents of duggan, whose inscriptions show that the duggan were used to transport daily administrative records to the central archives.
That said, “duggan” is an odd shape for a Sumerian noun. Geminates are especially unsual, but Akkadian glosses are pretty insistent on that geminate “kk/gg.” But there’s a lot we don’t know about Sumerian phonology…
Also, cumin may or may not be Sumerian (Sum. “gamun,” “kumul,” and/or “gumul”). It’s either that or Akkadian (kamûnu). Not sure if there’s evidence for a proto-Semitic reconstruction or if it spread from Sumerian to the rest of the Semitic languages through Akkadian—wanderwort is always an option too.
Noonan on cumin:
In Siwi and a couple of other Berber languages, “cumin” is yamǝn. The y could be from *k’ under appropriate assumptions about the environment, but it clearly does not derive from Arabic kammūn, and I don’t see a convincing way of getting it from the Greek form either even though it’s closer. Guess there’s a missing intermediate somewhere…
In Mishnaic Hebrew דּוּכָן dūḵān is a raised area, like a podium, and in later (only Israeli?) Hebrew also ‘bench’, ‘market stall’ or such. Klein thinks the latter and the former are connected to each other and to the root דוך dwk ‘pound, flatten’.
Just to clarify for LH readers, the word that is proposed to show up Swahili duka via Aramaic דוּכָן dukkān, Arabic دكان dukkān, Akkadian dakkannu(m), etc., is Sumerian daggan, spelled 𒆠𒍇, etc. Akkadian dakkannu is entered by the Chicago Assyrian Dictionary under takkannu (“chamber, niche, bench”), I suppose because most of the spellings use the ambiguous CVC sign 𒁖, which can be tag-, tak-, or taq-, or dag-, dak-, or daq- according to context here. The descendants require suggest reading dak-, which is how Black’s Concise Akkadian Dictionary and von Soden’s Akkadisches Handwörterbuch lists it. Von Soden also gives an example of a spelling da-ak-ka-nu-um, I see.
I wrote the following, “most sources, like the Chicago Assyrian Dictionary, indicate that Akkadian takkannu is a loanword from Sumerian”, because Sumerian daggan seems relatively late and it has an uncharacteristic geminate, as ə de vivre notes. Gutian? But why should a word for a kind of room be from the language of a nomadic people?
Oops, wrong dVkkan word
. Oh well, I guess we learned something about 3rd millennium bags for nothing.
Sorry… “show up as Swahili duka“.
Oops, wrong dVkkan word
. Oh well, I guess we learned something about 3rd millennium bags for nothing.
We can always stand to learn more about Mesopotamian bags, a topic which has been explored before at LH!
About x in Russian: this k is aspirated in Caucasian languages (whose aspirated-voice-ejective rows are so reminiscent of proto-Semites)…
But I can’t remember a [kʰ > x] loan from Caucasian to Russian.
@drasvi, Xerib
I was going to try to answer Xerib’s question about when Turkic k or q becomes Russian k or x in loanwords, and whether this correlates with the sound in any corresponding Iranian etymon, but I was unable to conclude anything more than “sometimes it becomes k, sometimes x (sometimes there are even doublets); where I can find a corresponding Iranian etymon, this is more likely to be the letter g in Latin transliteration(GAMMA?).” Are Turkic loanwords through Caucasian intermediary typical?
Couldn’t the k/x alternation be a function of the Turkic language that the loan word filtered through? Some Turkic languages weaken velar stops between back vowels, so depending on the given Turkic languages lentition processes and the specific vowel environment, you’d get either [k] or [x] in the most proximate source of the word?
“I wonder what the echt Coptic equivalent of haikal would be…”
Burmester, in his ‘The Egyptian or Coptic Church; a detailed description of her liturgical services and the rites and ceremonies observed in the administration of her sacraments’, gives ‘erphei’ (in Bohairic. Sahidic ‘rpe’, the normal word for temple) as the Coptic equivalent of ‘haikal’ when presenting various architectural terms at the beginning of the book, but without further discussion. I also read a comment in some article I found online (I can’t seem to find it again now) stating that the first use of ‘haikal’ to describe this part of a Coptic church is found in a work called The History of the Patriarchs of Alexandria (written in Arabic), but that earlier terms used were ‘iradyun’ (Gr. ἱερατεῖον ) and ‘askina’ (Gr. σκηνή).
The etymologically literal morpheme-by-morpheme equivalent of haikal “great house” in Coptic is actually ⲡⲣⲣⲟ “the king” (= “Pharaoh”), which in Coptic has got reanalysed as the masculine singular definite article ⲡ + ⲣⲣⲟ.
gives ‘erphei’ (in Bohairic. Sahidic ‘rpe’, the normal word for temple)
earlier terms used were ‘iradyun’ (Gr. ἱερατεῖον) and ‘askina’ (Gr. σκηνή)
Excellent! Thank you Andy for ferreting these out!
With Bohairic ⲉⲣⲫⲉⲓ, Sahidic ⲣⲡⲉ ‘temple’, from Egyptian r(ꜣ).pr, we circle back to the notion of the temple being simply the “house of the god” again, like Sumerian 𒂍 É ‘house’ in 𒂍𒃲 É.GAL ‘palace’ (> haikal) above. I looked briefly to try to find out what the original semantic contribution of the r element in r.pr was, but it seems that there is little consensus (as on p. 5 here). Maybe someone can direct us to a good study of the term r.pr.
Whence the Egyptian surname Heikal? Is it associated with any particular community?
(Like this former editor of Al-Ahram, or this egyptologist.)
When I was in Budva (a town in Montenegro that has more Russians than Serbs/Montenegrins) some Russian gallerist organized an exhibition of a certain artist in a tonnel in a promontory between two beaches. The artist illustrated Serbian alphebet. I can’t recommend the artist (I am all for drawings that look like children drawings – but it does not mean that I like all of them:) Same is true for more classical art) but anyway, a dukhan in Tbilisi by the same artist, illusrating letter რ.
a dukhan in Tbilisi
This got me thinking that maybe Russian духан is a crossing of Georgian დუკანი dukʰani with old-fashioned Georgian ხანა xana ‘house’, from Persian خانه xāna, as in name of the institution of the میخانه mayxāna ‘tavern’ (literally, ‘wine house’), Azeri meyxanə, Turkish meyhane, etc. (cf. also Georgian დუქანხანა dukʰanxana ‘inn with rooms for rent’). Or a crossing of Azeri dükan /dycɑn/ with xanə /xɑnæ/ ‘house’ in the process of transmission to Russian?
Azeri being the lingua franca of pre-Soviet Transcaucasia, what about Kumyk, the Turkic language that was the lingua franca of Ciscaucasia before the Soviet deportation of the Kumyk population? According to this fine site, the word shows up as тюкен, which I believe is to be interpreted as /dʉwken/ or /dʉβ̞ken/. I haven’t come across any descriptions of Kumyk indicating that the /k/ (before a front vowel here!) would be spirantized here, but then, I haven’t looked very much.
Maybe there were Azeri dialects spirantizing /c/ (ک , modern written Azerbaijani k) intervocalically to [ç] within a word, but I haven’t heard any. Question: Would Russians even have assigned such a [ç] to their /x/ phoneme? On the surface, this seems to be the case with German loanwords with the ich-laut ch [ç] in Standard German: архив, кирха, кнехт, рейх, фенхель, maybe пихта (Fichte, if it is not rather a Finnic form like Finnish pihka ‘resin, pitch, gum’, also with [ç], presumably); an interesting case is цейхгауз (Zeughaus). But in these, maybe the Russian outcome shows the influence of the orthography or German varieties where ch is [x] ~ [χ] everywhere?
-xâna is well known in Modern Russian,
xân “inn” is not (though خوان xân is).
But it does not matter, what matters is how popular the word was in the Caucasus back then….
Perhaps xân was not so popular (was not the default for ‘inn’) if it did not enter Russian. And perhaps “khan” the title blocked it?
I am now curious how exactly this pair xân “house” – xāna(g) “house” could have arisen in Persian:/
Sorry, cutting and pasting error… The Kumyk is presumably /tʉwken/ or /tʉβ̞ken/ for тюкен.
“Would Russians even have assigned such a [ç] to their /x/ phoneme?”
I need to hear this ç.
We have palatalised x (forvo.com хи-хи, хихикать, and before /e/ хер).
But хя is not too comfortable… If Arabic distinguished between x and χ as it distinguishes between k and q, Arabists would have adopted хя for xa and хы for χi (just like they scare the native speaker with кясра “kasra” and кыбла “qibla”). But it has x, gh, 7, 3 and h instead. 3a-3a-3a.
Also Russian has [ʝ] as a possible pronunciation of /j/, and when it is devoiced it becomes a voiceless fricative. When I am dealing with a language with a ç (distinct from both Russian sounds…), it is one of those situations when I want more IPA signs:(
So,
“do Russians hear ç as /xʲ /?” – depends on ç, but at least sometimes yes.
“would Russians render what they hear as (not too comfortable) /хян/ as /хан/?” – I don’t know… But it seems possible.
an interesting case is цейхгауз (Zeughaus)
That reflects the pronunciation of final “g” as [x / ç] that is usual in many German dialects (what is taught as Standard pronunciation has this only in the suffix -ig) plus the traditional rendering of German [h] as Russian “г”.
Nişanyan’s entry for dükkan cites an entry in the Codex Cumanicus, “apotheca [dükkân] – Fa: duχan – Tr: tugan,” that has a fricative in the Persian (Farsça) form of the word, suggesting that Persian dialects might be the source of the alternation.
Wild mountain Jewess. Pleasant guttural voice. Tattooed faces, Red Fingernalis .
GT:
“Even the very name of the owner of the dukhan where we landed, Beniogu, breathed something patriarchal, Semito-Arabic, and did not at all resemble the current Itzkas and Sruls, just as the Caucasian mountain Jews do not at all resemble their civilized counterparts in Europe, being a tribe in the highest degree attractive…
…..
From the door of the sakli came a girl who, rightly, could be mistaken for a fairy. If Heine were here, if this charming vision caught his eye, we would have a charming mountain legend. I myself felt ashamed that I had opened my mouth to a beautiful savage. Imagine a narrow oval face, thin and graceful. Large black eyes with tonsils look at you somehow timidly and submissively. This is the look of an oriental woman. Black eyebrows seem to be slightly pointed with a brush – their bend is so correct; a graceful nose with thin pink nostrils that swell slightly even from ordinary breathing, and a small mouth; slightly swollen bright scarlet lips, the upper one slightly upturned, not ugly, but just enough to show the small pearls of the teeth. And although her hair was hidden by an ugly silk bag behind, the strands of them were knocked out over a small forehead and intrusive, small curls framed thin pink, see-through ears. Dissolve two or three drops of blood over the matte, passionate dark skin of this face, enough to lightly color only the cheeks, draw his barely noticeable blue lines of veins over the finished sketch – and you will understand how amazed I was at first minute. A yellow silk shirt with wide sleeves, intercepted at the waist, made it possible to see the sloping shoulders and waist of the dragonfly, which at once almost turned into broad, luxurious lines of strongly developed hips. The narrow, beautiful hand was so perfectly fine that it became involuntarily annoyed at the blue warts of Persian turquoise, in the form of rings, sitting on the longish fingers of the Jewess …
She spoke in a guttural, pleasant voice. I shook my head, of course. A mocking smile flickered across her face for an instant and disappeared… And again those expressive eyes look timidly and submissively.
Alas! The fairy of the Kaibulag gorge we met was the first and last beauty among the Jewish population of Dagestan. Maybe I didn’t come across others, but as a conscientious tourist, I must tell you that all the other ladies of militant Israel were very unattractive. Between them, moreover, come across real tattooed savages. The already rough faces are still painted with different figures, lines, circles, triangles. The most disgusting thing is when these signs are pointed with bright red paint. Painted exactly in blood. I even happened to meet an old woman who did not deny herself the pleasure of drawing a few black lines across her face, which, in her opinion, gave her a special nobility and grace. Dappers with dyed eyebrows and hair are not uncommon. Lately, according to the Persian custom, local ladies in some auls put crimson colors on carefully grown nails. As if the ends of the fingers were soaked in blood. All these frivolous Evas are dressed very badly on weekdays; not to mention the poor, and the rich flaunt rags … Moreover, they are very dirty. An Englishman who measured the degree of civilization by the amount of soap consumed by a given country would have been horrified by the Mountain Jews.
