Reading Leslie Jamison’s NYT Magazine article on Turkish baths (hammams), I hit the following sentence (I have reproduced the italics from the printed version; online, they have gone missing):

A woman named Gamze rubbed down my body with the kessa, a rough glove made from woven goat hair, and then draped my raw skin in the cascading bubbles of the swinging torba, a fine mesh towel dipped in copper tubs of olive-oil soap to heap shimmering white hills along the knobs of my spine, feathery and fizzy against my scrubbed skin, silken and gentle where the kessa had been vigorous and bracing.

It was obvious to me at a glance that “kessa” was not a Turkish word, in that spelling at least, and I grew provisionally irritated but of course had to investigate further. It turns out that 1) the Turkish word is kese (from Persian کیسه‎ [kise], ultimately from Akkadian ???????????????? [kīsu]), and 2) the usual English rendering seems to be “kessa,” as in the article. Which is understandable, because “kese” looks to an English speaker as if it should rhyme with “cheese,” but presents a problem: is it an English word, in which case it should not be italicized, or a foreign term that happens to be spelled in a different way than it is spelled in the original language? I suppose the latter, since it’s not in any of my English dictionaries nor in the OED, which mentions the Turkish word only s.v. purse:

In sense 4 [“In the Ottoman Empire: a specific sum of money, a certain number of piastres; spec. a unit of account equal to about 500 piastres”] after French bource (1665 in this sense, in the passage translated in quot. 1687), itself after Ottoman Turkish kise, (in later sources) kese, literally ‘purse’, denoting a unit of account (see note at definition; Turkish kese; < Persian kīsa or its etymon Arabic kīs (now kīsa; reborrowed < Turkish), both in sense ‘purse, bag’).

(Note that they take the Persian word from Arabic, whereas Wiktionary derives it from Akkadian; perhaps Xerîb will weigh in on this.) I’m provisionally withdrawing my irritation about the italics and deciding once again that language is endlessly confusing. I am, however, still irritated that they leave the essential cedillas off the name of the Çemberlitaş Hamamı, spelling it “Cemberlitas.” (I have had baths both there and at the NYC Russian and Turkish Baths she mentions at the start of the piece, and they were splendid experiences.)


  1. This reminds me of the restaurant in DC named Zaytinya. For a long time I pronounced it with /aɪ/ in the first syllable, until I discovered it was a phonetic spelling of zeytinyağı,Turkish for olive oil.

  2. When a writer introduces a word that it’s reasonably expected will be a new one to the reader, some sort of typographical signal is justified, even if not obligatory. Whether it’s technically a foreign word, or a foreign-origin loanword not widely known in English (with or without a changed spelling from the source language), or a within-English coinage not widely known, shouldn’t really matter. If you are following a stylebook that claims these differences are crucial or dispositive of whether you should italicize or otherwise mark it specially, burn your stylebook and get a better one.

  3. Nişanyan mostly agrees with Wiktionary, although it has Arabic and Persian acquiring the word independently from Aramaic:

    ~ Fa kīse كيسه torba, özellikle para torbası, cenin kesesi ~ Aram kīsā כיס a.a. ≈ Akad kīsu para kesesi
    Not: Ar kīs (a.a.) Aramcadan alınmıştır. Erm ksag քսակ (a.a.) Orta Farsçadan alınmıştır.

    The Aramaic intermediary would explain how Persian wound up with “e” (or “a” according to the OED) as the final vowel. By the time the Persians arrived, final short vowels were no longer pronounced in Akkadian, so they would’ve been confronted with either “kīs” or a learned pronunciation of “kīsu.” The Persian form would reflect the Eastern Aramaic definite suffix “-ā” that was eventually semantically bleached into a singular noun marker.

    The “purse” meaning is first attested in 1303 in the Codex Cumanicus and the “hammam implement” sense in 1665 in Evliya Çelebi’s Seyahatname.

    The Chicago Assyrian Dictionary defines “kīsu” (alternative forms, “kēsu,” “kīšu”) as a “leather bag for stone weights and for a merchant’s silver.” The “bag for stone weights” may have been the primary sense, later enlarged to cover coins, since it was often spelled logographically KUŠ.NIG2.NA4, reflecting the Sumerian equivalent “niŋ-na,” “thing (for) stones,” with KUŠ as a determinative for “leather.”

  4. Thanks!

  5. Andrej Bjelaković says

    Of course, both ‘kesa’ and ‘torba’ are everday words in BCSM. They both mean a bag, with ‘kesa’ being used for plastic or paper bags, and torba being the general word for a bag (e.g. a leather bag, etc.).

  6. Keith, you could do worse. I read a few days ago in a respectable Israeli paper an article on India (which many Israelis have visited for years), which mentions the city of לוקנוב /luknov/, i.e. Lucknow.

  7. @Andrej Bjelaković. What is BCSM?

  8. (Bosnian,) Serbo-Croatian (and MacedonianMontenegrin).

    [Ed.: Oops! –LH]

  9. It’s one less than BDSM

    (I’m going to pretend hat’s post isn’t there because I didn’t think of my joke in time)

  10. Incidentally, I just saw a reasonably convincing etymology of Russian dialectal дроля ‘beloved’ (in youth slang, doubtless now obsolete, ‘fool, oaf’): from SCr. (or, if you will, BCSM) drolja ‘whore.’

