Slavophone refugees to Greece

To check further on Kızderbent, I got hold of the 2001 book Γλωσσική Ετερότητα στη Ελλάδα [Linguistic Otherness in Greek], to see what it said about Trakatroukika. The book is a transcription of a series of panels on linguistic minorities in Greece. Most sessions passed without incident, except for the Vlach session (which the organisers unwisely held in situ in Thessaly, instead of in Athens).

The session on "Slavic dialects of Greece" includes a four page presentation by Leonidas Empirikos (son of the surrealist writer), reporting on a survey he did with Lambros Baltsiotis on the geographical distribution of Slavonic in present-day Greece. Trakatroukika is mentioned on p. 155. I give the citation here, because Kızderbent turns out not to be the only settlement of Slavonic speaking refugees to Greece:

I should also note that this is a large contiguous region [where Slavonic is spoken]; by that I mean that it is a continuous population of rural origin, to this day. A survey of Slavophony in Greece would not be complete without mentioning at least five special cases:

1. The Slavonic-speaking Muslims of Western Thrace, who were not subject to the population exchanges and are called Pomaks;
   
2. A few people in some of the Nestus villages of Stavroupoli, Xanthi;
   
3. The "Trakatroukides" with refugee origins (Slavonic-speakers from Nicomedia in Asia Minor, settled here and there in various regions of Greek Macedonia and Western Thrace);
   
4. The few Slavonic-speaking Patriarchists from Eastern Thrace who chose to settle in Greece;
   
5. Slavonic-speaking Patriarchist refugees, mainly from Strumica (who settled in Kilkis), but also from Petrich, Startsevo and Nevrokop [= Gotse Delchev] (who settled in Eastern Macedonia, from New Petritsi/Veterna to Drama, Iraklia and Prosotsani).

View Greek Slavophone refugees in a larger map

The last couple of populations need explanation. Modern Bulgarian nationalism started out as an assertion of ecclesiastical independence, with the establishment of an autonomous Bulgarian Exarchate within the Orthodox church. Identity in the Ottoman Empire was credal under the Millet system, and the local bishop administered the Orthodox Christians; if Bulgarians wanted to not be Greek under that framework, they had to have their own bishops. This led to contention between Exarchists, aligned with the newly autonomous Bulgarian church, and Patriarchists, remaining loyal to the Greek Patriarch.

Linguistic or ethnic identity did not necessarily align with credal identity—as in so many instances in the Ottoman Empire; so people could speak Bulgarian (or Macedonian Slavonic—or for that matter Aromanian), but be Patriarchists. I don't know of reverse cases, native speakers of Greek being Exarchists, but I wouldn't rule anything out. After all, not all the Greek-speakers of Eastern Rumelia left for Greece in 1919, either.

The dispute between Exarchists and Patriarchists turned violent in Macedonia, culminating in the guerilla conflict that Greeks call the Macedonian Struggle. But from Shishmanov's survey, the dispute had also spread to Asia Minor, with several villages reported as containing Grecomans (the Exarchist disparagement for Patriarchists: the term is intended to mean "wannabe Greeks"). We would in fact expect the dispute to have spread anywhere in the Ottoman Empire where Slavonic-speaking populations had a Greek clergy. There was no force field between Sofia and Nicomedia, to prevent people coming across to convince the locals they were Bulgarians, and not just Christians.

I think that's the subtext to surprise at Kızderbent that I perceive from the Bulgarian Wikipedia. Bulgarian researchers found out about the other Asia Minor villages in the 1860s, and went over to encourage their Bulgarian national sentiment; Kızderbent was known about in the West fifty years earlier, and a Bulgarian national sentiment does not seem to have taken root.

(Greeks reading this should not get too smug: blogger Doctor has posted at Sarantakos' blog reports from Greek officials in Eastern Thrace, annoyed that they too had to work on the locals' national sentiment, and convince the locals they were Greeks and not just Christians.)

So when time came for the population exchanges with Bulgaria and Turkey (and the preceding violence in Turkey), the criterion for who chose to go to Greece was not linguistic but ecclesiastical. Patriarchists, who were derided as wannabe Greeks, wanted to be Greeks, and chose to go to Greece. That applied to Bulgaria—Empirikos' fifth group. It also applied to settlers from Bulgaria and Macedonia in Turkey: Eastern Thrace, Empirikos' fourth group, but also the Trakatroukides, Empirikos' third group. From Shishmanov's survey, several other Bulgarian villages of Bithynia chose Greece as well, though they don't appear to have got attention.