— Makhlas, Makhlas! was heard from within the sakli.
The beauty spoke again in her guttural language, addressing the door, and walked past me, her eyes downcast.”
Why I quoted it:
– прекрасная дикарка “beautiful savage” is an unexpected description of a Jewess.
– guttural (and pleasant)
– I don’t know if he means they were actually tattooed or he just compared it to tattoo. But I did not know that women in the Caucasus did that.
Which reminds how little we know about the visual aspect of the past.
Greek temples and statues (naked marble today), Tunisians who draw pictures that I mentioned in another thread, wooden sculptures in Russian churches and headgear of Russian peasant women. And red triangles on faces of Caucasian Jewish women:-/
– Red nails. Savage!!!
—-
https://ru.wikisource.org/wiki/Израиль_воинствующий_(Немирович-Данченко)
The beauty spoke again in her guttural language
Irish or German, presumably.
P.S. of course it was гортанный – the adjective used to describe sounds like… ع .
Гортань (etym.)- an anatomical term, “larynx”
Горло – “throat”.
So it is not the same as “guttural” in that “guttural” is Latin and гортанный sounds native, but it is distributed very similarly.
“Large black eyes with tonsils….” I had to wonder what led to the translation “tonsils” there.
The Russian is “Крупные черные глаза миндалинами смотрят на вас как-то робко и покорно,” which DeepL renders “The large black amygdala eyes look at you somehow timid and submissive.” The word миндалина can mean ‘almond,’ ‘tonsil,’ ‘amygdala,’ or ‘geode’; I leave it to you to decide which is the most likely comparandum for the beautiful savage’s eyes.
Nişanyan’s entry for dükkan cites an entry in the Codex Cumanicus, “apotheca [dükkân] – Fa: duχan – Tr: tugan,” that has a fricative in the Persian (Farsça) form of the word
For readers who aren’t familiar with the Codex Cumanicus, part of it consists of a trilingual dictionary or glossary, with Latin words in the left column, the Persian words in the middle column, and Cuman words in the right column. It is written in the Latin alphabet. The page with duchan can be seen here. The relevant entry is the second line from the bottom of the page (in this instance, there are two Cuman terms):
In the Latin column, apotecha is Latin apotheca, of course. Note the inconsistent (to say the least) orthography throughout the different parts of the Codex Cumanicus.
In the Cuman column, chebit ‘shop, booth, stall’ is a Turkic word still continued today in Tatar кибет (dialectal кибит) and Bashkir кибет, ‘small shop’, for example, and also widely borrowed outside of Turkic, as in Mari кевыт ‘shop’, Udmurt кебит ‘smithy’, Middle Mongolian kebid ‘shop’. A reflex of the word can also be seen in Russian кибитка. (In much of the rest of Turkic, this word has been replaced by our word, Persian dukkān. The Turkic kebit itself is probably a loanword from Sogdian qpyδ ‘shop, stall’, from Greek καπηλεῖον.) As is apparent, the ch in chebit most likely spells a front k.
And tuga̅ (that is, tugan) is just our word again, Persian dukkān (variant dukān), Arabic dukkān (u̅l̅, scribal abbreviation for Latin vel, ‘or’).
It worthwhile comparing other uses of the digraph ch in the Persian column of this part of the Codex. Here is a selection of clearer cases that I made quickly scanning the columns in the first few pages above our entry:
In light of this, it seems to me that the easiest interpretation of the ch of duchan is /k/. I am going to write Sevan and suggest that he look into this entry again. I use his dictionary every day and am grateful that he offers this wonderful resource free to everyone, especially to the often impoverished university students of Turkey.
For readers who aren’t familiar with the Codex Cumanicus
Thanks on my own behalf and on behalf of the other two. (They know who they are.)
*raises hand furtively*
In my dream I saw Google that showed the order of Lenin (as when you enter one word and it offers a definition) with a text: “the Order of Lenin. Only sorcerers of the second rank can perform magic rites with it”.
I found it peculiar (for the same reason why I would find it peculiar now) and wanted to post a link here.
“…As is apparent, the ch in chebit most likely spells a front k….”
“…In light of this, it seems to me that the easiest interpretation of the ch of duchan is /k/. …”
Ch (Africha) is a common spelling. I wonder how it arose and how it was distributed. Presumable students of medieval romance know…
Why for example, affricha (the city), africha (the land), chastel (Castilia), mare ochceanum, mecha (Mecca), domasch, but cafsa (Gafsa, qafṣah), cumania and castels (castles) in the Catalan Atlas?
Google Translate and eyes:
It is singulative of миндаль “almond” (cf. картофель – картофелина “one potato”). In modern literary Russian uncommon, but colloquially I think one can form it.
—- “almond-shaped eyes” always confused me (particularly when I was a child and wanted to know what it means), because any eyes (human eyes) are almond shaped:/ Two arcs joined at two corners…
I guess the comparison originated in lands where almond grows.
Kennst du das Land, wo die Mandelbäume blühen?
His nose was heigh, his eyen were cytryne,
His lippes rounde, his colour was sangwyn
It seems Chinese has it too:
杏仁 – apricot kernel, almond
杏眼桃腮 – almond-shaped eyes and peach-red cheeks.
杏眼圓睜 – her almond eyes glared round with rage.
(some 19th century dictionary).
Wiktionary says 杏 is “apricot”. And apprarently untranslatable 圓睜 is best illustrated with google images.
On the Romani origin of divvy: as mentioned at Lake Talk, divia ‘crazy’ also appears in the Romani-influenced argot of Nonantum, Massachusetts.
I don’t know if he means they were actually tattooed
it’s hard to know! my understanding is that kurdish and amazigh jewish women traditionally did have similar tattoos to other kurdish and amazigh women, but that those practices have largely been wiped out as part of the general flattening-out of diasporic jewish cultures since european colonialism arrived in those regions (which has, of course, accelerated with the rise of zionism). but that might or might not be relevant, since “mountain jews” – which usually refers to juhuri ([judeo-]tat-speaking) communities these days – has been used to refer to all the caucasus jewish communities (juhuri, georgian, kurdish, and others) at various times*.
but it’s also quite clear that henna, indigo, and turmeric were used for temporary and semi-permanent skin decoration in those communities and others. i would expect juhuri “mountain jews” to be among them, as part of the greater persian cultural world. noam sienna’s henna-centered “eshkol hakofer” blog (moribund, afaik) is a great collection of materials on that, with some nice philological tidbits along the way.
and, of course, very little popular writing about any of this is reliable at all, since it almost invariably starts from the post-khurbn anti-tattooing stance of most european-descended jews and takes prohibitions on tattooing by maimonides and joseph karo as if they were universally accepted both in theory and in practice.
.
* i wonder whether it’s ever been applied to the subbotnik and other “judaizing” heterodox-/post-christian communities that wound up in the caucasus.
Tattooed Mountain Women and Spoonboxes of Daghestan (academia)
!!
[ç] and [xʲ] sound almost the same. If you find a language described as having one of these, always wonder if it’s the other.
[ç]-less varieties of German: in and around Switzerland, /x/ is [χ] with no allophony; on the middle Rhine, [ç] has shifted to [ɕ] or thereabouts and may have merged into /ʃ/ in some places.
Googly-eyed madness?
When I was 16 (or about that) I wondered if крутой would sound the same if it were крутохь (and experimented with my friends).
Both sounds are basically allophones of other phonemes.
Soft consonants (including хь) are considered independent phonemes, but velars almost never occur before -ы, (and гы хы кы are difficult to pronounse, ги хи ки are preferable) or before -я (гя хя кя are also not too convenient, га ха ка are preferable again) or as -гь -хь кь.
I suppose these issues with distribution make it somewhat less phonemic.
[ç] only arises from devoicing of one possible realisation of /j/, but it still arises in my speech sometimes.
So it’s a situation when two allophones (or almost allophones) approach each other in sound closely from two different directions (morphonological directions)… while still maintaining slightly different articulation. Weird.
Awesome.
(And a known phenomenon in historical linguistics.)
JC (2015): “OED … doesn’t bring the chiton/tunic doublet past its immediate sources”
OED revised tunic in September 2023 and now shows the link to chiton, though they’re still not willing to go beyond Semitic: “classical Latin tunica … probably a borrowing from a Semitic language (compare Hebrew keṯōneṯ), from a word also borrowed into ancient Greek as χιτών chiton n.” Chiton is yet to be revised.
The revision also points out a gap of several centuries between the rare attestations of tunic in Old English and its reappearance in the Early Modern period: “Originally < classical Latin … probably subsequently reborrowed” from French and/or Latin.
The OED splash page has a list of recently updated words, among them divan. (The entry should be accessible to everyone at least for a while.) Here is the relevant part of the etymology:
There is still no Elamite tippi, no Akkadian ṭuppum, no Sumerian DUB, just their mealy-mouthed ‘…same Indo-European base as…’ for the second element of the compound. The reader may well ask, ‘But what is the first element? Why are you being so mysterious about it?’. How much more straightforward the exposition would be (especially for the phonological changes), and how much more complete and interesting the story, if they just gave the reconstructed forms like *dipi-vahana- or *dipi-vān, and went from there! (Something like “…Old Persian *dipi- ‘inscription, document’ < Elamite tippi < Akkadian ṭuppum ‘tablet’ < Sumerian dub”.) I believe the etymology with *-vahana- ‘dwelling, house’ (actually attested in Old Persian āvahana- ‘settlement, village’ with preverb ā-) is from Erik Samuel Nyberg (1974) here, p. 64. A different account of the second member of the compound is preferred by François de Blois in an up-to-date (revised 2011) article on dīvān in the Encyclopædia Iranica here, based on H. Bailey (1933) ‘Iranian Studies II’, p. 76f, here on JSTOR.
Good lord, that’s pathetic. All the space they want, and that’s the best they can do? Thanks for bringing it here!
Also on the OED splash page is a list of recently added words: Boricua, carniceria, guyliner, etc. And also dalgona, ‘A Korean confection made by adding baking soda to melted sugar, typically sold by street vendors in the form of a flat disc with a simple shape such as a heart, star, etc., carved on its surface’. (This entry temporarily accesssible to all visitors as part of the new additions—the inclusion of dalgona is no doubt due to the dalgona coffee fad a few years ago and the TV series Squid Game.) The OED etymology:
Really? That’s it? Even I, with the minimal Korean of a mere dilettante, can spot the root dal- ‘sweet’ (citation form 달다 dalda ‘it is sweet, delicious’) in 달고나 dalgona. What is the 고나 -gona, then? I do not know, but five minutes of Googling and speculation have already been promising.
Simple Googling finds an explanation that it is an alteration of 달구나 dalguna ‘Wow, it’s sweet!’, ‘I didn’t know it was so sweet!’, ‘I wasn’t expecting it to be sweet!’. That looks like a good start. This 달구나 dalguna is a mirative of dalda, formed with the informal mirative ending -구나 -guna. What would explain the alteration to -gona then? I was thinking that it might be accomodation to Korean vowel harmony, in which some affective words like interjections and phonaesthetic adverbs have all vowels belonging to one or the other of two sets: the bright, light 陽 yang set (including a and o) and the dark, heavy 陰 yin (Korean eum) set (including eo and u). Changing dalguna to dalgona makes the word entirely yang and bright and positive. Maybe street vendor cries had something to do with it? Or is it simply a dialect pronunciation that that has been generalized in standard Korean? I don’t know at all. It will take some time to find out. But I wish the OED had done the work for me. One gets the feeling that etyma in European languages get a full treatment under the OED3 editorial plan, even back to ‘the same Indo-European base as…’, but non-European etyma… Coreanum est, non potest intellegi. Gotta churn out those entries to meet the 1st quarter goals! No time to write two or three of the hundreds of linguists and philologists working on Korean with a specific question and wait a few days or a week for an answer. (I know from experience, most scholars are delighted to receive such enquiries.)
Sad. I guess “meet the 1st quarter goals” is a universal imperative now.
More carping about recent OED etymologies… I just spent a bit of time looking some stuff up before going to sleep and I thought some of it might be of interest to LH readers, so here it is.