  11. Well, if the Croats had chosen the Roman Imperial name of Dalmatia for their shiny new country in 1918, things would have been jusssst-perfect.

  12. PlasticPaddy says

    Sorry jc, dalmatian is already in use for another language:

  13. ‘torba’ used in everyday Russian to mean a tote bag or satchel…

  14. Terry K., Hat: I believe BCSM stands for “Bosnian, Croatian, Serbian, Montenegrin” (Macedonian, like Slovenian, is a separate language from BCSM). I have also seen BCSM referred to as FYLSC (Former Yugoslav language of Serbo-Croatian).

    John Cowan (Echoing PlasticPaddy’s comment above): If “Dalmatia” had become the name for the country inhabited by (mostly) Catholic speakers of BCSM, it would have made things rather confusing: how would one have differentiated the dominant Slavic language of the country (what we call Croatian) from the extinct Romance varieties we call “Dalmatian”?

    (Oh, and by the way, recent scholarship has established that there are no grounds for using a label such as “Dalmatian” to refer to all the extinct Romance varieties once spoken in Central Europe and the Balkans along the Adriatic: it turns out that these Romance varieties shared a grand total of zero common innovations, so that technically “Dalmatian”, whether we take it to refer to a single language or to a subgroup within Romance, did not in fact exist. Yes, I can supply references to interested readers).

  15. Ahem. As the coiner of the acronym FYLOSC I should like to gently advise Etienne of the standard spelling while less gently congratulating him on helping expend its spread beyond the usage of the enthusiastic-but-lonely early adopter David Marjanović. The inspiration for the name is or was standardly given as FYROM, not FYRM. Which of course can serve as a reminder that the amount of confusion inherent in using “Macedonian” to refer both to a now-extant Slavic language and a now-extinct non-Slavic language previously spoken in some of the same geographical region is apparently not intolerable.

  16. David Marjanović says

    The M in BCSM is Montenegrin, not Macedonian!

    Croatia is traditionally “Croatia, Dalmatia and Slavonia”; extending “Dalmatia” to the whole would have been met with considerable resistance.

  17. *slaps self*

    How could I forget little Montenegro down on the Adriatic Sea!

  18. David Marjanović says

    “Illyria” was actually in 18th-century use for varying parts of the region, and “Illyrian” for various parts of FYLOSC. But later, identity with the Ancient Illyrians became an important part of Albanian nationalism.

    (…to the point that Ilir is a common first name there, even though Albanian has an [y] that’s even spelled y.)

  19. In addition to the general meaning “bag, sack, purse,” the Persian word کیسه kise is also used for the rough glove used for scrubbing in the hammam (or shower–it’s a common enough item in Iranian bathrooms).

  20. i enjoyed that article, though it made me wonder when i’ll get to go back to the 10th St Baths (thanks, de blasio & cuomo, for setting aside your feud to ensure that open schools will feed the slowly building COVID spike)…

    but i was impressed that it managed not only to never mention the other russian baths of nyc (which share a structure and style with 10th St, which istanbul hammams do not), but even to avoid using any of the basic new york terms for them. it’s not a bathhouse, it’s either a shvitz or a banya, depending on when your family left odessa.

    earlier from yiddish (שװיץ – sweat [verb]) vs. later from russian (wiktionary gives ба́ня, cognate with bain, bagno, and other descendents of latin balneum).

    after looking at wiktionary, i’m wondering if anyone has thoughts about the polish/ukrainian/belarusian alternative: łaźnia / ла́зня, which i haven’t met in the wild…

    (i’ve only ever heard the other main nyc option, the korean version, be called its full name: Spa Castle)

  21. “Dalmatian”, whether we take it to refer to a single language or to a subgroup within Romance, did not in fact exist. Yes, I can supply references to interested readers.

    Yes, please.

  22. after looking at wiktionary, i’m wondering if anyone has thoughts about the polish/ukrainian/belarusian alternative: łaźnia / ла́зня, which i haven’t met in the wild…

    Interesting; I don’t remember running into it either in my many years in NYC, though I probably heard Ukrainians mention it in their мова without knowing what they were saying.

  23. Attn Rozele and others. The 10th St. baths are apparently still closed because they supposedly haven’t figured out a pandemic-compatible way to reopen, but the downtown banya on Fulton St. (less old-timey but definitely unassimilated post-Soviet, at least when I was last there which was probably no more recently than 2012) is per the website back open with limited capacity and certain features (like the really cold pool and the Russian restaurant) closed.

  24. If they’re really unassimilated, you can probably get into the closed things by speaking Russian and bribing the guy at the door.

  25. The Republic of Macedonia decided a couple years ago that NATO (and eventual EU) membership was more important than keeping their name, so they finally gave in Greek demands and changed their name to the Republic of North Macedonia, effective February of last year. It seems to be unclear what will happen to the language name, since the official “Macedonian” language of what is now North Macedonia has traditionally been considered a a FYLOSC-influenced Bulgarian dialect.

  26. PlasticPaddy says

    @rozele re laznja
    This goes back to a Proto-Slavic *lazьńa. The explanation I have seen is that early saunas had no door but were entered by a hatch you “climbed” through or a low entry/exit tunnel you “crept” through. But maybe someone has something better. Vasmer seems to reject borrowing from old Icelandic and connection with a Greek term λάσανον.