I don't know anything about Empirikos' second group, btw, but populations did move around in the Ottoman Empire; there's a village of Albanian Christians in Serres prefecture (Koimisi, in Irakleia municipality), for example, who are refugees from Eastern Thrace—like Empirikos' fourth group. So the Stavroupoli Bulgarians would have moved there from elsewhere in the Bulgaro-Macedonian continuum.

...Read more

The Trakatroukides and their language

This is a complicated story no matter how one tells it and from what vantage point, but I've put this off a few months too many. So, following on from the preceding post on the Trakatroukides, let's take the story from the top. Again, thanks to Butcher of Yore for doing all the research.

After the Greek-Turkish war of 1919–1922 and population exchanges of 1924, thousands of refugees settled in Greece. These refugees were deemed Greek because they were Greek Orthodox, but they did not all speak a form of Greek. As I wrote at the Other Place, a good number of them spoke Turkish: over 100,000 first language speakers out of 1.3 million refugees—191,254 reporting Turkish as their language in the 1928 census (Google Books excerpt below), minus 84,669 indigenous Turks in Thrace (p. 17 of Featherstone & Papadimitriou's presentation):

But language was not what the Ottomans or their subjects used to define identity.

The villages that were abandoned are now dwelled in by the Muslims that stayed, and one such village is Kızderbent, "Maidenpass", in the region Greeks call Bithynia, 28 km away from Nicaea, which the Turks call İznik.

View Larger Map

The refugees from Kızderbent, who go by the name of Trakatroukides or Rokatzides, were scattered in Greece. There are two Facebook groups for refugees from the village (1, 2), and per this post on the second, most settled in Valtotopi, Kilkis prefecture; others ended up in the nearby villages of Ryzia, Gerakona, Fillyria, and Polypetro, but also further afield, in Nikomidino, Kallithea (Thessalonica prefecture), New Triglia (Chalkidike), and Roditis (Rodopi): this local newspaper article is on a get-together of the Rodopi Trakatroukides.

View Trakatroukides in a larger map
(Touchingly, the Greek Facebook page for those who have left Kızderbent links across to the Turkish Facebook page, for those who live there now.)

The Greek villagers of Kızderbent call it Κιζδερβέντ, Kizdhervent, and there is a copious variety of transliterations: Kizderved Kisderved Kizdervent Kisdervent Kizderbed Kisderbed Kizderbent Kisderbent.

And Къздервент. Because there's something notable about the language of Maidenpass. It wasn't Greek; or at least, it wasn't just Greek. We have little left of their language—and I'd certainly never heard of it during my preoccupation with Balkan linguistics; but from the Motley Word, a site I reviewed a while back, we have some 500 words. (To see them all, click on Όλες in the interface.) A lot of the words are Slavonic, and a lot of the words are Turkish, and none are noticeably Greek. That could be a selection bias—the Motley Work lexicographer (who is still going!) may just be singling out the Slavonic and Turkish words. But from the evidence I've seen elsewhere (see the followup post), I doubt there's a substantial amount of Greek vocabulary.

There was also discussion of the language of Maidenpass on a thread at Sarantakos', which triggered this (and which I've already posted): the indication there was that there was no mutual intelligibility between Trakatroukika and Greek. From that thread, we also have a sentence of Trakatroukika, a song verse: tákar tákar ke priataláh duí sulísperi zebé khantúm

The indications are that the language of this village was somehow a mixture of Balkan Slavonic, and Turkish. At least one glancing mention, from the website of the municopality of New Triglia, implies Armenian was there as well, but I have not seen a direct confirmation. (I have seen something more indirect, on which see further down.) The Slavonic component is conventionally called Bulgarian, and certainly when Slavists visited Kızderbent in the 19th century, noone was making a distinction between Bulgarian and Macedonian; but there is at least one account saying the village was settled from Ohrid, and the language spoken in Ohrid is not now called Bulgarian.

The name the villagers went by, Trakatroukides, doesn't give us much of a hint: as TAK anticipated in comments, it derives from traka truka, an onomatopoeia of firecrackers (as the Triantafyllidis dictionary gives it)—so "bang bang". The onomatopoeia is also associated with grinding knives (carnival song: τράκα τρούκα τη μαχαίρα, a verse that has wandered into Rebetika referring to knives), and a children's song about two steamships colliding ("The Greek boat has lemons. The Turkish boat has eggs. The boats go 'bang gang'. The sea is avgolemono").

The alternate name Rokatzides doesn't help either: ρόκα means spindle or rocket, arugula (Italian rocca—hence also English rocket; or ruca). It's tempting to associate "rocket" with "bang bang", but I'd say that's an anachronism—the association of spindles with rockets is late and Western.