The publicity materials for the OED update include a note on the word virago. The note has link to the entry for virago, which should be accessible to everyone, at least for the moment as part of the publicity push. Here is the etymology given there:
For a second I was like, what? Hebrew his ‘man’ and hissa ‘woman’? What is this family of Hebrew hapax legomena that I have never heard of? Then I realized that the verse the OED is talking about is Genesis 2:23:
Now, I can understand that the OED might want to avoid quoting forms in the Hebrew script like אִישׁ and אִשָּׁה peppered with Masoretic pointing (although the OED gives ἀνήρ and ἄνδρις untransliterated in this very entry…), and even to avoid scholarly transliterations like ʾîš and ʾiššāh. But if the concern is accessibility to the general reader, why not use romanizations like ish and ishah or ishshah, or something like that? Where did these perverse spellings his and hissa come from? A little searching reveals the following passage (available on archive.org here) from the Venerable Bede, commenting on the Vulgate version of Gen 2:23 (with my quick expedient translation of the medieval Latin):
The etymology of Israel that Bede relates is not etymologically correct. However, this interpretation of the name was current in later antiquity, and Jerome knew it (Liber de nominibus hebraicis), and it helps put Bede’s comment in context. In any case, is this passage the place where the editors of this OED etymology got the ‘Hebrew’ forms his and hissa? I venerate the Venerable Bede as much as anyone, but nowadays, most of us have access to better Hebrew lexicographical resources than could be found at Jarrow in the 8th century, and OED editors could have availed themselves of such… The question of how much knowledge Western European Christian scholars had of Hebrew in the medieval period , and how they pronounced the language, is interesting in itself, but does this justify the OED in citing the Hebrew words as his and hissa? If you tried to discuss the OED etymology with your average educated speaker of Israeli Hebrew or with your average Biblical scholar working as an instructor in a university, I suspect they would be baffled by his and hissa at first, until the penny dropped: ‘Oh, they mean ish and ishah!’
(In fact, Hebrew ʾîš is not originally etymologically related to ʾiššāh, however much native speakers of Hebrew—including the composers of Genesis—may have felt them to be so synchronically. A bit like English male and female. Maybe the OED etymology should also note this?)
Maybe there are also some other medieval sources for this ‘Hebrew’ his and hissa besides Bede. I leave it to other LH readers to find them if interested in the topic.
Good lord, this is getting worse and worse. OED, we hardly knew ye!
So you’ve finally descended to the level of TV Tropes. We Hardly Knew Ye [Characters that die, get eaten, go on a long trip, spontaneously vanish, or otherwise exit the show — and all before we get to properly know them.] Yes, I know it has been put to music.
I would imagine that the h was in fact silent in Bede’s his. His attempted etymologising of “Israhel” only makes sense on the assumption that in his Latin, h was at most only a sign of hiatus. As for everyone else who was writing Latin at that time.
In the more-or-less contemporary Old Irish orthography, written h is everywhere silent except in digraphs, even though the language actually had a /h/. It tends to get added to vowel-initial words that would otherwise have been very short.
The Old Welsh “Surrexit” memorandum seems to have h in the right places for /h/, (“and” may have actually been /ha/, as in Cornish):
https://en.wikipedia.org/wiki/Old_Welsh
but the words involved are all pretty short. Dunno about other Old Welsh sources.
“words that would otherwise have been very short.”
Irish:)))))
The “Computus” fragment also seems to have h in the right places for /h/ pretty often, though it’s not easy to decode:
http://www.ancienttexts.org/library/celtic/ctexts/computus.html
On the one hand, I see hanaud for anodd “difficult” in the very first line. But ys “is”, yn “in” and y(r) “the” seem to be always written is, in, ir, short words though they certainly are.
If Old Welsh orthography did (mostly) use h for /h/, this would be a point against the argument that I was recently making, to the effect that Old Irish spelling was more likely to reflect Old Welsh/Brythonic spelling conventions than an idiosyncratic British pronunciation of Church Latin. The Computus fragment is quite a bit later than either Bede or the creation of Irish orthography, though.
h for [ʔ] does sometimes show up in Old English poetry; evidently, some scribes found it necessary to write the alliterating consonant even by desperate means.
(Hypercorrection for aitch-dropping is not an option – [h] remained stable for a few more centuries.)
I suspect “[I/we] hardly knew you” is nowadays more often used to mean A “never got to know you properly” rather than B “barely recognise you now that you have changed so much”. The old song was B; I guess A was originally a kind of pun or jocular reinterpretation of B which has gained autonomy.
Stumbled across this when looking for Old Welsh sources online:
https://www.repository.cam.ac.uk/bitstream/handle/1810/252838/Russell%202015%20Studia%20Celtica.pdf
It doesn’t actually bear on the h issue, as it was in fact composed very late in the Old Welsh period (about contemporary with the compilation of the Four Branches, in fact, which we actually have in Middle Welsh orthography. The division between Old and Middle Welsh is as much orthographic as anything.)
But it’s interesting in that the text was overwritten by successive Middle-Welsh speaking scribes to make it more comprehensible (or perhaps, just better adapted to be recited aloud), and you can still see where they’ve done it in many cases.
Old Welsh “and” definitely was in fact ha(c), too. I’d forgotten. So the examples I cited above in Old Welsh texts where it’s written that way are all cases of h for /h/ too. (The revisers of this document have actually sometimes altered the text’s hac to ac.)
It occurs to me that in Old Welsh, /h/ must have had a much higher functional load than in Old Irish, in which it turns up basically just as a result of lenition and some kindred phenomena.
So on one level it’s not surprising that Welsh speakers should be keener on rehabilitating the silent letter than the Irish were.
Germanic folk seem to have done that pretty much from the start once they adopted Latin letters. When you think about it, that is not quite the obvious step that it now seems in retrospect. What was their model? Descriptions of the sound by the classical grammarians?
Etienne? (You know this sort of thing …)
So you’ve finally descended to the level of TV Tropes.
Don’t teach your grandmother to suck eggs. I was singing “When on the road to sweet Athy” before TV Tropes was a gleam in the eye of some fucking millennial.
Hey! I like TV Tropes.
But then, I am a millennial. (First millennium, of course, not that dreadful new-fangled so-called “second” millennium. Bloody Normans all over Europe. To say nothing of that religious maniac Gregory VII. Things have never been right since.)
I didn’t say I didn’t like it, but I don’t need to ransack it for cultural detritus. I grow my own detritus.
It occurs to me that in Old Welsh, /h/ must …
Um, now that we speak of approximants and fricatives I have a question about Erdoğan. We hear that /ğ/ as if it were /w/, ugye? I understand that the full /ğ/ is supposed to weaken away between those vowels – but does it properly weaken to a /w/, or is that simply the best we anglophones can come up with? A truly bare /oa/ would be unlikely for us and perhaps most other speakers; so what to interpolate? Wouldn’t /h/, for example, be anatomically nearer to the yumuşak ge than /w/ is? Wikipedia tells that /ğ/ becomes silent or is rendered as a /w/:
Is this to be trusted? Could it really ever be truly “silent” in that location, or might speakers of Turkish resort to a glottal stop or to some more benign lightening of the canonic /ğ/?
[With apologies for my many terminological and notational approximants.]
And for that matter, how are we all pronouncing “periplus” – and why? I would especially welcome answers that mention “[ex] cathedra”.
in Old Welsh, /h/ must have had a much higher functional load …
With h and /h/ bouncing around in my head, I was in a traffic queue behind an ambulance.
‘Saint John’ in Māori is ‘Hato Hone’. Also ‘Church’ is ‘Hāhi’, but ‘Jesus Christ’ is ‘Ihu Karaiti’. (These are all loans.)
I’m bemused how ‘Jesus’ got loaned without the initial consonant. Would the C19th preachers have been saying ‘Jesu’ with a yud? We have perfectly native words ‘hēhē’ (various negative senses, also breathing), ‘Hihi’ (a bird), ‘Hahei’ (a beach/village), ‘huhu’ (a grub and roasted delicacy, allegedly).
AntC: my guess is that an initial [i] might already have some frication noise, which would mask the preceding [h]. I wonder if there are any native words where initial /hi/ became /i/.
The 1996 Kusaal New Testament replaced the “Jesus” of the 1976 version with “Yesu”, which looks less foreign, but is in fact the form from the Twi Bible, which doesn’t seem a great improvement to me.
The actual form I heard people saying in real life was “Ayiisa” (with the “A” that precedes all personal names.) I suspect this was rejected as “too Muslim”, which seems unfortunate to me, if so. But it’s their language …
Dunno how the Twi form came about. I dimly recall that 19th century missionaries in the Pacific did sometimes deliberately use Hebraizing forms of Bible names as more amenable to local phonology. But the Twi one is probably simply from German (the Basel Mission was a major player in the missionary work there.)
Hat: Sad. I guess “meet the 1st quarter goals” is a universal imperative now.
Thou shalt meet the goals set for you in the 1st quarter.
Set for thee. How did I miss that?
@Noetica:
I use [ɣ], but I only have that because my L1 is Spanish.
Most folks in the UK seem to have no idea that <ğ> is supposed to represent anything but /g/, to be honest. Newscasters tend to go for something like /ˈɜːd.wæn/, but I don’t think many of them are particularly aware of how that maps to the spelling.
Out of the phonological inventory of English, I agree that /h/ comes closest, but perhaps folks feel the need to keep the rounding somewhere, and if it’s lost from the vowel then it must be taken up by the consonant?
The mystery extends to Gothic 𐌷 – and to early medieval Latin texts with Germanic words like the Salic Law, which use ch more or less consistently for… well, was it [h] yet in initial position, or was it still [x] even there…
However, the Wikipedia article on Pannonian Latin claims “disappearance of /h/ from the 4th century onwards”. If that’s true, it’s late enough for Wulfila to pick it up.
It is the most common rendering in German.
Some varieties of English don’t allow /h/ at the beginning of an unstressed syllable. Those are the people who actually say “an historical”. In vehicle I was even taught not to pronounce it.
Of course. Some languages, and some varieties of English, don’t allow vowel clusters, but others do.
I understand that the full /ğ/ is supposed to weaken away between those vowels – but does it properly weaken to a /w/, or is that simply the best we anglophones can come up with? A truly bare /oa/ would be unlikely for us and perhaps most other speakers; so what to interpolate?
From what I know, it is in fact /oa/, and I’m not sure why you find that so unlikely. For what it’s worth, Wiktionary says /ˈæɾdoan/, [ˈæɾdoɑ̯n].
I use [ɣ], but I only have that because my L1 is Spanish.
You shouldn’t be using it, because ğ is basically silent.
The Mooré Bible calls Jesus “a Zeezi.” While one can see how this happened, I still feel that some sort of prosecution is warranted.
Welsh /ɣ/ became zero in the transition from Old to Middle Welsh, thus proving the Sun Language theory.
Some varieties of English don’t allow /h/ at the beginning of an unstressed syllable. Those are the people who actually say “an historical”. In vehicle I was even taught not to pronounce it.
In my experience, Americans who say “an historical” have an /h/ in it. It would be interesting to hear what indefinite article they have before iess highfalutin words such as “hilarious” and “Hawaiian”. Google ngrams has some not necessarily dispositive thoughts.
You were taught correctly about “vehicle” (he said vehemently). However, I think that after a morpheme boundary, there’s an /h/, as in “ahistorical”. One could probably come up with a complete rule.
SCIYENCE! Some kinds of English insert [j] or [w] – or, if non-rhotic, [ɹ] – into every vowel cluster; basically an exaggeration of the first vowel. There are other languages that do the same or go for [ʔ].
Apparently, “glide insertion” is even taught to some people as part of English as a foreign language.
Here is Geoff Lindsey talking about it – as a universal feature of English, except that the accent he’s using to tell us that lacks it.
In my experience, some do, some don’t.
Yes, that one, too.
Ağ!
Some varieties of English don’t allow /h/ at the beginning of an unstressed syllable
I don’t think this is true for any variety of English that actually still has [h]; but it is true for non-word-initial /h/ before a syllable with the lowest level of stress in the complicated English four-level system, hence “vehicular” versus “vehicle”, but still “billhook”, “childhood.”
Because non-morpheme-initial /h/ rarely occurs at all in English (of any variety), there aren’t all that many cases like “vehicle.” (I have no [h] in “vehement”, myself; I suspect that those who do are using a spelling pronunciation.)
The “an historical” thing is a mere affectation, like “an hotel.” It’s not phonological. Nobody who has [h] says “an hexagonal shape.” Nobody normal, anyhow.
Loss before unstressed syllables is true for non-initial /h/ in Welsh: hence brenin “king” (Middle Welsh brenhin) but plural brenhinoedd. The pattern has been generalised to cases where /h/ doesn’t actually belong historically at all, like cenedl “nation”, plural cenhedloedd.