  27. Brett, well, they didn’t change the name of the country to the “Republic of Southwest Bulgaria,” did they? It’s actually sort of an interesting case study because you can believe that historically the notion of a distinct “Macedonian” Slavic ethnic identity was a cynical propaganda initiative directed from Belgrade aimed at undercutting local support for Bulgarian irredentism, but after a few generations it may have taken sufficient root among the populace as to survive the termination of rule-from-Belgrade. Apparently in the post-Cold-War dispensation you can subdivide former polities but you can’t expand them – Romania hasn’t even managed enosis (or however you say it in Eastern Romance) with post-Soviet Moldova.

  28. I still recommend the Danforth book I mentioned in this ancient post.

  29. how would one have differentiated the dominant Slavic language of the country (what we call Croatian) from the extinct Romance varieties we call “Dalmatian”?

    Where’s the problem? We manage somehow to distinguish the dominant language of Ireland from the language we (and they) call “Irish”.

    there are no grounds for using a label such as “Dalmatian” to refer to all the extinct Romance varieties once spoken in Central Europe and the Balkans along the Adriatic

    There are, of course, grounds geographic and historic, simply not linguistic. The Gurage people of Southern Ethiopia speak some unknown number of languages from three different small subfamilies of Southern Ethiopic all of which include non-Gurage languages, but both Gurage and non-Gurage consider their languages, but not their relatives, a single language called Guragē Af ‘Gurage language’ in Amharic.

    both to a now-extant Slavic language and a now-extinct non-Slavic language previously spoken in some of the same geographical region

    As well as a non-extinct variety of a non-Slavic language also spoken there, characterized by such features as the loss of unstressed high vowels (and the raising of unstressed mid vowels to replace them) and an American-style dark /l/.

    Croatia is traditionally “Croatia, Dalmatia and Slavonia”

    So if Croats are not Slavs, who are they?

    How could I forget little Montenegro down on the Adriatic Sea!

    I know it’s Gatsby, but it still sounds like the refrain of an old pop song.

    it’s not a bathhouse, it’s either a shvitz or a banya, depending on when your family left odessa.

    As Mr. K*A*P*L*A*N said when informed that the first name of “Moishe Elman, famous on fiddle” was in fact Mischa, it depends on who’s pronouncing. There have been and still are many men frequenting bathhouses in NYC who don’t use any other (formal) term for them. I would not wish to see them erased.

  30. So if Croats are not Slavs, who are they?

    I don’t understand the question. Unless you’re making a joke about “Slavonia”?

  31. David Marjanović says

    The Slovenes, Slovaks, Slavonians and Slovincians are those Slavs who never got a distinct name for themselves.

  32. @PP: thanks!

    @John: you’re absolutely right! and i would not wish any of their -lects to disappear.


    i happened to be in ohrid during the 2006 macedonian elections, and although the political parties seemed to be pretty strictly ethnic units, i’m not sure i’ve ever been in a place that gave me such a taste of everyday (post)ottoman cosmopolitanism. we rented a room from a local (who was home on shore leave from cruise ship work; her uncle turned out to live in astoria, queens, of course) who greeted almost everyone she passed on the street while taking us to the house. that involved three languages (and i have no idea whether there were multiple dialects at play) – albanian, (slavic) macedonian, and macedonian turkish – with many of the exchanges being multilingual. being a worldly person, she of course had other languages: istanbul turkish, english, and i believe one or two more. we should all get to live that way!

  33. I envy you and them!

  34. It turns out the question isn’t as dumb as I thought. From the WIkt page on *xъrvatъ, from which we get both Croat and cravat:


    Uncertain. Possibly from the Old Persian *xaraxwat-, attested by the Old Iranian toponym Haraxvaitī- [known to the Indians as Sarasvatī-], the native name of Arachosia [a Persian satrapy in modern southern Afghanistan now known as Kandahar Province, as its capital of the same name was founded by Man Who Names Cities After Himself].

    Common theories from the 20th century derive it from an Iranian origin, the root word being a third-century Scytho-Sarmatian form attested in the Tanais Tablets as Χοροάθος (Khoroáthos, alternate forms comprise Khoróatos and Khoroúathos).

  35. On historical grounds, I believe identification of Albanians with ancient Illyrians is a mistake.

    Everything we know about them tells me that they were a newly pastoral (of transhumance variety) people of the same origin as Vlachs.

    Roughly – sometime during post-Hunnic and Avar period, groups of refugees largely consisting of former Roman citizens fled to mountains and took up pastoral lifestyle (which was more appropriate for survival during Dark Ages). Some of them were Daco-Thracian speakers and some spoke Vulgar Latin (reflecting linguistic composition of the population in Roman provinces of Eastern Balkans).

    Eventually they colonized most mountainous areas in the Balkans and beyond.

    The Thracian speakers became Albanians, the Vulgar Latin speakers Vlachs and Romanians. Some of them switched to Slavic language which was lingua franca of the Avar Khaganate and ended up becoming part of modern Slavic nations (not just South Slavic, there are Vlachs and groups derived from them in Moravia, Poland, Western Ukraine, etc)

    As for Illyrian, I think it died out without leaving direct descendants. Perhaps even during the Roman period.

  36. Disappointingly the North German word for pillow – Kissen – is apparently a borrowing from old French that derives ultimately from Latin culcita (pillow).