The web site of New Triglia has this to say on the origins and language of the Trakatroukides of Kızderbent—the Bangbangers of Maidenpass:

There are three views on the founding of Kizdhervent. One claims that Greeks from Ohrid and Monastir [Bitola] migrated to Constantinople seeking a place to settle. The other opinion is that they did not migrate but were exiled because they robbed the nearby inhabitants. The third is that they were prisoners of Mehmed the Conquerer, who transported them to Constantinople and from there to Kis-Han, as he passed by Ohrid returning from Kroja. [...]
Later on, because of the privileges the Sultan gave the area, Greek-speakers, Turkish-speakers, Armenian-speakers and Bulgarian-speakers settled the region, and that's how Kizdhervent was created. The mixture of these peoples is also the reason why its peculiar language came about, spoken by no other village in Asia Minor. For that reason they were called Trakatroukides.
Their customs and religious habits are the same as other Greeks', and their religion is Christian Orthodox.

The Bulgarian Wikipedia article has a somewhat different story, which is worth giving in detail:

Kăzdervent was the first village of Asia Minor Bulgarians attested by written evidence in the 19th century. [Accounts from 1807 on, almost all of which describe the village as speaking Bulgarian.]
According to stories by the Kăzderventsi the village existed in the 18th century, and is one of 12 villages donated by the Grand Vizier to a lady in the seraglio: she sold them to an Armenian banker, and he sold them back to the Turks one by one.
Konstantin Jireček later wrote of Kăzdervent, and inferred that they originated from Momim Prohod, since that is what Kăzdervent means in Turkish. According to Jireček (1899) the villagers "speak Turkish and half are already on the way to losing their nationality entirely." Vasil Kănchov writes that the village had 440 houses, and believes that the Kăzderventsi were involuntary migrants from Ohrid.
At the end of the 19th century Kăzdervent was gradually Turcisised linguistically and Hellenised in national consciousness as a result of intermarriage and influence from other Karamanlides' villages.

So we have in Kızderbent a village which spoke Bulgarian (or Macedonian Slavonic); which ended speaking something between Bulgarian and Turkish (somehow), and was notorious in its region for "talking funny". (We have a report the Greeks and Turks in the area thought so; did the Bulgarians?) By 1914 however, the villagers did not consider themselves Bulgarians or Turks, but Greeks, and did not follow the Anatolian Bulgarians' exodus from Turkey; they left Turkey with the other Greeks in 1922.

The history behind that is obviously complicated, and needs a separate post. Here, I'm going to pose some linguistic questions. I don't have the answers (although I do hope someone who knows their South Slavic does go through the Motley Word list, and the language samples from the next post). For all I know, there have been Bulgarian studies of Trakatroukika; but I'll speculate anyway.

The most important question is, what sort of a mixture was Trakatroukika? If the situation was something like the New Triglia article paints, with Armenians, Greeks, Turks and Bulgarians cohabiting—and no single language dominating, we might end up with a pidgin. Pidgins arise as contact languages all the time, between people in recurrent but not close contact. For such a pidgin to become the main language of a village though, it would have to make its way into the hearth. That would mean routine intermarriage (possible for the Greeks, Bulgarians, and Turkish-speakers if they were Christian too).

It would also only make sense if multilingualism was impractical. That happens on slave plantations perhaps, which is how pidgins become creoles. But it didn't happen in Papua New Guinea, because the hearth can be multilingual; and although it may not now look like it, the peoples of the Balkans used to be multilingual as well. So the drastic breakdown of language which you get with pidgins—and the rebuilding you get with creoles—seem to me unlikely for Kızderbent. We'd know whether that had happened of course the minute a Slavist sees an actual sentence of Trakatroukika (like the one I pasted above). A pidgin or a creole would have very little left of the morphology of Bulgarian, and even less of its syntax.

True mixed languages, like Michif—the verbs are Cree, the nouns are French—are even more rare, more spectacular, and certainly more controversial. But they don't result from language breakdown, but fluent multilingualism: enough villagers speak both languages that not only words, but grammar ends up seeping between the two. It's worth noting that Wikipedia's list is short, but includes Cappadocian Greek—because Cappadocian was not just borrowing words from Turkish, but vowel harmony and inflections: elements of the basic machinery of Turkish.

That said, mixed languages are rarely stable anyway, and at least some of them (I'd say including Cappadocian) are an attention-grabbing twist on language shift—language death, if you prefer. That's certainly how Jireček read Kızderbent: their Bulgarian was mixed with Turkish, not because it was finding a new equilibrium, like Mbugu (Cushitic vocabulary, Bantu morphology) or Erromintxela (Romany vocabulary, Basque grammar). It was mixed because it was yielding to Turkish. Of course, if Trakatroukika really was like Michif or Erromintxela, Jireček may not have realised it, and I don't know if we'd have enough data now to work it out. This was the time of duelling aphorisms on mixed languages (Müller 1875: "There's no such thing as a mixed language"; Schuchardt 1882/83: "There's no such thing as a completely unmixed language"), but linguistics was only starting to be sophisticated enough to realise a mixed language when it saw it.