I have no [h] in “behemoth”, either.
I don’t care! If it means so much to you, go and be a moth!
Nobody normal, anyhow.
There is rude ambiguity in “normal”. “People who talk like that are statistically insignificant” would be more polite, and more accurate to boot. It makes clear that it’s big us against small them. It’s a numbers game.
I am not infrequently persecuted by insignificant others. They can be so annoying, despite being so few.
Not numbers, and no ambiguity: “normal” means “like me.”
There are few of us normal people left in these dark times. Increasingly, many people have never met a normal person.
The Good and Normal scoffs at mere counting arguments ? You are familiar with the works of one G.E. Moore ?
#
Good as a non-natural property
In addition to categorising ‘good’ as indefinable, Moore also emphasized that it is a non-natural property. This means that it cannot be empirically or scientifically tested or verified—it is not analyzable by “natural science”.
#
Can’t believe I read all that crap as a teenager. I suspect there just weren’t enough boys readily available to take my mind off higher things.
I recently noticed the phrase “an hexagonal wrist watch” in “Black Orchids” (1942) by Rex Stout. It’s not in dialog, but in the narration in the voice of Nero Wolfe’s assistant Archie Goodwin.
You are familiar with the works of one G.E. Moore ?
Indeed. I once even attempted to be persuaded by Principia Ethica.
However, “normal” is conceptually distinct from “good.” You are understandably confused by the fact that in my own case the attributes coincide. But this is simply a contingent fact about the universe we live in. It might have been otherwise without logical contradiction. There is a possible world in which I am not Good.
@Keith Ivey:
A neat example of my point. Clearly this “Archie Goodwin” is Good (being an assistant to Nero Wolfe), yet he is evidently not Normal.
Grammatical role seems to play a strong role in when an is used with words beginning with [h]. It seems most natural with adjectives. My perception of different levels of acceptability is approximately:
They are all affected, and I can only imagine someone like Ted Kennedy (who is indeed nobody normal) using the ones with question marks.
@Keith Ivey: That Nero Wolfe story provided the name of the last “pure historical” Doctor Who story. The First Doctor had a mixture of science fiction and historical adventures, with a science teacher and a history teacher along to explain things. Stories set in Earth history, with no SF elements apart from the presence of the main characters, were dropped after the early Second Doctor story “The Highlanders,” except for “Black Orchid,” a short Fifth Doctor whodunnit set in the 1920s.
I must concede that the data do support the notion that “an” before a non-silent “h” is most likely with a word that doesn’t have its principal stress on the first syllable. Still an affectation (unlike the “vehicle” phenomenon) but not one that has nothing to do with phonology, as I rashly asserted.
In Early Modern English, “an” was usual before “h” everywhere; similarly with “mine”, “thine” as “possessive adjectives.” As in “mine host” …
“An historical” wasn’t part of my mother’s vocabulary, but she did say “an hibachi.”
And re “Erdoğan”: The other day I was looking at “Chaadayev” and puzzling over it, when it hit me: “Chaghatay!”
Brett, all of those except “hexagonal” (which you rate as merely questionable) can be explained by stress, with no need to bring in part of speech. How do you rate the acceptability of “an historically significant find” or “an historian”?
DE, I’m not sure whether either Goodwin or Wolfe actually counts as good, but they are definitely not normal (though Goodwin, like most people, is more normal than Wolfe).
Call the ğ a [ɰ], which you can hear or ignore, and stop bickering.
I’m trying to say I’ve heard “an ‘istorical” without [h] from at least one American who was under no suspicion of general h-dropping.
It’s the version with [nh] that… started out as an affectation in any case.
That’s what the abovementioned Spanish “[ɣ]” is.
That’s Alon’s [ɣ]. I’m not sure how open it is.
Alon:
I use [ɣ], but I only have that because my L1 is Spanish.
Hat:
You shouldn’t be using it, because ğ is basically silent.
Y:
Call the ğ a [ɰ], which you can hear or ignore, and stop bickering.
David M:
That’s what the abovementioned Spanish “[ɣ]” is.
Y:
That’s Alon’s [ɣ]. I’m not sure how open it is.
I’m glad we cleared that up. I knew there’d be a straightforward answer that eluded me. Hatters may need to be reminded that Karl Zimmer and Orhan Orgun (contributing to Handbook of the IPA, 1999) gloss ğ in both uğraşmaktan and olduğunu as [ɣ].
I suppose we should add [j] to the inventory of plausible intervocalic realisations, on anatomical grounds. As for silence, well … more could be said.
Now, anyone like to have a go at “periplus” (with optional but esteemed reference to “[ex] cathedra”)?
The letter ğ was basically a concession to Ottoman spelling conventions. Not even Atatürk could completely prevail against the Forces of Orthographic Reaction. Sometimes you just have to know when to pick your battles.
I’ve heard my Hispanic son called (to my Anglophone ears) something pretty much like [miˈwel] by his friends.
Now, anyone like to have a go at “periplus”
I have always said /pɛˈɹɪplu:s/, à la grecque (περίπλους), but Wiktionary tells me it’s supposed to be /ˈpɛɹɪˌplʌs/ (perry-plus), which sounds stupid and I refuse to say it.
Noetica, in the handbook they gloss it as a phoneme (/…/, not […]), even morphophoneme, and in the paragraph above this gloss they say when intervocalic (and not between front vowels) it is a phonetical zero.
(however I think this depends on the dialect).
it is a phonetical zero.
Yup, what he said. It’s purely graphic.
https://archive.org/download/intonation-practice/Handbook_of_the_IPA.pdf
pp 167-8 (155-6)
он был увулярный дрожащий,
она – фонетический нуль…
(“он был титулярный советник / она – генеральская дочь”)
Doğru (“true”, /doɰˈɾu/, [d̪o̞ːˈɾʊ]), Dravsi. The wording itself:
So Wikipedia is quite misleading. A mere zero (or a lengthener), except between front vowels. And looking to the standardly available English options, even elsewhere intervocalically it appears that there is no greater justification for substituting /w/ (front-velar approximant) than for my proposed /j/ (palatal approximant).
Wiktionary gives silences for these:
But interestingly, at Wiktionary soğan (“onion”) gets a non-/w/ approximant in Azerbaijani and Salar, and apparently also in Crimean Tatar.
Wiktionary gives “/ˈæɾdoan/, [ˈæɾdoɑ̯n]” for Erdoğan.
More can be said about silence? Yes. The Wiktionary spoken clip for boğa sounds very like a superannuated English way of saying boa [as in “boa constrictor”], given as /bɔː/ at Wiktionary. Both sound monosyllabic, but there is a definite morphing to /a/ in boğa that we wouldn’t necessarily get in that older English boa. How does the psycholinguistic notion of syllable work with this Turkish /ɣ/? If it separates two vowel elements each of which might have been taken as syllabic (and where are syllabic in Azerbaijani, say), are two syllables considered retained no matter how /ɣ/ is realised phonetically, or only if the realisation is non-silent?
Hat:
I have always said /pɛˈɹɪplu:s/, à la grecque (περίπλους)
I thought you might; but tell me: is your preference due to the ί diacritic (which many would care not a fig for), or because of the configuration of sylalbles (à la latine) – in which case the situation is less straightforward?
= “(and which are syllabic in Azerbaijani, say)”
= “of syllables”
When Shakespeare makes two syllables of hours, as in Sonnet 5, this must be understood as necessarily involving an intervocalic approximant, ugye?
We can hardly scan the first line other than as including “howers” (the actual Quarto spelling).
Being Irish, I sound as many haitches as anybody, but if I find myself reading “an historical” or the like, I sound the /n/ and not the /h/. I recall reading a suggestion that A New English Dictionary on Historical Principles was going to be called An Historical Dictionary of English until James Murray, being Scottish, changed the “An” to an “A”, so they compromised/fudged.
Apparently, “glide insertion” is even taught to some people as part of English as a foreign language.
John Wells found out about this in 2010 and had thoughts: see linking semivowels? and linking semivowels (ii):
If talk of linking semivowels is a useful technique for getting EFL students to avoid prevocalic glottal stops (“hard attack”), then so be it.
No one addresses the topic of prevocalic linking glottal stops. I mean, if they are not there to link, what the hell else are they there for ? They don’t stop, they slyly signify by being there. What you guys call phonemes (I think).
I find these can be extremely monkey-wrenching when I listen. There are many short phrases in Heartstopper, spoken by actors with “Posh London accents”, that I still can’t understand after listening for dozens of times all told.
Glot rot.
But interestingly, at Wiktionary soğan (“onion”) gets a non-/w/ approximant in Azerbaijani and Salar, and apparently also in Crimean Tatar
But those are different languages. From what I read, there are also dialects of Turkish where ğ has different, “fuller” realizations than in Standard Turkish.
Just to add an example, in my experience the first name Uğur is pronounced not *[‘u?ur] or *[‘uwur], but [‘u:r].
is your preference due to the ί diacritic
My preference is because I know Greek, so I pronounce the Greek word in the Greek way. It’s the same reason I pronounce Nabokov with the Russian stress.
Fair enough, Hat. A fine principle to apply; and applying it, you will undoubtedly stress phronesis on its first syllable. In Latin it would fall on the second; but it’s a Greek word (φρόνησις) before it’s a Latin word (phronēsis), just as περίπλους precedes Latin periplūs – which might take stress on the second also, but on different and far less secure grounds.
If we apply that same principle to the name of the Turkish kleptocrat, we’d not insert a /w/. I’ll use a silence myself from now on – or try to, but I’ll still feel awkward about the sense that there is a reduction to just two syllables in the name.
I pronounce the Greek word in the Greek way
Classicizing or modern? I’ve encountered Greeks who insist that the classicizing pronunciation was made up as a joke by Our German Oppressors. And then there’s Oxford, about which the less said the better.
Classicizing or modern?
This is one of the rare cases where it doesn’t make much difference — the moderns have changed the pitch accent to a stress accent, and that’s about it.
I’ve encountered Greeks who claim that the classicizing pronunciation was made up as a joke by Our German Oppressors.
I wrote about that here (three months into the blog’s existence!):
Classical Greek may have had predictable word stress distinct from the pitch accent. There have been attempts to account for Greek metres on that basis, though they seem to have fallen out of favour:
https://www.academia.edu/43608786/Stress_in_Greek_A_Re_Evaluation_of_Ancient_Greek_Accentual_Typology
(The author’s name seems familiar somehow …)
FWIW (not a lot) I’m not impressed by the point that this would be typologically unheard of. There are examples among African languages: the northwest Bantu language Eton is one. (Also Kusaal, but I imagine that everyone assumed that already.)
I am quite certain that there was perceptible (not phonemic) stress on “heavy” syllables; otherwise the poetry simply wouldn’t have worked.
@DE: Classical Greek may have had predictable word stress distinct from the pitch accent.
This seems opportune enough to mention an upcoming talk by Kofi Yakpo at this online-accessible conference, along with a number of other broadly comparative Africanist (and other) talks.
Thanks!
Not at all.
Yes, but this w is not a [w] for a lot of people (and, unlike some others, it never was one etymologically).
Some of the commenters pointed out that there are accents where this does happen. Putting the way in situation seems to be particularly common, and I’ve also heard US as [jʉ̯ʊˈwæs] a few times (part of an accent that merges /ɛ/ into /æ/ in a few environments, of which a following /s/ is one). The Civilization II science advisor’s “SCIYENCE!” and Popeye’s “I yam what I yam” are probably not made up out of nowhere either.
But of course he’s right it doesn’t happen in Standard Southern British/RP.
I wondered about this (for Latin) long ago and concluded that rigorous mora timing makes it work fine even in a total monotone.* (I don’t think it ever was performed in a total monotone – especially not in Greek as long as the pitches were still phonemic – but stress is not part of how the system works.) Before I read the paper, here’s the first line of the Odyssey (in decidedly preclassical Greek):
ἄνδρα μοι ἔννεπε, μοῦσα, πολύτροπον, ὃς μάλα πολλὰ
a-n-dra-mɔ-jɛ-n-nɛ-pɛ-mɔ-ɔ-sa-pɔ-lʊ-t-rɔ-pɔ-nɔ-s-ma-la-pɔ-l-la
It is really hard for me to remember that the syllable-final consonants have to be taken seriously as as long, or as abstractly important, as a short vowel – that VC can’t be as short, or “light”, as V. But that’s because I’m at home in stress-timed languages that don’t even have phonemic vowel length etc. etc. etc.