  37. Andrej Bjelaković says:
    “Of course, both ‘kesa’ and ‘torba’ are everday words in BCSM. They both mean a bag, with ‘kesa’ being used for plastic or paper bags, and torba being the general word for a bag (e.g. a leather bag, etc.).”

    That would be the lumper view. Taking the splitter view gives a bit more nuance:
    – kesa and torba are not synonyms, at least not in Croatian. They might be in Serbian, who knows.
    – In Croatian vreća is the usual word for a bag or sack. Regionally, the Turkicism kesa is used.
    – vrećica is used for a grocery bag in Croatian
    – Torba is the usual word for a handbag in Croatian. This is from Turkish. Another word for this is taška, regionally – tašna which is from German.
    – in Bosnian, apart from kesa, there is also the form ćesa. This reflects more accurately the Turkish pronunciation. Ćesa was also a unit of account, which changed over time, decreasing from 30,000 akçe or 10,000 ducats in the time of Mehmed the Conqueror to 500 kuruş (piastres) in later times.

    John Cowan says: “Well, if the Croats had chosen the Roman Imperial name of Dalmatia for their shiny new country in 1918, things would have been jusssst-perfect.”
    Illyrian and Dalmatian were used for Croatians and the Croatian language since at least the renaissance, if not before.

    In Latin, it was quite common for classical names to be used, especially for languages: Batavian or Belgian for Dutch, Lusatian for Portuguese, Etruscan for Tuscan. And this still continues in English with Spanish for Castillan and German for Deutsch.

  38. The BMAC is something entirely other: Bactria-Margiana Archeological Complex. I think there’s a museum in Jászberény (Hungary) containing relics of the Iazygians, who were Sarmatian and therefore Iranian, and who might have contributed language elements to sundry wandering Slavs. They were certainly still around, stubbornly, into the Middle Ages. Connecting the Croats to Sarasvati is such an astonishing reach that the mind boggles.

  39. There’s discussion of BMAC in this thread from a couple of years ago.

  40. I’ve also read that “Hrvati” is “Carpathi” plus Grimm’s Law. There were Chrobati in southern Poland at one time, I seem to recall. What the Germans have to do with it, I don’t know.

  41. Lusatian for Portuguese

    Surely Lusitanian?

  42. Surely Lusitanian?

    Apologies for the typo

    Re etymology for Croat: One thing that is clear is that it is NOT clear. The mainstream view seems to be that it is Iranian.
    Apparently there are place names with the word hrvat from Ukraine up to Poland & Czech and all the way down to the Peloponnese.

  43. David Marjanović says

    I see your West-Eurasia-wide Iranian Croats and raise you Eurasia-wide Iranian Serbs.


    apart from kesa, there is also the form ćesa. This reflects more accurately the Turkish pronunciation.

    Well, in some kinds of Turkish, /k g/ in front-vowel words become [kʲ gʲ] (…just like in mainstream Greek *diving under table*). Those seem to have been equated with apparently identical sounds present in central Macedonian, and those are cognate with the alveolopalatal FYLOSC affricates which are used in a lot of Turkish borrowings, e.g. köfte (sg.) > ćufte (f. pl.).

    They were certainly still around, stubbornly, into the Middle Ages.

    Those were different Iranians – Alans/Ossetes.

    I’ve also read that “Hrvati” is “Carpathi” plus Grimm’s Law.

    But without Verner’s, and I can’t see how to explain the first vowel (ъ).

  44. Man Who Names Cities After Himself

    Speaking of which, I note that the modern names of Alexandria-in-Egypt include al-ʾIskandariyya in Modern Standard Arabic and the regular derivative Eskenderiyye in Egyptian Arabic, with corresponding names in most Wikipedia languages except in the Persianate zone, where like most things Alexander it uses the stem Iskender-. In Coptic, however, it has the unique name Rakodi < Egyptian rꜥ-qdy.t ‘the built-up [area]’, referring to a fishing village that eventually became the Egyptian quarter of the ancient city.

  45. Respect!

  46. Вulgarian “кесия” is a normal derivation from (early) Ottoman Turkish which means “purse” (one that is made of leather/cotton, small in size, an made to contain coins) and торба for “bag” (a similar container, bigger in size, made to contain not-coins). The latter is even identical phonologically.

  47. torba is also the Polish word for “bag”.

  48. -ия is the normal Bulgarian ending from Turkis that end in [ɯ], at least the ones from the 17/18th centuries.

  49. al-ʾIskandariyya in Modern Standard Arabic

    I what they make of Al-‘Aska and Al-‘Abama

  50. David Marjanović says: “raise you Eurasia-wide Iranian Serbs”

    “Srbija do Tokija”

  51. In Tatar, кесә means only ‘pocket’ and, by extension, ‘money’.

  52. The words for “pocket” in Croatian and Hungarian are džep and zseb respectively, also deriving from Turkish.

  53. Also Greek τσέπη; all from Ottoman Turkish جیب‎ (ceb, cep), from Arabic جَيْب‎ (jayb). (See the link for a bunch of other words derived from the Turkish.)