A scenario like Cappadocian though makes sense, and I suspect (based on no real information) that that's what happened in Kızderbent as well. The village language in SW Cappadocia has a Greek core, but is bilingual: several villages had both Muslim and Christian inhabitants (in one in fact the Muslims spoke Greek), and Cappadocia was also home to the Turkish-speaking Christian Karamanlides. The Greek-speaking villages themselves were shifting towards Turkish: the presence of Muslim Turks and/or Karamanlides made it that much easier. Because Turkish was gaining ground, and everyone was fluent in Turkish, there are more and more elements of Turkish grammar, and not just Turkish vocabulary, in the language the village spoke. In the Balkans, that same bilingualism meant that the grammars of the Balkan languages ended up looking fairly similar, even if their lexica did not. In Cappadocian, it was read—probably correctly—as a precursor to Greek yielding to Turkish correctly. Substitute Greek with Bulgarian, and what you get sounds plausible, at least.

To put the Cappadocian scenario in play in Kızderbent, it would help if we knew that the Turkish-speakers were on intimate terms with the Bulgarian-speakers. The Bulgarian Wikipedia (building on Kănchov? on that 2001 study on Anatolian Bulgarians?) thinks it was Karamanlides, who were Christian and so could intermarry. I think Bithynia is a bit too far West for Karamanlides, but obviously something Turcicised the Trakatroukides' language.

Outside the main question, of how Trakatroukika actually worked, are the historical questions of what went into it. For starters of course, the question that Jireček and Kănchov worked on in their visits to the region: was Kızderbent's Slavonic closer to Momim Prohod or Ohrid? (Or for all I know, a mixture of the two—although I presume these settlements were block moves and not mergers of multiple populations.) From the Bulgarian Wikipedia, I see we know a lot about Anatolian Bulgarian in general, and that it was mostly southern Bulgarian; it was possible to pinpoint some villages to particular towns—including Stengelköy, settled from "Kostur in Southwestern Bulgaria". (The hyperlinked town on the page is Kastoria, now in Greece, and its Slavonic has not been called Bulgarian for a while; but there are a couple of Kosturs in Bulgaria as well, so I'll assume lazy hyperlinking. :-)

It'd also be worth checking whether there is actually any Armenian in the language mix, as New Triglia's site hints. Bithynia is next door to Phrygia, and Phrygian is suspected of being related to Armenian—or a missing link between Armenian and Greek. Phrygia is a long way from the Armenian heartland, but Armenians did travel west during the Ottoman Empire; and the Wikipedia listing of Armenian dialects, from Acharian's 1909 Classification des dialects arméniens, includes Nicomedia (İzmit) and İznik (Nicaea). So Armenian was spoken in Bithynia at the time (and there's confirmatin next post too). The creation story of the village reported in the Bulgarian Wikipedia refers to Armenian involvement.

New Triglia also drops the hint of a Greek contribution, both in the four constituents of the village, and in its reference to "Greeks from Ohrid and Monastir". (If your reaction is, "*what* Greeks from Ohrid and Monastir?"—well, hold that thought.) As mentioned, there is no evidence of mutual intelligibility with Greek, and the Motley Word lists don't feature a single word that looks more Greek than Bulgarian and Turkish. The Motley Word is a Greek dialect wiktionary, so Hellenic words should have still been distinctive enough to include. (Of course there's a selection bias, and I'll come back to that.)

As a language of culture and trade (and the church), Greek had pervasive influence through the Balkans and into Asia Minor, including where Trakatroukika was spoken; and Greek was spoken by other communities in Bithynia:

• Danguitsis, C. 1943. Étude Descriptive du Dialecte de Démirdési (Brousse, Asie Mineure). Paris: G.-P. Maisonneuve.
  
• Δεληγιάννης, Κ.Π. 2002. Κουβουκλιώτικα: Το γλωσσικό ιδίωμα των Κουβουκλίων Προύσας. Αδελαΐδα.

From the Bulgarian Wikipedia again, we know both Turkish and Greek were encroaching on Anatolian Bulgarian in general, not just in Kızderbent: "Surrounded by many Greek and Turkish villages, the Bulgarians were forced to adapt and to learn the languages of the majority. According to the recollections of the older sons of Anatolians, almost all young and old settlers in Anatolia also used the Greek and Turkish languages, though many of them were illiterate." But if Bithynian Greek had an effect on Trakatroukika, I suspect it would have been apparent to me by now.