* You haven’t heard a total monotone if you haven’t heard Mandarin with the tones stripped off by native speakers. It happens in some kind of… rap, I think, and it’s a pretty strange experience.
In Latin hexameters, there is a well-known tendency for the metre and word stress to come into alignment toward the end of the line. This only makes sense if word-stress was significant in reading the verse, though the relationship is certainly more complex than in English metres. There is no reason why it should happen if mora-counting was the whole story.
I’ve also heard US as [jʉ̯ʊˈwæs] a few times (part of an accent that merges /ɛ/ into /æ/ in a few environments, of which a following /s/ is one).
Are you sure about that? There’s an accent that lowers the /ɛ/ close enough to the /æ/ to be confusing, until you get used to it; they don’t quite merge. I lived with someone who talked with such an accent.
The paper says “stress-accent language” when it means “stress-timed language”. It is simply not the case that mora-timed languages can’t possibly make use of loudness as a correlate of stress; one-pitch Finnish is the most dramatic example.
(There is three-pitch Hungarian and two-pitch Hungarian; there is two-pitch Finnish and one-pitch Finnish. The two-pitch versions have high pitch for stressed = word-initial syllables and low pitch for everything else; three-pitch Hungarian adds a third for sentence-level stress; one-pitch Finnish is spoken at a constant pitch for several seconds without interruption and marks the stressed = word-initial syllables by loudness alone.)
Yes; I think switching this on and off was a conventionalized artistic choice that had something to do with the facts that 1) the mora-counting meters were imported from Greek rather than a homegrown development and 2) in Greek, “ignoring” the stress did not erase the positions of the stressed syllables because the pitches were still there, while in Latin this erasure was necessarily complete, so more conspicuous and more artificial.
No; I probably haven’t heard enough to exclude the possibility that the merger isn’t complete. However, the [æ] I heard in US was really wide open, so if /æ/ remained distinct I find it hard to imagine where it went.
As for “an h*”, in formulating rules it is a distracting complication to consider cases like “an hour”, “an honest woman”, “an hotel” (if that is pronounced without /h/); or cases where the h* word is effectively realised as a hj* word: “a[n] humungous error”, “a[n] humanitarian gesture”. Setting those complications aside, I allow myself a stylistically optional “an h*” according to this rule:
My rule accounts for a good deal of observed real-world practice. If extended slightly it covers the practice of those speakers who omit the /h/ after the “an”. It allows “an whoˈever-may-care attitude”, and “on an ‘whomˈever is available’ basis” (ignoring the ungrammatical “whomever”, of course).
David M:
Not at all.
At all. If we remove every last trace of intervocalic approximant in that aberrant “howers”, the whole inevitably drifts back into a perceived monosyllabicity – which is exactly what the prevailing scansion must avoid. The [w] is our anglophone psycholinguistic guarantor that a new syllable is in play. So it seems to me, after much Freud-like self-audience.
This is not inevitable either globally or across English as a whole. It is inevitable in some kinds of English, and in some other languages.
There’s a poem by Hopkins in which he writes “hoürs” to emphasize the meter he wants.
“I wake and feel the fell of dark, not day”
This is slightly mixed up, but there really was a dispute over “a(n) historical”, according to Mugglestone’s Lost for Words and Gilliver’s The Making of the Oxford English Dictionary. The project had no specific title beyond “New English Dictionary” until Murray was hired. Murray’s title (as of the first page proofs in 1882) was A New English Dictionary on a Historical Basis, and that title appeared in some early advertisements, but Oxford higher-ups kept sticking their fingers in. Late in 1883, with the first section almost ready for the printers, Jowett and Liddell complained about founded chiefly on materials collected by The Philological Society (they didn’t like a “basis” being “founded” on something else), while Fitzedward Hall objected that a historical “grates on my ear painfully”. Finally on Historical Principles was approved by the Delegates — although they forgot to change “A New English Dictionary on a Historical Basis” at the top of the first page of entries, and it remained there in reprints into the 1970s.
Many thanks, Rodger and Stu. That masterpiece is new to me. It deserves to be quoted in full for our admiration and at least a little analysis. I present it in quatrains and tercets, to show the structure more pellucidly:
Ach! The form is pure Petrarch – as is a great deal of the sense, despite the infusion of much that is Dantesque and much that is more late-nineteenth-century anguish.
What masterly play of consonants and vowels. Where best to pick examples? It glitters with them.
Which of the meanings of fell falls most to the fore? I favour the Spenserian, the least known: “gall, bitterness” (from Latin neuter noun fel, gen. fellis). This wondrously prefigures “I am gall, I am heartburn. God’s most deep decree / Bitter would have me taste: my taste was me”; but the “pelt” meaning would perhaps come to mind on a first reading, given that he feels the fell – and there is so far no hint of bile. That said, the fell of dark surely conjures up the fall of dark, and the adjectival fell (“fierce, grim”) also bristles ominously forth.
The poet’s intent is surely to deliver ten-syllable lines, given the respect he shows (at least here: it’s Hopkins, remember) for tradition and formal purity. But deferring full consideration of line 2, line 9 calls for comment:
You’d have to read each “I am” as “I’m”, to scan this right. And alerted by the metrical fastidiousness of hoürs (assuming that is indeed from Hopkins; see below), we would so read them.
Line 12 too:
Selfyeasting sourdough, a strongly Dickinsonian trope; but while this yeast may swell the dough it shrinks the дух, and spirit becomes one syllable.
Now what about line 2? Not all editions have the diaeresis – which originally, of course, meant a splitting into two syllables and soon came to be applied to the sign (¨) that marked such a split or separateness. Nor do readers of Hopkins these days take any notice of it, on the evidence of the YouTube efforts I’ve surveyed. Is it original in Hopkins, or just editorial? His editor Robert Bridges may have introduced it to help get the scansion right, for all I know. (Does any Hatter know? Do we have an MS?)
I’m temporarily without access to OED, but SOED makes hour disyllabic anyway: /ˈaʊə/. So does Longman Pronunciation Dictionary (LMP): [ˈaʊ‿ə] for BrE, and something similar for AmE. The undertie (‿) is not explained in LMP’s glossary of symbols, but here it must mean the absence of any break between syllables (just as it marks liaison in analysis of French).
How did Hopkins himself pronounce hour? One uniform way always? Was he even aware of there being more than one way? Did Bridges, preparing this poem for publication in 1918, introduce the diaeresis because it had recently begun to have two pronunciations (perhaps one disyllabic and the other affectedly or hypercorrectly monosyllabic)? Hopkins and Bridges were both born in 1844, but Hopkins composed this poem in or before 1885.
One hypothesis: for Hopkins it was disyllabic, but negotiable just as spirit is negotiable (like spirit itelf, but that’s for another Hopkins analysis …).
I have listened to (and in case made a spectrogram) of some readers’ saying hours in this poem (not that they distinguish two ways, as I have said). And where I can I’ve watched their lips. By my pycholinguistically conditioned interpretation of what I observe, some make it plainly one syllable, some plainly two, some randomly one then the other, some somewhere in between. Myself, whenever I hear it as two syllables I see some narrowing of the lips, and hear some sort of a /w/ introduced. The things about approximants is – as we see from ample evidence upthread – that they are approximate.
Hard phonetic facts-on-the-ground may be clear, and I would willingly sit at David M’s feet with slate and chalk to learn. But the perception and the making of syllables, and their theorising, is another and far more disputed matter. Even I know that: and I experience it, as a keen listener and speaker.
Heh. Corrections:
“I have listened to some readers saying hours in this poem, and in one case made a spectrogram …”
I’m mildly surprised that the poem is new to you, and delighted that you discovered it here at the Hattery (Home of High-Quality
PotteryPoteryPoetry)!The stunning part of this is that both pronunciations occur in the same line.
In the Lay of Igor’s Campaign, Мъстиславъ sometimes scans the new way (/mstiˈslav/, two syllables; today spelled Мстислав) and sometimes the old way (/ˌməstiˈslavə/, four syllables). I doubt that happens in the same line, though.
You heart, you!
It’s not “and ten low words oft creep in one dull line”, but the stress that should be put on the first “you” still breaks the meter.
Perhaps, but they may be just as complex as you observed. In languages that don’t put a terrible amount of importance on syllables or morae, like English or German, it happens that the number of syllables in a word varies with stress and such for a single speaker, and also that the number isn’t an integer.
As an example, let’s take a simple Standard German word like… hören. There are people, many of them, who render it somewhere around [ˈhøːʁɵn] at least in clearer, slower pronunciation. Straightforward enough, right? No. It turns out they do this because they live in, or close enough to, places where the -n drops off more easily than the -e- drops out. (This has happened in Dutch and Swiss, and Hans has said it also happens in between.)
Where I’m from, the -e- has dropped out (except optionally in singing and the like). This puts the -r- right in front of a consonant, so non-rhoticity kicks in. That’s where the fun starts. You’d expect *[ˈhøːɐ̯n̩], with a long diphthong [øːɐ̯] followed by a syllabic [n]. For some reason *waving in the general direction of some sonority hierarchy*, this never happens. (In really exaggerated pronunciation, the -e- is restored instead, and you get [ˈhøːːɐ̯ʀɛnː].) But the logical redistribution, *[ˈhøːɐn] with an unremarkable syllable [ɐn], also never happens (maybe because it would break the diphthong). What happens is… well… I would say a single, very long syllable with an outright stretched diphthong: [høːˑɐ̯ˑn], let’s say.
All this is with full stress. When the stress is taken away, as in aufhören or zuhören, things do approach sanity, e.g. -[høˑɐ̯n] (single syllable).
(Likewise lehren, leeren, führen, every -ieren, and so on.)
It gets even more eldritch if there’s no underlying vowel length. Depending on sentence-level stress, there is a difference, artificial though it is, between Herren (pl.) and Herrn (sg. except nominative). The latter is straightforwardly [hɛɐ̯n] (single syllable; BTW, that’s also how hören comes out in my dialect, where underlying vowel length isn’t a thing to begin with). The former comes out the same when it is unstressed enough; otherwise… I would say it grades seamlessly between a single syllable with a stretched diphthong, [hɛˑɐ̯ˑn], and the two-syllable version that doesn’t occur for hören, [ˈhɛɐ̯n̩].
Taking diphthongs out of the equation doesn’t change a lot. While there are, AFAIK, Standard accents that distinguish /aɐ̯/ and/or /aːɐ̯/ from /aː/, most don’t, and so fahren comes out as [ˈfaːn̩] there, two syllables, when it occupies a stressed foot (or something) all by itself. Mess with this, as stress-timing routinely does, and the syllable boundary weakens and soon disappears, and then fahren, fahren, fahren / auf der Autobahn rhymes. (…The Bahn in Autobahn is not normally stressed enough to maintain its full length, but here the tune takes care of this and even provides full length to the o.)
This suggests, BTW, a phonemic contrast between /n̩/ and /n/. Remarkable as that is in global context, it’s not so strange given the minimal pair Karl [ˈkaːl̩] and kahl [kaːl]. (Alas, no thanks to adjective declension, Charles the Bald – Karl der Kahle – is not himself a minimal pair.)
From the point of view of Chinese or Japanese, or Spanish, Italian or Polish, or French or even Russian, all the above is psychedelic. English, I think, does get there sometimes.
The lip-reading test wouldn’t work on me, though. There’s a lot more lip and even jaw movement in English than what I’m used to.
Me two days ago:
I once read a Japanese children’s rhyme. It has to be said this way or even more so, with the syllable-final consonants as outright syllables, or the meter collapses completely.
Also, for comparison, here’s what I want to make of the quoted line intuitively, i.e. being used to southeastern German:
ˈanː-dra-mɔ-ˈjɛnː-nɛ-pɛ-ˈmoːː-sa-pɔ-ˈlʊtː-trɔ-pɔ-ˈnɔsː-ma-la-ˈpɔlː-la
…and that, I think, would have sounded startlingly barbarous to any Homer or even Virgil. It would also make it hard to apply the pitches (some of them would be on unstressed syllables even though other syllables would be stressed) and probably ruin the lost tune. It does still scan, but at a cost.
I have a vague memory that in Ancient Greek s was voiced before voiced consonants (hence …ozm…), but I don’t know where I got this.