  54. Also (not included there):
    Adyghe: джыбэ (ǯ̍əbă)
    Afar: gib m
    Aromanian: gepi f, buzunar n
    Assamese: জেপ (zep)
    Erzya: зепе (zepe)
    komi zürjén: зеп (zep)
    Midob: céeb
    Moksha: зепе (zepe)
    Ossetian: дзыпп (ʒypp)
    Somali: kish m, jeeb m
    Tajik: кисса (kissa), киса (kisa), ҷайб (jayb), ҷайбак (jaybak)
    Telugu: జేబు (te) (jēbu)
    udmurt: ӟеп (dźep)
    Wolof: jiba (wo)

  55. My first thought was “How did Telugu wind up with it?” but duh, it’s from Hindi/Urdu जेब/جیب‎, from Persian جیب‎ (jeyb).

  56. Also Bulgarian джоб (job/dzhob) (pocket). Both джоб and кесия are likewise metonymic for “the amount of money one owns” like Tatar кесә. “джобни пари” means literally pocket money, but you can pay out of your own джоб (pocket), which is definitely not pocket money. And your кесия is, metonymically, a more abstract concept of the amount of money you have in general.

  57. Telugu speakers were ruled for centuries by Nizams of Hyderabad, Muslim dynasty whose ancestors came to Mughal India from Samarkand.

    So I imagine Telugu was greatly influenced by Perso-Arabic vocabulary.

  58. Midob… there’s one that didn’t register on sight. One of the Nubian languages apparently.

  59. In answer to Y’s request upthread: the reference (on there not having existed a Dalmatian language or subgroup):

    Chambon, Jean-Pierre. 2014. “Vers une seconde mort du dalmate? Note critique (du point de vue de la grammaire comparée) sur ‘un mythe de la linguistique romane'”. Revue de linguistique romane 78: 5-17.

    Incidentally, I *may* have found evidence (involving a borrowed Romance morpheme with a shared idiosyncratic meaning both in a northern variety of BCSM and in Albanian) that “Dalmatian” is in fact a legitimate term if we take “Dalmatian” as having been a language area within the Romance continuum, whose shared features (such as the odd semantic shift of the Romance morpheme mentioned above) are due to language contact between genetically distinct Romance dialects spoken along the Eastern shore of the Adriatic rather than to descent from an alleged “Proto-Dalmatian”. So do not take the above scholarly article as the last word on the subject…

    SFReader: there are good linguistic grounds for doubting the affiliation of Illyrian and Albanian too, which I presented here at the Hattery: see my December 9 10:34 PM comment at this fine thread:

  60. Thanks for reminding me of that very interesting thread! Here‘s a direct link to the comment in question.

  61. Thank you, Etienne. (This page has a link to that publication.)

  62. David Marjanović says

    No, it’s several more clicks away; I finally found it here after getting lost on the site you linked to.

  63. Dalmate: un terme qui reste à interroger.” How French!

  64. I believe the word ‘budget’ also originally refered to a purse-bag for carrying travel money.

  65. PlasticPaddy says
  66. Trond Engen says

    Why is “node in a phylo-genetic tree structure” or “reconstructible as a proto-language” necessary to make Dalmatian a useful (or “valid”) term in Romance linguistics? The phylogenetic tree structure is a simplification that works well for great time depths and distinct entities but badly for short timespans and linguistic continuums. If the Romance dialects spoken in Dalmatia shared traits that made them distinguishable as such to contemporary speakers, or to us today, that should be enough. What probably didn’t exist was a Dalmatian Dachsprache or common prestige dialect that gave the local dialects a shared center of gravity distinct from those of other Romance varieties.

  67. Trond: Oh, I quite agree that “Dalmatian” as a term need not refer to a genetic subgroup: it could indeed refer to an areal grouping. The problem, though, is that there is zero evidence for Dalmatian having existed either as a genetic subgroup or as an areal subgroup (leaving aside the one piece of evidence I may have found, as I mentioned above). And there “certainly” did not exist a Dalmatian Dachsprache: recent scholarship on words of “Dalmatian” origin in coastal Croatian dialects, relating chiefly to fishing and related activities, has revealed a degree of lexical diversity that is utterly incompatible with any postulated pan-Dalmatian prestige language/dialect (doubly so since, considering the insular nature of the area, we would expect maritime terminology to be one of the first to be affected by any incipient linguistic unification).

    Hat: Chambon’s writing is indeed very French in style, I agree, but unlike all too many French academics, there is at the core of his article a data-based argument (namely, that Ragusan and Vegliote, the two best-known “Dalmatian” varieties, do not have a single common phonological innovation).

    Some might like this article on the topic (in Catalan, but the maps are quite telling): one of its conclusions (page 183), that the highest number of original (i.e. alien to all other Romance varieties) lexical items is found in Southern Dalmatia, does seem to fit with the claim that Southern Dalmatia was the second Romance variety to break away from the Romance continuum (Sardinian being the first):

  68. Etienne, how unusual is it for someone to publish a paper on Romance historical linguistics written in Catalan, when it’s neither the focus of the paper nor the writer’s native language?

  69. Y: Not THAT unusual, inasmuch as a fair amount of work on circum-Mediterranean languages/language contact has been written in Catalan, which is appropriate since historically Catalan played a major role in Mediterranean linguistic history (I suspect the rise of Catalan nationalism has something to do with research on the topic and/or its being written in Catalan, what with its being much more of a Mediterranean language than Castilian). But it is not all that common, I will grant you that.