And finally, "*what* Greeks from Ohrid and Monastir?" Well, that's a messy question, and it's best left for The Other Place, where I take all discussion of complications in what being Greek (or Bulgarian, or Macedonian) is about. You can probably already predict what I'll say: the current Trakatroukides of Valtotopi or New Triglia identify as nothing but Greek—so they're nothing but Greek. The indication is that they share a bloodline and a linguistic heritage with people who have never been Greek. And one must have a very restrictive notion of national identity, to think that a paradox.

Because of course in the 17th century, or whenever Maidenpass was settled by the Bangbangers, there was no Bulgarian or Macedonian or Greek identity as we now know it; there was just the millet-i Rûm of Orthodox Christians, which encompassed Greeks, Albanians, Bulgarians, Arabs and Romanians alike. People knew they spoke different languages (and very different dialects within those languages), but that was not how they defined themselves: that came later.

So yes, it may be wishful thinking to say that settlers from Ohrid in the 17th century were Greek—particularly as Greek that far north was an urban-only language, and Kızderbent is unlikely to have been settled by townsfolk. But if you're going to deem them Bulgarian or Macedonian, and not just Slavonic-speaking, you'll need a very big asterisk; and to their descendants, such an identification is irrelevant. (If you make it alive out of the argument over whether they were Bulgarian or Macedonian—which goes to show how immaterial the question was back then.)

...Read more

How I found out about the Trakatroukides

This post has been put off for a long time, and its sequel even longer; but one of its protagonists has just asked for it here, so I will not put it off further. This is a translation of a derailed comment thread last October, over at the Magnificent Nikos Sarantakos' Blog, through which I discovered the existence of the Trakatroukides of Kızderbent (Kizdervent, Κίζδερβεντ, Къздервент). Kızderbent is a village in Turkey which spoke "Trakatroukika"; Trakatroukika is reportedly a heavily Turkicised Bulgarian (possibly with Greek and Armenian elements). The villagers of Kızderbent considered themselves Greek, and their descendants now live in Greece.

There will be two or three followup posts with more detail that I can find about the village and its language, but I wanted to get this out to tide me over. I'd like to thank here the two protagonists of the thread:

• In the right corner, Cornelius (Κορνήλιος), the polytonicist linguistic reactionary gadfly of the blog—and as it turns out, the grandson of a Trakatroukika speaker. It's worth noting, for those interested in Greek diglossia, that even if Cornelius is a linguistic reactionary, he can't quite use Puristic for the Cause in 2010—which underlines how thoroughly Puristic has lost ground by now. (That he is under forty doesn't help either.)
  
• In the left corner, A Butcher of Yore (Ένας Χασάπης Από Τα Παλιά). He's a Butcher Of Yore because, like many other regulars at Sarantakos', he used to be on the Hellas mailing list in the 90s, where his moniker was Abdullah the Butcher. Butcher of Yore has deep reserves of scholarship on Greek refugees, and I want to thank him for forcing me to look at the Trakatroukides more thoroughly than I intended to.

I'm going to translate the entire exchange, because it really did come out of nowhere, it went in several interesting directions (including whether it's worth preserving languages), and it has some great pointers. Like I said, I'll be supplementing this with more material in future posts (including links to a dialect lexicon, and to a conversation snippet on YouTube).

I've inserted some annotation in square brackets; I am editing out some of the ribbing and the language politics, because that would tread on sensitivities further than is necessary (including taking some linguists' names in vain). But the thread is openly accessible (in Greek).

• #20 Cornelius: My grandmother calls jars κιούπι, and γκιούμι for metal vessels. Once a gang of antiquities smugglers was arrested, and I was watching it on the television with my grandmother. They were showing some amphorae and my grandmother said, "ah, we got plenty of them κιούπιa."
  
• #21 Voulagx: Cornelius, do you know where γκιούμι comes from? I mean is it a Turkish word, Slavic, or something else?
  
• #22 Cornelius: I don't know. My grandmother in her childhood spoke a language with Turkish, Slavic, Bulgarian, Greek words etc in it—and she was of Asia Minor descent, at that. They'd say mátska for cat and kokóska for hen and kútse for dog and ódzak for fire and sófa for couch. Quite unacceptable. How could I not become an admirer of the learnèd tongue after that? Fortunately my grandmother has almost forgotten that language. Her mother spoke it; her father spoke Greek, and was at war for 10 months and 10 years. Of such a generation do I long to be. My grandmother also uses the word ílem: I don't know what it means, but I know how to use it instinctively.
  [More on Cornelius teaching himself polytonic at 11.]
  Also from a fairy tale my grandmother used to tell, I remember this song which was supposed to be spoken by a snail on a bridge while a cart went past. I don't know what language it is or even if I'm breaking the words apart correctly:
  tákar tákar ke priataláh duí sulísperi zebé khantúm
  