I was taught (so I qualify this, as often, because it was decades ago) that omicron was a close vowel and omega was an open one. A long close vowel was written /ou/, same as the diphthong. How this applies to the closeness of the short vowel, I suddenly don’t know.
On page XXXV (“5.3 Other Symbols”) it is explained as “possible compression”, referring the reader to p. 173 where the phenomenon of compression is explained and demonstrated (“Sometimes a sequence of sounds has two possible pronunciations: either as two separate syllables, or compressed into a single syllable”).
Catherine Phillips’ Oxford Authors edition of Hopkins, taking the text from Hopkins’ fair copy autograph, doesn’t; but in her annotations she points out that hours has a circumflex as a metrical mark, “to express that they are made to approach two syllables — he-ar etc” (Hopkins, apparently from a letter)
W. Sidney Allen, Vox Graeca, 3rd ed., p. 45/46, for example.
W. Sidney Allen denies the closeness of epsilon and omicron, but I don’t find his arguments very convincing.
W. Sidney Allen, Vox Graeca, 3rd ed., p. 45/46, for example.
Thanks, that must be it; Vox Graeca was my vade mecum once upon a time.
Yes, Ulr, there the undertie is – in the print third edition of LPD (“LMP”? What was I thinking?), and explained elsewhere under the rubric of compression but nowhere given a name. At my computer – once the fell of dark has given way to the hérissement of day – I’ll recheck my digital version. I suspect that it’s rendered there much smaller than in the pronunciations themselves, where like Molly’s drawers of India mull they are “cut on generous lines”. No such mismatch in the print version.
And a million thanks for the invaluable circumflex intelligence.
Hat:
I am mildly surprised that you would make that observation, since I know few poems and can recall at will only a handful of lines with any accuracy. I feel this as positive – though it would make a feller misery of prison in an Arctic Circle Russian prison, if such a recherché calamity were ever to befall me. It lends a freshness and immediacy to discoveries in later decades, and keeps undulled the stiletto of penetrating analysis – leavening the дух.
The readiness is all …
Missed the cut-off by one second: “time in an Arctic Circle Russian prison”.
I have listened to some readers saying hours in this poem, and in one case made a spectrogram
Puts me in mind of South Americans speaking of Pisco sah-wers.
I am mildly surprised that you would make that observation, since I know few poems and can recall at will only a handful of lines with any accuracy.
I’m afraid I have a fixed idea of you as a connoisseur of poetry, and I’m too old to alter my fixed ideas. You must forgive and endure.
late to several parties, but:
/ğ/
it was originally explained to me as a sign to lengthen the preceding vowel, which is what i think i hear turkophones doing with the fake-coup-impressario’s name.
insignificant others… so few
o stu, i was under quite the opposite impression!
hardly knew you
@mollymooly: that shift seems to me like it must run alongside (or result from) the fading of the sense of “knew” meaning “recognize” (which i want to say survives mostly in the present tense).
Oh. Yes. That’s the case in modern Greek and probably goes way back; I simply forgot.
That seems to be textbook wisdom or at least Wikipedia wisdom, and likewise for ε & η. The main, if not only, argument is that contractions of ε + ε and ο + ο from the loss of /j w h/ don’t produce η & ω in Attic & Ionic, but ει & ου. The fact that the main source of η in these dialects is long α (which is preserved as such in the other dialects) argues it must have been a rather open vowel at least in let’s-say-Homeric times, and the fact that ει merged into /iː/ (together with ου becoming a new /uː/) long before η did means it must have been less open; and from this it is extrapolated that ε & ο were less open than η & ω as well.
But interpreting this system as ε & ο being [e] and [o] produces a system that makes no synchronic or diachronic sense.
– If you leave an [a e i o u] inventory alone for a while, it’s going to change into [a ɛ i ɔ u] pretty quickly. This happened in Spanish: starting from an /a ɛ e i ɔ o u/ system, /ɛ/ and /ɔ/ quit by becoming diphthongs, but now most kinds of Spanish have [a ɛ i ɔ u] with open, not even mid, vowels.
– The rest of IE points at the main ancestor of ε having been *[ɛ] all the way back to Proto-Indo-Anatolian, and the main ancestor of ο having been *[ɔ] at least since Proto-Indo-Actually-European. Likewise, the other main ancestors of η come out as *[ɛː] and *[ɛh], and that of ω as *[ɔː].
– In the Modern Greek I’ve heard, ε/αι is an unremarkable [ɛ], while ο/ω is a striking mid vowel (a sound I’m not used to).
So here’s what I think happened:
Proto-Greek: [a ɛ i ɔ u aː ɛː iː ɔː uː ɛj ɔw]. Deceptively neat & simple. (Also [aj ɔj ɛw aw] that I can safely ignore here.)
Attic-Ionic starts by increasing the difference between /a/ and /aː/ by shifting the latter forward & upward:
[a ɛ i ɔ u æː ɛː iː ɔː uː ɛj ɔw]. There are a few inscriptions that seem to use different letters for /a/, /æː/ and /ɛː/.
Then, /ɛː/ merged into the more common /æː/, and /ɔː/ apparently became [ɒː] in solidarity:
[a ɛ i ɔ u æː iː ɒː uː ɛj ɔw].
Then, during the Homeric period, /j w h/ fell, ε + ε produced a new /ɛː/, and ο + ο produced a new /ɔː/. Next, /ɛj ɔw/ joined these new long vowels, making it possible to spell them ει & ου. Similarly, a rare new /aː/ resulted. The resulting system is how I’d read Epic Greek if I had to:
[a ɛ i ɔ u aː æː ɛː iː ɒː ɔː uː]
α ε ι ο υ α η ει ι ω ου υ
Soon thereafter, /u/ & /uː/ got fronted (in Attic-Ionic), and the really bottom-heavy system of long vowels spread out more evenly. This resulted in the (likely) Classical system:
[a ɛ i ɔ y aː ɛː eː iː ɔː oː yː]
α ε ι ο υ α η ει ι ω ου υ
Still a bit crowded and bottom-heavy, and with a hole for /u/ & /uː/. By Roman times (as seen from transcriptions into Latin and Gothic), ει had merged into the previously rare /iː/, and ου had become a new /uː/. Vowel length was also lost at some point:
[a ɛ e i ɔ o u y]
α ε η ει=ι ο ω ου υ
Still too much. I’m not sure why, but the next step was asymmetric; I suppose part of the reason is that the vowel chart is asymmetric, too. The frequencies of the different phonemes, which I don’t know enough about, must also have played a role. Medieval situation (and modern in Megara and Aigina):
[a ɛ i o̞ u y]
α ε η=ει=ι ο=ω ου υ
Then merge /y/ into /i/, and you get the modern standard.
David M: Thanks, that’s wonderful.
(Vox Graeca! Of course. Well, I read it 57 years ago …)
…and the reason I thought this is that zmaragdus had an entry near the end of my Latin dictionary in school; the dictionary didn’t cover post-imperial Latin.
Interesting. My dictionaries spell it smaragdus, and Georges’ Lexikon der lateinischen Wortformen (1890) tells us “in Inschrn. zmaragdus […] in Ausgaben gew. smaragdus“. Of course, Georges must have been aware that the orthography most editors use is what they think was the classical orthography (medieval manuscripts are usually not very helpful in that respect).
Mine made clear that was the usual rendering.
What’s interesting is that the inscriptions with z give us a terminus post quem for [s] > [z] before resonant in Italian. (And now I wonder if that comes from Greek.)
The σ- in σμάραγδος also got rendered as -z- in Aramaic, as for example אזמרגדין ⟨ʾzmrgdyn⟩ ‘emeralds’ (Targum Onkelos, Exodus 28 in the list of the gemstones of the priestly breastplate) and Syriac ܐܙܡܪܓܕܐ ⟨(ʾ)zmrgdʾ⟩ ‘emerald’ (Peshitta, Ezekiel 28:13). There would have been nothing wrong with a cluster -sm- ܣܡ ,סמ. Also eventually Middle Persian uzumburd, Persian and Arabic زمرد zumurrud, Russian изумруд, etc. (Where is the initial и- from? Alteration of Turkish zümrüt with influence from измарагд?)
Today the OED’s rolling ‘Recently Added’ word list includes a word apparently from a Western Oti-Volta language, azonto:
With etymology:
I wonder if our resident expert on Western Oti-Volta can comment on the Frafra etymology? After seeing so many etymologies in the OED like divan, dalgona, and virago (as in the thread above), my knee-jerk reaction to newer OED etymologies now is that they are most likely deficient in some way.
From the article ‘Azonto – The New Music and Dance Craze in Ghana’ (2012) on ModernGhana.com:
From p. 97 in Akosua Darkwah and Dzodzi Tsikata, ‘The Changing Conceptions of Work and the Language of Work in Ghana: Towards a Research Agenda’, in Deborah Hill and Felix K. Ameka, eds. (2022) Languages, Linguistics and Development Practices:
And so, a Gbe form like *azɔntɔ : *à-, noun prefix + *zɔ̃́ ‘work’ (?; Ewe dɔ̌ ) + *-tɔ́, agent noun suffix. Cf. Gun àzọ́n ‘work’.
Did the OED confuse Frafra with Phla–Pherá ?
*sighs deeply, awaits DE*
For the curious, sound file of Ninkare Frafra zonko ‘hair’, pl. zomto in Ninkare Frafra here.
Zonko is “(one) human head hair”, as opposed to kõbgɔ “human body hair; animal fur/fleece; bird feather.” Kusaal zuobʋg, kɔnbʋg respectively. Zonto just means “human head hair”; the gloss “natural (curly) hair” is just embroidery (and a clear sign that the writer hasn’t consulted any actual sources on the language.)
A is (of course) not a “masculine nominal prefix”; it’s the same as the A which precedes Kusaal and Mooré personal names (male or female.) It’s also used as a sort of ad-hoc impersonal clause nominaliser, as in Kusaal a-sʋ’ʋ-m-mɛŋ “independence” (A-“own-my-self.”) It appears in a few Farefare placenames (but not in Kusaal placenames) where the places seem to be named after (mythological) people.
A Gbe origin seems vastly more probable on pretty much every ground, linguistic and soiciological.
“Frafra” is, apparently, dispreferred by the actual people themselves (properly called “Gurense”), though oddly enough they seem happy with “Farefare.” The exonym is based on a Gurenne greeting (as with “Jamsay.”)
Ninkare is a different dialect of Gurenne from (Bolgatanga) Farefare, in which the word in question is in fact zuomto. Ninkare is from the Burkina side of the border, making it even less likely as a source of the name of a Ghanaian dance craze.
A is (of course) not a “masculine nominal prefix”; it’s the same as the A which precedes Kusaal and Mooré personal names (male or female.)
I had guessed that already — your work here has not been in vain!
The actual form zonko is weird, incidentally: the stem is from proto-WOV *zo:b- everywhere else, eg. Mooré zóobgó.
The stem has evidently added a derivational suffix *m, which is indeed a thing that happens for no discernable reason in Oti-Volta languages sometimes. But even more, the class suffix form is strange: the suffix is basically *go, and you usually only get k as a result of the internal sandhi change *gg > kk, e.g. dʋkɔ “pot”, plural dʋgrɔ.
This may be a case of a mysterious interchange between the derivational suffixes *m and *g that turns up within the paradigms of a few WOV nouns.
Thanks as always for all those details, DE!
That is a very charitable expression!
Does this devoicing occur with the other geminated voiced consonants, too?
To circle back to the original topic of this thread, Oti-Volta dʋg- is obviously Sumerian 𒂁 dug.
The OED’s ‘Word of the Day’ today is amusing: rounce robble hobble ‘The sound of a clap of thunder’.
Here is the full passage quoted in the first citation from Richard Stanyhurst. Rounce robble hobble fills out the Adonic clausula. What bathos! The second citation makes it clear to me that I must try read all I can from Thomas Nashe, a feeling I have had vaguely for some time now.
I see there is typo caelodeicit in the attribution of the first citation in the electronic OED. I wonder if that is in the print OED2 or crept in during electronic conversion. Here is the passage in question, 8.424–432:
Rounce Robble-Hobble would be a nice name for a dog. Rounce for short.
Does this devoicing occur with the other geminated voiced consonants, too?
Yes: *gg > kk, *dd > tt, *bb > pp. The rule must actually go back to proto-WOV, but secondary voicing of postvocalic stops has obscured the situation in Dagbani and especially in Dagaare (the French of Western Oti-Volta.) Most WOV languages (including Farefare) have also changed postvocalic *d > r, too, so the rule ends up looking like *rr > tt.