  70. David Marjanović says

    a fair amount of work on circum-Mediterranean languages/language contact has been written in Catalan

    I am absolutely amazed.

    But then, it really doesn’t matter for accessibility, does it. If you can read one western Romance language, you can read them all.

  71. Trond Engen says

    I’m reminded of Mikko Heikkilä’s thesis on the linguistic pre-history of Scandinavia written in Swedish,

  72. David Marjanović says

    That’s less amazing than a whole field of research just being in Catalan, even though Heikkilä’s thesis is about the linguistic prehistory of Finland and not of Scandinavia.

    Also, I find Catalan and, as I’ve been experiencing lately, Czech easier to read than Swedish. Four years of Russian in school, ending 20 years ago, are enough to make Czech at least as tractable as Swedish is after being fluent in the two most divergent West Germanic languages, one of which has left plenty of loans and calques in Swedish.

    (What I’ve been reading in Czech, slowly, is a MSc thesis that redescribes a bunch of important fossils and casually reports a bunch of surprises.)

  73. David: about your statement-

    “If you can read one western Romance language, you can read them all”

    -I’m afraid this is only true if you are either a historical linguist specializing in Romance historical linguistics or at least familiar with Latin and with the major phonological changes leading to each Romance language. In either case being a native speaker of a Western Romance language is very helpful, of more than one even more helpful (I once knew a doctoral student, bilingual and biliterate in French and (European) Portuguese, who despite being a literary scholar with no knowledge of or interest in of linguistics found reading Spanish easy but needed my help with Italian), and even then it will prove necessary to learn some basic vocabulary and grammar (especially involving words/morphemes not widely shared within Romance).

    In this light I found it very telling that the LRL (Lexikon der romanistischen Linguistik) makes use of the four major Romance languages (along with German) in its various articles, plus Galician, but excluding both Catalan and Romanian. I think the underlying logic is that anyone who can read Spanish and Portuguese can read Galician (which is true, if my experience is any guide: even in this case, though, some aspects of Galician spelling and grammar do need to be learned directly), whereas Catalan and Romanian are much more challenging (too challenging, apparently, according to the LRL Editors) to speakers/readers of other Romance languages. Even ones who could be expected to know the basics of historical Romance linguistics (which is certainly the case with the expected readership of the LRL).

    Incidentally, my own experience is that Romanian is sometimes easier to read than Catalan: but as a native francophone I am biased, so to speak, since Romanian has been so heavily gallicized in its more bookish registers that at times it can give the impression of simply being French with phonetic spelling and exotic morphology.

    Oh, and on the topic: I have been slowly but surely reading a book on linguistic matters written in Haitian Creole, and what I wrote above (on needing to know the basic vocabulary and structure) is even truer, by and large: if I hadn’t already learned (via the linguistic literature) the basics of the structure of the language (the tense + mood + aspect markers, plural marking, the definite marker especially) I doubt I could have deciphered a single paragraph of the book. As it is I have managed to read several chapters…

  74. Jen in Edinburgh says

    I’ve never tried reading Romanian, but I discovered in museums in Barcelona that I find it easier to read Catalan than (Castillian) Spanish, despite speaking a bit of the latter and none of the former…

  75. Jen in Edinburgh: what I suspect is happening is that some specifically Castilian word choices/phonological features which Catalan lacks make the former more opaque to you than the latter. This reminds me of a fact written about by a linguist: apparently Catalan-speaking tourists in Southern Spain often mistake local (pre-Castilian Romance, AKA Mozarabic) place names for Catalan ones: the similarity they perceive typically involves the absence, in these place-names, of certain Castilian features (i.e. preservation of Latin F in Mozarabic and Catalan, versus its aspiration/loss in Castilian, for example) unshared with Catalan and Mozarabic (which needn’t imply any special relationship between these non-Castilian varieties, of course, beyond the obvious fact they are Romance).

  76. David Marjanović says

    I’m afraid this is only true if you are either a historical linguist specializing in Romance historical linguistics or at least familiar with Latin and with the major phonological changes leading to each Romance language.

    Well, it’s all for a certain value of “read”. I read the article you linked to – but I didn’t quite understand every sentence, and I got 2/3 through before I remembered that the preposition amb means “with” (I was guessing “around” or something). If I really needed to work with that paper, I would at least fire up Google Translate once every few sentences. And then, of course, it’s a scientific paper, so the topic is restricted and the vocabulary is more international than the rest of the language. I’m sure I could not meaningfully read literature or an opinion article in a newspaper.

    I don’t think Latin helped a lot. It made ou easier to recognize as “egg”, but that’s irrelevant to what the paper is about.

    I’m sure I’d find the French-based creoles much harder because, as you say, their grammar is built from scratch out of words which are sometimes reduced beyond recognition, sometimes misleading because they’re ordinary lexical words in French, and they sometimes express TAM categories that simply aren’t known in Europe, so I wouldn’t expect them to exist.

    My Czech reading experience is similar to my Catalan one: I understand most sentences well enough for my purposes. For instance, when I encountered vrstva in the geological part, it had to be related to Polish warstwa “layer” (I don’t even know of there’s a Russian cognate), and the derivative souvrství looks like it could mean “formation“; I don’t need to understand the geology of that coal mine in any detail, so I left it at that and just moved on.