• #22 Butcher Of Yore: Don't say it so loud, Cornelius; no matter where you think your grandmother's descent is from, Tremopoulos will deem you an Oppressed Macedonian, and then you're in trouble! matska and kutse, he says—why Gruevski will be looking for you as a Grkoman example to avoid! :-)
  In all seriousness, if your grandmother's mother was truly from Asia Minor, the only thing I can think of is that she was one of the Trakatroukides of Kızderbent. But there were very few of them.
  Otherwise, should I presume her mother was a local Slavophone, and her father was from Asia Minor?
  
• #24 (Me thanking Butcher for the information, and threatening Cornelius with posting about Trakatroukika)
  
• #25 Butcher Of Yore: Well someone should work on the Trakatroukides before their language dies out (as with the speakers of Cappadocian, Propontis Tsakonian, etc.) Beyond the aforementioned municipality in Chalcidica, you can also head towards Polykastro in Kilkis and find them there.
  They're organised: [Facebook group], and there's a thesis on them. [NN: More links in following posts]
  
• #27, #29, #44 (Me burbling further on what I found on Wikipedia, including the claim that they were Bulgarians) [More on that in following posts too]
  
• #31 Nikos Sarantakos: (γκιούμι ~ γκιουγκιούμι < Turkish güğüm < Greek κουκούμι(ον) < Latin cucuma) So what is the word? Greek, Latin, Turkish? It turns out to be common property.
  
• #33 Voulagx: Sarant, it was an honest question, because we also call it γκιούμι in Aromanian (giumi-a, where -a is the feminine article). I like "common property", but you're making me wonder: what if the partners start arguing about their percentage of ownership? Thank you.
  
• #34 SophiaOik: Cornelius, I suspect you're trying to trick everyone with your grandmother; it's not as if those parts had so large a population, or were fertile enough, for us to bump into their descendants as easily as that.
  
• #40 Cornelius: I'm not tricking anyone, I'm telling the truth, and she probably was a Trakatroukis, because I have heard it said, that they spoke "trakatroukika".
  
• #44 #45 #46 (Me and Cornelius about me posting about Trakatroukika)
  
• #48 Cornelius: I used to note whatever my grandmother remembered of that language in a notepad, but where might it be? I'll ask my grandmother when I go to the village on a weekend. They had two versions of "fish", balúk and ríba. My mother says that my great-grandmother never learned Greek; so when she would go as a child to my great-grandmother's village, her grandmother, she must have learned some words; but she doesn't remember them. But I found this [a lexicon at the Motley Word site], and I think I remember ókhtse, I must have noted it in the notepad.
  
• #49 Cornelius: In fact they came from the region around Nicomedia to Greece. [Kızderbent is near İznik, Greek Nicomedia.] First they took them to Komotini. [NN: Some Trakatroukides still live in Roditis, 5 km from Komotini.] But my great-grandfather did not want to see mosques, they were sick of them from Asia Minor, and they moved on. It's all so fitting for me now!
  
• #50 E-fufutos: You're carrying such an imperial treasure with you, and instead of preserving it, you exorcise it with caricatures of the Ancients? Haven't you heard stories of aunts (like mine or "the Aunt From Chicago"), who thoughtlessly threw out their grandmothers' furniture, carved from massif wood, to buy fake European stuff out of plywood, only to regret it now?
  I hunt after things like that, maybe because I feel the lack more intensely. My folk have their peculiar origins hidden more deeply.
  
• #51 Maria: Cornelius, if your great-grandmother didn't speak Greek, then your grandmother must have spoken Trakatroukika. She can't have needed an interpreter to communicate with her.
  
• #52 Cornelius: Anyone can consider whatever they want to be a treasure. I don't consider the leftovers of Ottoman Rule and Bulgarians a treasure. What I consider a treasure are Doric lexical remnants still surviving in the mouths of illiterate villagers. As for old furniture, they are my weakness after old books.
  
• #52 SophiaOik: I wouldn't say they're an imperial treasure, because their vocabulary would be limited to everyday objects with few abstract notions. Such a mixture might have linguistic interest, but such halfway languages seldom survive long, because they have few speakers and are not spoken outside the group that creates them. If Cornelius works on linguistics, he might write a couple of papers on it, maybe even a PhD. If not, he need not feel guilty.
  