In Kusaal. geminate devoicing remains entirely transparent and synchronically active:
gik *gɩgga “dumb” plural gigis (cf biig “child”, plural biis)
bʋd “sow” imperfective bʋt *bʋdda (cf kʋ “kill”, imperfective kʋʋd)
sɔb “write”, gerund sɔp *sɔbbɔ (cf kʋ “kill”, gerund kʋʋb)
To circle back to the original topic of this thread, Oti-Volta dʋg- is obviously Sumerian 𒂁 dug.
No doubt about it …
Its immediate source, however, is from the verb dʋgɛ “cook”: Nõotre dɔɣe, Kusaal dʋg; Mooré dʋge; outside WOV, Buli digi, Konni dig, Nawdm duug-.
There is a proto-Bantu *dúg- “cook” which looks tantalisingly like a cognate, allowing for a reconstruction to proto-Volta-Congo itself, but I have my doubts: the Oti-Volta cognates all belong to what I now call the “Outer Oti-Volta” subgroup, with no good evidence for it going back farther, to proto-Oti-Volta, and the proto-Bantu form in turn seems to be regional rather than solidly pan-Bantu.
So probably two independent episodes of borrowing from Sumerian …
(hey, the timescales work: roving bands of Sumerian refugee potters fleeing to Africa after the collapse of Ur III …)
The OED’s ‘Word of the Day’ today is amusing: rounce robble hobble ‘The sound of a clap of thunder’.
Thanks for that! It’s apparently derived from rounce ‘intransitive To behave in an agitated, boisterous, or noisy manner; transitive To shake.”
DE: So should I go into your earlier comment and change the direction of the arrow in “you usually only get k as a result of the internal sandhi change *gg > kk”? I originally thought Xerîb was misreading you when he talked about devoicing, but since you confirm that, I presume the arrow should have gone the other way.
No: voiced geminate stops become devoiced, so *gg > kk is the right way round. I hadn’t noticed before, but Xerîb has accidentally transposed my original > to < there.
Just to confuse the issue (and why not?), from a historical point of view, WOV g d b after vowels actually arose from earlier *k *t *p, so the geminates may in fact preserve the original lack of voice. But that’s gratuitously complicated for a synchronic description, and also runs into problems because the voicing of *k *t *p after vowels can’t all have happened at the same time: *Vp > Vb had probably already happened in proto-Oti-Volta, for example: cf “bone”: Mooré kóbre, Buli kōbī, Nawdm kówŕ, Gulmancema kpábílī, Waama kɔ́bídē but proto-Bantu *-kʊ́pà.
Argh! I meant to quote Xerîb’s version but then I quoted yours and didn’t notice the arrow went the other way.
[emily litella]
Never mind!
[/emily litella]
I’m actually in the throes of reworking my ideas about the fate of non-initial stops in proto-Oti-Volta, initially driven by the fact that what I’ve currently reconstructed as non-initial *d probably lumps together two distinct consonants (both of them also distinct from *t.)
My current treatment kinda works, but at the cost of proposing some implausible-looking internal sandhi rules for the individual branches/languages, which I’m starting to think are basically epicycles …
The problem is complicated by the fact that the Atakora Sprachbund, in which all the Eastern Oti-Volta languages now live, doesn’t like voiced obstruents and has definitely devoiced quite a few, but in patterns that don’t always lend themselves to nice neat Neogrammarian sound laws.
Unless the storm is so bad that Batman appears. In that case it’s KRAKOOM.
Apologies for mixing up > and < above.
Another OED entry I noticed this morning, from the ‘World Englishes’ publicity push in the rolling ‘Recently Added’ list on the OED splash page: Philippine English ate ‘An elder sister. Also used as a respectful title or form of address for an older woman’. (When I stay at my brother’s bilingual English- and Tagalog-speaking household, I use this word all the time to refer to the older of my two nieces in the presence of my younger niece—from the point of view of the younger niece, the older niece is ate. The Tagalog term for the youngest child in the family is bunso, which is what we call the younger niece.) Here is the whole OED etymology:
Really, that’s it? It is well established that Tagalog ate is a borrowing of Hokkien 阿姊 a-chí, a-ché (entry in Wiktionary here). Compare Tagalog kuya ‘elder brother’, also from Hokkien, for which the etymon is usually given as 哥仔 ko͘-iá, but some Philippine Hokkien pages give it as 哥兄 ko͘-hiaⁿ. (I am not qualified to say which, offhand.) In any case, made up of 哥 ko ‘elder brother’ + 仔 -(i)á, diminutive and hypocoristic suffix, apparently a reduced form of 囝 kiáⁿ ‘son, child’, or 兄 hiaⁿ ‘elder brother’, but used in Hokkien as a honorific suffix for a male older than the speaker (for the loss of final nasalization in kuya in this latter case, compare lumpia ‘spring roll, lumpia’ from Hokkien 潤餅 lūn-piáⁿ).
Gloria Chan-Yap (1980) Hokkien Chinese borrowings in Tagalog, p. 65, note 13 (here, referring to p. 31, §2.3.3) remarks that we should expect Tagalog *atse from the Hokkien 阿姊, but she offers no explanation for the actual form ate. Speculating on my part… perhaps simplification of ts [tʃ] in baby-talk? The cluster ts is found mostly in loanwords.
I wonder why the OED editors did not include any of this. Or did they even bother to look? Consider the OED entry for Philippine English kuya, mentioned above, with no Hokkien etymology given either. Nothing beyond Tagalog. But this word 哥 ‘elder brother’ (as in Mandarin哥哥 gēge ‘older brother’, etc.) is a loanword into Middle Chinese from a Central Asian source—compare Old Turkic aqa ‘older brother’ and the word in the (Para-Mongolic?) Xianbei language recorded as 啊干, to be read in Early Middle Chinese as something like ʔa-kan. This 哥 partially replaced earlier 兄 (Mandarin xiōng, Hokkien hiaⁿ). Old Turkic aqa shows up in Anatolian Turkish as ağa ‘agha’. Similar words can be seen all over central and northeast Asia: Middle Mongolian aqa ‘elder brother’; Evenki akā; Yukaghir акаа (borrowed from Tungusic), etc. This borrowing occurred in a period marked by the relative rapid rise and fall of many powerful ethnic confederations of Turkic, (Para-)Mongolic and other peoples of the eastern Eurasian steppes that came to rule parts of China at times. The OED could even cross-reference the entry for kuya to the entry for aga, agha, where all these facts could be mentioned.
All this information would be available to the OED editors through a few minutes of Googling with some linguistic common sense. It is easy to discover that Tagalog has many interesting layers of loanwords. I have the feeling that if the proximate etyma of Tagalog ate and kuya had been European words, instead of Chinese, we would have seen much fuller etymologies. Compare the current OED entry for Philippine English tiyo ‘uncle’ (included under tio), which takes the word back via Spanish to Greek, with further remarks on nursery language and a reference to Isidore’s comment on the word in the 7th century (Tius graecum est). No such depth for ate or kuya, despite the great general historical interest of the etymologies.
Good lord, that’s depressing. What’s happened to the OED?
Of course, nerds (specialising in something else) often produce better results than overly confident specialists who aren’t nerds. But it is a good practice to have a specialist, so the question is why they don’t have one for…er, very Oriental langauges.
(I don’t know how to call East which is not, say, Maghreb and not even the Near East. Technically it is the Far East, but people rarely call China “Far East”)
People unaware of the Russian Far East (i.e. unaware that there’s something behind Siberia) do use “Far East” for China, Japan and the like.
DM, yes, the Russian usage is confusing.
But do you mean that (Western) Europeans will think of (whole) China and Philippines if I say “the Far East”?
If so, that’s good.
P.S. Though in Wikipedia Kazakhstan is not “Far East” anyway:( It makes some sense if our Near East extends as fa as Islam does… But.
Hm. I’d say “Far East” is an unremarkable term in my American experience and China and Japan are the first places people would think of, followed by everything from there to the Philippines and Indonesia. Some particularly knowledgeable people might include eastern Siberia, and I’ve seen Mongolia included.
But do you mean that (Western) Europeans will think of (whole) China and Philippines if I say “the Far East”?
I can’t speak for Europeans, but Americans certainly do.
Hispanophones also speak and think in terms of “el lejano oriente.”
Same with German Fernost.
“Far East” is still current in Celto-Anglic English; not as much as “Middle East” but far more than “Near East”.
I think the boundaries of the region are fuzzy and vary between speakers. Wikipedia defines “Far East” as encompassing “East Asia”, “Northeast Asia”, and “Southeast Asia”, but its definitions of the latter two subregions overlap with the first. For some anglophones I guess “Far East” means East Asia, including the overlapping parts of Northeast and Southeast Asia.
“Far East” to me suggests Asia east of India, with China and Japan as the prototypical examples. Central Asia, not really, and definitely not “South Asia” (India/Pakistan/Bangladesh/Sri Lanka.)
It’s not altogether different from “Sinosphere”, for me, though I would grant the Russian Far East honorary membership.
“Middle East” is similarly a cultural notion masquerading as geography. Morocco is “Middle East.”
“Eastern Europe”, too, was formerly a politico-cultural rather than strictly geographic concept. Finland – Western Europe; Czechoslovakia and the DDR – Eastern Europe.
It’s not altogether different from “Sinosphere”, for me
Same here.
In my English, I don’t include Southeast Asia in “East Asia”, except maybe Singapore.
@How dywan has wandered to the floor in Poland.
Unfortunately dywan was not among the words included in this dictionary, so I managed to find only some post quoting Etymological dictionary of Polish language by Andrzej Bańkowski from 2000 and an entry from dictionary by Aleksander Brückner published in 1927. Both seem to suggest that dywan in the sense of “council of viziers” firstly changed to “chamber where divans are convened” and as such rooms were richly decorated with carpets, then the name has been transferred to them.
Is Southeast Asia included in South Asia?
Burma is, but definitely not Vietnam. Thailand, Laos and Cambodia are probably “oh, yes, them too, now you mention it” members of the Far East, like the Philippines and Indonesia. And Vladivostok.
For German Speakers it’s also fuzzy – most would include all of Asian Russia up to the Bering strait into Sibirien, not Fernost; the Philippines and Indonesia can get a similar “now-that-you-mention-them” inclusion into Fernost, but if you talk pecifically about that region, people would probably rather say Südostasien. Indochina, Thailand and Burma traditionally are Hinterindien “Behind-India”, but that term, still used in my parents’ geography textbooks from the 50s, seem to have fallen out of fashion.
@Hans, made me think of Zahumlje (medieval land or principality in Croatia and Bosnia), застолье and numerous Russian toponyms (also по-: подунавье, при-: приднестровье…)
It is tempting to translate them (literally) as, e.g. hinderhillness or hinderhillity (the land behind the hill, Zahumlje), because neuter abstract or mass nouns do sound abstract:)
My German school atlas (1967, Diercke) doesn’t mention “Ferner Osten” (and there is no “Naher Osten” either, although then as now this was a commonly used term in the daily news).A map of China, Korea and Japan is entitled “Mittel u. Ostasien”. Southeast Asia, Indonesia and the Phillipines are “Südasien”. “Sibirien” is everything in the Soviet Empire east of the Ural and north of Kazakhstan (I don’t remember the latter ever being mentioned, whether in school or in the news). I don’t think “Fernost” is much used these days, and back in the day it was more a cultural than a geographic term – an amalgam of the popular impressions of Chinese and Japanese culture.
made me think of Zahumlje (medieval land or principality in Croatia and Bosnia), застолье and numerous Russian toponyms
Заволочье.
Kazakhstan (I don’t remember the latter ever being mentioned, whether in school or in the news)
Probably not; when I started working in Kazakhstan in 1993, most people didn’t even know that it existed, never mind where it was. Friends and neighbors of my parents used to ask “Does your son still work in Russia?”
The Arab world (Muslim, Christian or Jewish) is the south-eastern part of the Roman empire.
China is far. (From Europe). Independent agriculture, Chinese characters, people and their bodies.
And there is a problem with “East”, “Orient”, “Easterners” and “Orientals”. Arabs are from there, Chinese are from there. They understand each other well, eh? Eat same food, wear same dress…
Usually, when someone speaks about the East, she means no more than one of these two.
A decade after the dissolution of the USSR, the continued obscurity of Kazakhstan was part of the joke of Borat.
@languagehat
If you could, please retrieve my previous comment from shadow realm.