    Swedish is harder: the lexical and phonological divergence from both German and English (spelling) is wide enough that more words are unrecognizable.

    Where historical linguistics is really important for me is reading Dutch: miss the important sound changes (e.g. the High German consonant shift) once, and you drown in false friends and nothing makes sense anymore. With Catalan I can coast much more on superficial similarity to French (spelling).

    BTW, Nikola Vuletić has also published a few papers in Galician – here’s one on fraseoloxía dálmata, so he published it in the Cadernos de fraseoloxía galega, logically –, in addition to French, Italian, Castilian, English and Croatian.

  77. For instance, when I encountered vrstva in the geological part, it had to be related to Polish warstwa “layer” (I don’t even know of there’s a Russian cognate)

    Yes, верста ‘verst, milestone’ (from Proto-Slavic *vьrtěti, from Proto-Indo-European *wert-); see here.

  78. That is a great link LH. Though, it saddens me to see the results of Wikipedia’s denialism of Croatian. I did a double take at the “Serbo-Croatian” meaning.
    First they leave out the Croatian meaning.
    Second they do a disservice to Serbian by labelling it Serbo-Croatian.
    Third, they do a disservice to anyone who might get the impression that Croatian is written in cyrillic.

    The definition in the Wikidictionary should be enhanced by inserting the Croatian word “vrsta”, meaning sort, kind, species, rank; and also: verst, Russian unit of length.

  79. There is also vârstă, which is the Romanian word for “age” (borrowed from OCS).

  80. David Marjanović says

    …I knew the unit of length but had no idea how to make the semantics work! This is like clean < “neat” > klein “small”.

  81. David Eddyshaw says

    I once managed to make sense of a Catalan linguistics thesis fairly well, aided chiefly by the written resemblance to French, but most of all, of course, by the fact that the genre was so familiar anyway, so that I was generally going not “What on earth does that mean?”, but “Aha! That’s how you say it in Catalan, then.” (I don’t know Castilian at all.)

    It’s not very difficult to read Dagbani or Mampruli on the basis of knowing Kusaal, though this judgment may be somewhat skewed by that fact that there isn’t a lot available to read apart from the Bible, and I know what that says already. Mooré is much harder, though, and Dagaare (the Romanian of Western Oti-Volta) pretty much impossible.

  82. Note that they take the Persian word from Arabic, whereas Wiktionary derives it from Akkadian

    The Persian form would reflect the Eastern Aramaic definite suffix “-ā”

    One point to note in this regard is that final Persian ه -a often goes back to Middle Persian -ag, from earlier Iranian *-aka-, and this is probably the case in کیسه kīsa. (For LH readers who aren’t familiar with Persian orthography and phonology: final Persian ه -a in words like کیسه kīsa is pronounced [e] in standard contemporary Iranian Persian. This pronunciation is an innovation. Dari preserves the original lower sound, /a/ (often [ɐ]), as in this sound file containing kīsa, as does Tadjik with /a/ in киса and its variant кисса.)

    This origin of final -a in *-aka- is most evident in plurals of animate nouns, such as زاده zāda “child”, with plural زادگان zādagān, in which g, lost in final position in the singular, has been preserved before the plural ending. Operating mechanically from kīsa we can restore a Middle Persian *kīsag. Such a reconstruction is supported by Armenian քսակ k῾sak “pouch, sack, purse; money”, doubtless a borrowing from a Middle Iranian source akin to the source of New Persian. (The loss of the ī, an original high vowel in an unaccented syllable, is regular in the history of Armenian.)

    In this regard we can also mention the form in Balochi, a language which is archaic in many respects despite its late attestation. Balochi has kissag with -ag. To my knowledge, it is an open question whether the regular outcome of Proto-Iranian *-aka- is -ak or -ag in Balochi—or perhaps it is both, and there are subtle conditioning factors. If the regular outcome is -ak, the suffix -ag could have entered pre-Balochi as part of Middle Persian loanwords and then enjoyed a certain productivity within Balochi, perhaps continuing in the adoption of loanwords from New Persian. (The double -ss- in kissag is due to a kind of quantitative metathesis, akin to the Latin littera-rule. This is a widespread phenomenon across Balochi dialects: جُرّاب jorrāb beside Persian jōrāb “stocking”; کونّ kunn “ass” beside Persian kūn; etc., etc.).

    Here are some other examples of Middle Persian -ag corresponding to Persian -a, Balochi -ag, and Armenian -ak in Iranian loanwords to show the network of correspondences that Persian kīsa fits into:

    Middle Persian dastag written ⟨dstk’⟩ “bundle (bunch grabbed in the hand)”; Persian دسته dasta “handful, shock, bouquet; handle, haft” Balochi دستگ dastag “handle”; Armenian dastak դաստակ “handle”.

    Middle Persian mādag written ⟨mʾtk’⟩ “female”, Persian ماده māda; Balochi مادگ mādag “female; heifer”; Armenian մատակ matak “mare”

    Middle Persian tābag written ⟨tʾp̄k’⟩ “frying pan”, Persian تاوه tāva and تابه tāba “frying pan”; Balochi‎ تاپگ tāpag “griddle, baking stone”; Armenian տապակ tapak “frying pan, skillet”

    However, it is not the case that Aramaic nouns were all automatically fitted out with Iranian *-aka-, Middle Persian -ag, as part of their adoption into Iranian, as *kīsag was. This can be observed in an inventory of some Middle and New Persian words of Aramaic origin in Shaul Shaked (2009), “Aramaic Loan Words in Middle Persian”.