• #54 Nick Nicholas: Let's lay off people: all treasures end up faded and ignored. If someone doesn't want to be a hostage to his heritage, well, we have to respect that. I do it too, and all of us who are modernised do. If he prefers grave accents [referring to Cornelius and the polytonic], good for him. I prefer English, after all. [By which I meant, I don't feel an atavistic obligation to use the language of my forefathers in my everyday dealings.]
  I remember a talk on remedies and folk botany in Indonesia in my linguistics department. All us westerners were saying how cute and picturesque it was. But the Indonesian student was irritated: he did not come to the West from Aceh to keep hearing about primitive nonsense. And of course he was under no obligation to stay "picturesque".
  But Cornelius kept a notepad and submitted entries to the Motley Word. [I was wrong about the latter.] I'm hardly going to tell him off; quite the opposite.
  
• #55 Butcher Of Yore: Well, I've impressed myself with my correct guess on where Cornelius' great-grandmother was from. See what obsessing with the Macedonian Issue does to a Southern Greek? Admittedly the Trakatroukides had particularly impressed me when I first found out about them. I'm curious to hear if they really did also have Armenian linguistic antecedents.
  Of course the Bulgarians consider them all their own, Nick :-) —they consider the Slavic-speakers of the Albanian Prespa lakes Bulgarians too. (And some though not all of them do call themselves Bulgarians, as their great-grandfathers did.) The same goes for the (Muslim) Gorani of Kosovo. The Bulgarians would even issue them with passports, at least until recently.
  But if you've read about Vasil Kanchov [who had researched the Bulgarians of Asia Minor], you immediately realise that these populations had been placed into the game of Balkan nationalisms, since Basil was the ethnographer par excellence of Bulgarian nationalism. (And his work was quite detailed—Lithoxoou treats him as a saint.) So politically Bulgaria clearly tried to stamp them as her own, to use them as a minority, either as leverage and a pressure point, or as a population to exchange (as largely happened). Somewhat like how Rumania used the Rumanian-identifying Vlachs (since there was of course no possibility of advocating for a Rumania reaching the Aegean). [See "Awakening of the Aromanian identity and Romanian sponsorship "]
  What's interesting to research further is whether the enmity between the Patriarchate and the Exarchate [which marked the ecclesiastical starting point of Bulgarian nationalism] had spread further from Macedonia, Thrace and Constantinople to Bithynia [the region of Kızderbent], and how extreme or even violent it ended up as in Asia Minor. It's not impossible that those who left for Bulgaria were already pro-Exarchate, and those who stayed behind and were expelled [to Greece] with the population exchanges were pro-Patriarchate—meaning, the national/ecclesiastic split in the Ottoman Balkans had also passed to Asia Minor.
  Trakatroukika have unfortunately gone down the road of all the remaining dialects and languages of the Greek people, though given the small number of its population and its scattering in Greece, it is closer to dying out than the other dialects and languages. It would be good for it to be studied and recorded; I don't consider it practical to revive it at this late stage, nor do I think it desirable, if it will only be promoted by certain activists, and the majority of the descendants of the Trakatroukides are indifferent to it. But we owe it to the tradition of our people not to let the recollections of local idioms and their treasures be lost entirely—the more so when they are particularly unusual.
  [...]
  Cornelius, seriously, gather whatever you can preserve. From my grandfather's Andriote Arvanitika, I can only vaguely remember some curse words (because that's the only time he'd use it), and I am unhappy about it.
  
• #56 Cornelius: I am not denying my heritage, nor am I acting modernised. And don't think I want those words and their meaning to be lost permanently from people's memory. When I was noting them in my notepad I thought with some degree of melancholy that one day there would be noone alive who had heard those words used. But that's one thing, and grave accents are another. Trakatroukika can't be an official language or a language of literature.
  
• #59 Nick Nicholas: True, the idiosyncratic language mixture of one village can't be an official language. (Of course, Puristic was an idiosyncratic mixture too, but never mind.) But a formal language becomes such through extra-linguistic circumstances. I mean, if we could accept 150 years of Puristic, under different circumstances we could have ended up with an official Trakatroukika.
  So long as we weren't obsessed with grave accents. [i.e. prioritising continuity with Ancient Greece.] But that's an extralinguistic factor too.
  
• #60 Languagelust: balúk (balık) and ríba (риба)! The first Turkish, the second Slavic. You're giving us these one by one, Cornelius!
  
• #76 E-Fufutos: I don't believe that languages, dialects and idioms that have started heading towards oblivion should be preserved in formaldehyde at all costs, either, just so we don't get called nationalists, or for reasons of folklore. If the speakers themselves are reluctant and it doesn't sit with them, we should let things develop on their own. You have to like something to want to retain it.
  ...
  I also have relatives with some Arvanite roots, and nothing has survived, not even a recollection from a grandmother or a great-grandfather who used to speak Arvanitika proverbs; and I consider that natural.
  