Done!
And thanks; this makes sense:
Both seem to suggest that dywan in the sense of “council of viziers” firstly changed to “chamber where divans are convened” and as such rooms were richly decorated with carpets, then the name has been transferred to them.
For Russians the least known “stan” is, of course, Pakistan.
A Belgium of the East (but Pakistan is known to have fundamentalists, generals (at least one…), conflicts with India and the bomb).
Belgium of the East
In the sense, “united against a neighbour of very similar general culture only by religious affiliation”? (Didn’t work in Pakistan, notoriously dysfunctional in the case of Belgium. American “Christian Nationalists” should take note; but learning anything at all from uncomfortable facts is not part of their mindset, or even within their capacity. Deus vult!)
If we’re dealing with stereotypes, Belgium’s is “boring but makes good chocolate”, distinguished from the other one of that stereotype by being flat. Not a good match at all to Pakistan.
Belguim is not that flat. I mean compared to Bulgaria it is. But compared to the midwest.
@languagehat
Furthermore, the older word for carpet, kobierzec, may also be apparently derived from some Turkic loanword.
As I’ve mentioned before, I still had Vorderindien (“Front India”, i.e. India) and Hinterindien (“Rear India”, i.e. Indochina) in my school atlas of 1993 on a physical map of Asia, but that was just cartographic tradition at that point. Naher Osten and Ferner Osten weren’t; the former is in the news daily, the latter is not. The shorter forms, in Nahost and in Fernost, always seemed rare to me, except of course in compounds (der Nahostkonflikt, yay).
Oh. Задняя Индия (Rear India) would sound as “a shithole” in Russian. And not because of “Индия”.
(я нашёл себе жену / на Кольском полуострове
сиськи есть и жопа есть / и слава тебе, Господи!
I think there’s some easy way to change it to “где-то в задней Индии”, but I’m too sleepy (and dumb?) to invent a rhyme to “Индия” as good as полуострове-Господи*)
* I wonder if the hiatus in poluóstrove rhymes with the weird sounds heard instead of /g/ in Góspodi.
“weird sounds” [ɦ] or something between [ɦ] and [ʕ], but sometimes weakened. [ʔ̞]?
Or also with lenghtening or diphtongisation of o?
More carping… I thought this word was interesting, so I’ll just write up what I found after a few minutes of googling for LH readers who might be interested. A while ago the OED’s ‘Word of the Day’ was anting-anting, which I just found scrolling through my Twitter feed (the entry should be accessible to everyone at least for while from the OED’s search page as part of the dictionary’s World English publicity push):
No farther than Tagalog? The word anting ‘earring’ is a widely and—it seems—recently disseminated word in the Southeast Asian Archipelago. For more on the institution of anting-anting and magical earrings, LH readers can consult the Wikipedia article on agimat.
Blust’s online Austronesian Comparative Dictionary says that the ultimate source of all forms is Malay anting, under the rubric earring here. For the Malay, see for example anting ‘pendent, hanging down and swinging’ on p. 41 in Wilkinson here.
Could anting ‘hanging down and swinging’ belong to a larger family of Malay words including untai ‘hanging loosely and dangling’, untang ‘pendulous, swaying’, and untang-anting ‘swaying; all alone in the world’? The vowels in this last collocation remind one of typical Malay and Indonesian reduplicative formations like bulak balik ‘back and forth; round trip (ticket)’, gerak gerik ‘various movements’, etc. I was wondering if Malay anting could have been extracted from a reduplication like untang-anting. Blust gives a cognate for untai in Ilocano ontáy.
The Wiktionary, on the other hand, says that the Malay word is borrowed from an Aslian language. Compare Aslian words for ‘ear’ such as Jahai ʔanteŋ, Kensiu ʔəntiŋ, Temiar gɛntɔk, Semaq Beri ntəŋ, Semelai təŋ, etc. See no. 555, p. 190, in Harry Shorto (2006) A Mon-Khmer comparative dictionary, where the Aslian terms are derived from a Mon-Khmer *knt₁əŋ ‘to hear’. (I believe Austroasiatic *t₁ generally has the reflex t throughout Mon-Khmer, where *t₂ has the reflex s in some positions in Munda and sibilant and lateral reflexes in some northern Mon-Khmer groups.) The person who added the etymology in the Wiktionary also made changes to other Mon-Khmer related words. I haven’t been able to follow this Aslian etymology up to a published source, but Geoffrey Benjamin (2012) ‘Why you should study Aslian languages’ proposes other Aslian loans in Malay: cōmon ‘nephew’ (Bowrey 1701; cf. Temiar kmɔn ‘nephew, niece’); ketam ‘crab’ (vs. Indonesian kepiting), helang ‘eagle’, semut ‘ant’, cucu ‘grandchild’, merak ‘peacock’, pergam ‘imperial pigeon’, dekan ‘bamboo-rat’, jenut ‘salt lick’ (cf. Temiar jnɔɔd, from jɔɔd ‘to lick’; love that infixing Mon-Khmer morphology!). I do not have the expertise to evaluate an Aslian etymology of a Malay term, but there is there is a hitch in the semantic gitalong when we consider the mismatch between Malay anting, apparently originally simply ‘hanging down, pendent’, and Aslian ‘ear’. As far as I have been able to discover, in recent times Aslian peoples have not practiced any extreme modification of the earlobes such as is found among some peoples of Borneo, for example. In any case, if you’re going to feature an entry as Word of the Day, why not fill out the etymology? I guess I will have to write Benjamin…
All of this was available to me just now after a few minutes of Googling, and would be to the OED editors too. And they could follow up by writing Benjamin, who as been working on Aslian for decades, it seems. In any case, it would be nice if the OED included at least the Malay etymon and a remark about how widely the Malay term has apparently spread. If anting were a western European word, the OED would quote all the cognate forms in all the Western European languages, perhaps at tiresome length in all their spelling variations with their respective dates of attestation…
I love your carping, and I share your exasperation!
Today, as part of the OED’s Word English publicity push, the Nigerian English word
wahala ‘trouble; inconvenience; problem, bother’ appeared in the rolling ‘Recently Updated’ list on the OED splash page. The OED etymology:
Is that all that can really be said? Many different sources say that Hausa wàhalā̀ ‘trouble, affliction’ is from Arabic وهلة wahla ‘fright, terror’. Already Robinson in 1899 under woha(l)la suggests an Arabic origin for the Hausa. And Heath apparently gives alwaxala ‘grand problème’ for Koroboro Senni in his dictionary of that Songhay language. I don’t understand the x for Arabic h in that case. (Has there been some crossing with Arabic وحل waḥl ‘getting stuck in the mud, sinking in mire’, وحل waḥal ‘mud, mire, morass’, with mediation through Tamashek?)
The Hausa word is obviously from Arabic.
On reading the snippet more carefully, I see that the OED do actually understand that the Yoruba word is borrowed from Hausa: they are just wondering if the word got to Nigerian English via Yoruba alone or partly direct from Hausa.
I’ve heard unequivocally-Hausa discourse particles in Nigerian Pidgin, at least in Kano, so direct influence seems entirely possible to me.
I was often surprised to find southern Nigerians who knew quite a bit of Hausa; the situation is evidently not like Ghana, where Hausa is very much associated almost entirely with the north, with Islam, and with the zongos (Muslim trading quarters in all the big towns and cities, including in the south. One of the hospital drivers I knew in Ghana was a L1 Hausa-speaker from Kumasi.)
Seems to me that the OED staff could do with a bit of a crash course on African linguistics.
I haven’t got Heath’s dictionary of Koyroboro Senni, but his grammar doesn’t list /x/ as a phoneme at all. He does say that other sounds turn up in words that are only partially assimilated, though (but he also says that none of them are common in Gao.)
Lameen will know (of course.)
Today on its rolling new entry list featuring World English words, the OED has otak-otak, ‘A Southeast Asian dish consisting of ground fish or other seafood mixed with spices and coconut milk, wrapped in banana or palm leaves, and cooked by steaming or grilling over an open charcoal fire’.
The OED’s etymology:
Really? That’s it? First of all, the general meaning of otak is ‘brain’. Note also in R.J. Wilkinson (1901) A Malay-English dictionary, p. 53, column A here : otak tulang ‘marrow’ (tulang ‘bone’; analogizing ‘brain’ and ‘marrow’ is well-known in Indo-European; for convenience, see here in the Wiktionary). These fish cakes are soft and spongy like brains (sesos de cordero in tacos, for instance). And the cakes come out of their leaf package and can be likened to brains out of the skull, or cooked marrow out of a bone. Hence the name otak-otak, it would seem, with the reduplication indicating similarity. (Some other examples of such reduplication of nouns: orang ‘person, human being’, orang-orang ‘puppet, doll; scarecrow’; jari ‘finger’, jari-jari ‘spoke; radius’; langit ‘sky’, langit-langit ‘bed canopy; ceiling; palate, roof of the mouth’ (love that metaphor ‘palate’ = ‘sky’; also in Kurdish).
The Proto-Malayo-Polynesian antecedents of Malay otak are not exactly clear and involve disjunct reconstructions. The 2023 version of the Blust’s Austronesian Comparative Dictionary Online puts the Malay with *qutek ‘brain, marrow’, but this exists beside PMP *hutek. (R.L. Trask’s gloss of disjunct: “The term used by Blust (1970, 1980) for either of two reconstructed forms which are supported by overlapping cognate sets. For example, Fijian kumi ‘beard’ can be derived either from Proto-Austronesian *gumi or from Proto-Austronesian *kumis, both of which are independently required.”)
palate, roof of the mouth’ (love that metaphor ‘palate’ = ‘sky’; also in Kurdish)
Until I saw a disambiguating note in a French-language dictionary the other day, I don’t think I’d quite registered that “palace” and “palate” are both the exact same palais in French. The potential for confusion is clearly enormous.
Also:
https://en.wiktionary.org/wiki/palatum#Latin
In my father’s palace there are many places of articulation. He rents them by the hour.
love that metaphor ‘palate’ = ‘sky’; also in Kurdish
Russian has that as well, нёбо “palate” is the expected Russian outcome of PSl. nebo, while небо “sky, heaven” without the development of /e/ > /o/ before non-palatalized consonant must be due to Church Slavonic influence.
is often the only word spelled with ё in scientific papers; all others are left to context.
I thought about this pair of words when I was a child.
But I didn’t know that е is the ecclesiastical reading and ё is the vernacular reflex (совершенный “perfect” совершённый “done”, Алексей dim Алёша).
Njobo is great- I would have defined it as the (neuter!) state of abstinence from sex.
I think it implies an adj. *jobyj, jobaja.
Though when I was 20, a [male] friend of mine frequently said about this or that young lady that she is jebabelnaja.
Лёша is the name of another friend of mine. When I speak about him to my (or our) foreign friends in text, I use “Lesha”. And of course when I say his name, I say it with jo.
Leads to confusion (which is a good thing).
This is off-topic even by Hattic standards; justified (if you can even call it that) only by the fact that we’ve been discussing obscure etymologies.
But …
Anyone (by which I mean Xerîb and Lameen, but anyone can play) have any bright ideas about the origin of Kusaal takoro “window”?
The dictionaries, FWIW (not a lot) say Hausa taga, which comes, probably via Kanuri, from Arabic طَاقَة, itself (interestingly) from Iranian, if Wiktionary is to be believed, and ultimately related to Latin toga and Welsh ty etc etc.
Now, certainly it looks as if takoro must be connected with the Hausa word somehow, but it’s clearly not a direct loan (the dictionaries handwave vaguely in the direction of the Hausa plural tagogi, which I reckon doesn’t actually help at all.)
The k is odd, as is this -oro bit.
The word in that form is naturally not confined to Kusaal, but doesn’t seem all that widespread geographically.
Mampruli has takɔrɔ, Dagbani takɔro, Dagaare tókóróó. Farefare has at least made the effort to fit the word into the noun-class system, with takolle, plural takɔla. Moba has tákóbónn̂, where bònǹ is “hole, orifice.”
Farther afield, Kasem has it, as təkoro, and Dendi Songhay has tɔ̀gɔ̀rɔ́. Bisa has tokoreyɛr, where yɛr is “eye, hole.”
So the *takoro form is a local Wanderwort; the Dagbani reflex with r suggests it’s more recent than Songhay loans like bilichina “noble person.”
But I’ve no idea where it comes from, or how (if at all) it’s connected with Hausa taga.