    As far as the ulterior etymology of Hebrew כִּיס kîs “bag, purse” and Aramaic כיסא kīsā, it is commonly accepted that they are loanwords from Akkadian, whatever the ultimate etymology of the Akkadian word is. Arabic kīs would be a borrowing from Aramaic, and Geʽez ኪስ kis “purse, pocket” would be a borrowing from Arabic. According to Akkadische Fremdwörter als Beweis für babylonischen KultureinflussZimmern, the noun would have diffused from Akkadian through commercial contacts—a very plausible scenario. This view is maintained by the Hebrew and Aramaic Lexicon of the Old Testament (1996) and by Wolf Leslau in his Comparative dictionary of Ge‛ez (1987), for instance.

    However, an interesting new addition to the dossier of the family of Akkadian kīsu was made recently by Richard C. Steiner in his monograph, The Nefesh in Israel and Kindred Spirits in the Ancient Near East, with an Appendix on the Katumuwa Inscription, which overflows with philological delights. I recommend it to anyone who can read a few Hebrew words when cited without transliteration. Steiner takes up the difficult verse Ezekiel13:18:

    וְאָמַרְתָּ כֹּה־אָמַר אֲדֹנָי יְהוִה הֹוי לִֽמְתַפְּרֹות כְּסָתֹות עַל כָּל־אַצִּילֵי יָדַי וְעֹשֹׂות הַמִּסְפָּחֹות עַל־רֹאשׁ כָּל־קֹומָה לְצֹודֵד נְפָשֹׁות

    wə’āmartā kōh-’āmar ’ăḏōnāy YHWH hōwy limṯappərôṯ kəsāṯôṯ ʽal kol-’aṣṣîlê yāḏay, wəʽōśôṯ hammispāḥôṯ ʽal-rōš kol-qômāh ləṣôḏēḏ nəp̄āšōṯ

    rendered in the KJV as

    And say, Thus saith the Lord GOD; Woe to the women that sew pillows to all armholes, and make kerchiefs upon the head of every stature to hunt souls!

    On pages 31ff, Steiner, has proposes keset, here rendered “pillow” reflects earlier k&imacron;st, a feminine corresponding to masculine which became *kist by shortening in this position, and then keset by the regular change seen in numerous Hebrew segolate nouns:

    The view of Ezekiel’s כסתות [kǝsātȏt] as pillow casings (rather than pillows) yields an etymology far better than the ones suggested by modern biblical scholars. The etymology is hinted at by Joseph Qara’s gloss [as “pouches”], cited above, and by the suffixed form כִּיסָתוֹ (rather than the expected כִּסְתּוֹ ) that appears in one early vocalized manuscript of the Mishnah. These two pieces of evidence suggest that כֶּסֶת [keset] < *kistu is nothing other than the feminine form of כִּיס [kîs] < *kīsu, with the expected vowel shortening in a closed syllable. In other words, the vowel alternation in כֶּסֶת ~ כִּיס has the same origin as that in שַׁחַת ~ שִׁיחָה , in מַזְכֶּרֶת ~ מַזְכִּיר , in גְּבֶרֶת ~ גְּבִיר (pausal גְּבָרֶת ), in ~ אַדִּיר אַדֶּרֶת * (pausal אַדָּרֶת ), and probably in קֶרֶת ~ קִיר מוֹאָב * (pausal קָרֶת ). It is true that one would have expected the plural to be כִּיסוֹת (cf. שִׁיחוֹת and מַזְכִּירוֹת ) instead of כְּסָתוֹת , but there are other examples of the feminine ending -t being incorporated into the root by metanalysis; Gesenius compares דְּלָתוֹת and קְשָׁתוֹת to which we may add שִׂפְתוֹת.

    Although Steiner does not mention it in this connection, the Modern South Arabian languages offer some interesting data as well. In his Mehri Lexicon and Mehri-English Word-List, T.M. Johnstone notes Mehri kəst, “bag” alongside eastern Jibbali kist and western Jibbali sikt, presumably metathetic. These have a feminine -t too… I do not know whether scholars know enough about the adoption of Arabic loanwords in the Modern South Arabian language to rule out the hypothesis that kəst could be from the colloquial Neo-Arabic feminine kīsa, construct kīsat- (the Wortatlas der arabischen Dialekte under “Sack” gives a Bahraini variant čīsa, for instance).

    I’ll leave others to draw inferences on the possible Proto-Semitic status of *kīsu in light of these data.

  83. Very interesting indeed; thanks for doing all that research!

  84. January First-of-May says

    There is also vârstă, which is the Romanian word for “age” (borrowed from OCS).

    Whence (the OCS, not the Romanian) presumably also Russian сверстник “age-mate, peer”.

  85. ə de vivre says

    Would there be to tell if the attested forms of kīs were common inheritances or loans from Akkadian? It looks like PS *k and *s are pretty stable across the daughter branches so that there wouldn’t be much to distinguish between loaned and inherited forms.

  86. Lars Mathiesen says

    And now I suppose we are going to have the FFYROMLOM, pardon FNMLOM because Bulgaria.

Speak Your Mind