• #77 Angelos: We should not keep it alive with a respirator and an artificial kidney. But we should preserve it in formaldehyde, like we would piously preserve a mammoth if we found it close to death in Siberia, dying in our hands. Not "just so we don't get called nationalists, or for reasons of folklore". But because every manifestation of the human phenomenon par excellence we call language is interesting and worth recording.
  And if Trakatroukika are indeed a "mixture" of Bulgarian, Turkish, Greek and Armenian (and not just a South Slavic dialect which has absorbed a lot of Turkish, Greek and Armenian words), it would be a very rare if not unique linguistic phenomenon. That's one more reason to record and study it to the extent it is still spoken (if indeed it is at all), even if it's just ten old ladies.
  
• #78 Cornelius: I asked my grandmother today and she told me that ókhtse was the word for a big sheep, and a small sheep was called kuzía.

...Read more

Kozani: a stab at etymology

Language Hat asks in comments to the previous post about the Wikipedia etymologies of Kozani:

According to prevailing opinion, the name comes from the village of Epirus Kósdiani, the origin of settlers of Kozani in 1392. The settlement was first named Kózdiani, which then, it was changed into Kóziani, and in the end into Kozáni.[2]

The name "Kozani" probably may also derive from the South Slavic kožani < koža 'skin (goatskin)'.[3] The name of the city in South Slavic languages is Кожани (Kožani).

Am I wrong in thinking the Slavic etymology makes much more sense than the first one, with its strained reshapings? (We won't even get into what "probably may also" is supposed to mean.)

As often happens in Wikipedia, it's two opinions barely reconciled together on the same page. Now, Googling is no surrogate for actually knowing Slavonic (or Albanian), but let's see how I can help. (Damn you, Language Hat, I was supposed to be doing my taxes!)

The "Hellenic" etymology is sourced from the website of the city of Kozani, which as you'd expect wouldn't be eager to point to the parallel with that language. (Kozani prefecture turns out to be the furthest spread south of Slavonic in Modern times, with Greek and Slavonic villages mixed.)

[Non-Greeks will notice that I speak vaguely of Slavonic, when we all know *which* Slavonic language I'm talking about. But I'm not feeling like getting into needless argy-bargy with those of my readers who don't want to call it "Macedonian"; and since I'm talking about 1400 and not 1950, I may just get away with it...]

The thing is, Kozdiani may not look like Кожани, but it doesn't look particularly Hellenic either. Is there any way Kozdiani could originate from the same Slavonic form as Kožani? Although given it's Epirus (Northern Epirus, as it turns out), and given the names of the other villages cited below, Albanian is more promising as an etymology.

This is what the Kozani city website has to say:

In 1392 colonists from Premeti, Bithikukio, and Kozdiani of Epirus {Përmet and Vithkuq, Albania}, fled chased away by Muslim Albanians to the region north of Selitsa {Selitsa, now Eratyra, Greece}, which is to this day called Old Kozdiani, and subsequently migrated eastwards, encountering the Christian settlement of Kalyvia. {The quite Hellenically named "Huts"; I've found one near Konitsa, on the Greek side of the border.}

The inhabitants of Kalyvia did not reject them, but they obliged them to build their houses further east. The new inhabitants called the region Tzamouria {i.e. Çamëria), preserving the name of their old region.

Nowadays the region is called Tzambra. {[dzamurˈja] > [dzamˈrja] by high unstressed vowel deletion > [dzamˈbra] > [ˈdzambra]}. They called the rocky hill over Tzambra Skrika or Skirka (Sk'rka), which means a rocky elevation.

Though there are different opinions of where the city name came from, the dominant opinion is that these colonists from Epirus called the new settlement Kósdiani, which then became Kóziani, and later scholars [i.e. in Puristic Greek] transformed it to Kozáni.

A few years later families from Servia {town in Greece} and Drepano move to Kosdiani, augmenting it.

From another site: "People speculate that the name Kozani is due to either the place of origin of its first inhabitants, Kosdiani or Kostiani in Epirus, or their main occupation of tanning: "coza" in Epirot means goatskin." Tellingly, coza is given in Latin script, which you'd be unlikely to do if "Epirot" here meant "the Greek dialect of Epirus".

The account derives from a 1924 history of Kozani, cited at this blog:

The name of the folklore group (Kóziani), stressed on the antepenult, shows a hardy genuineness, like the hulls of Sk'rka (Slavic: rocky protuberance; Albanian: hill). And given the opportunity: the definitive interpretation of the name of the city is stilll under research. I'm opening up the History of Kozani by P. Lioufis, Athens 1924.

• The first inhabitants of the region one day saw ... a she-goat running through the trees, and called the village Kózani or Kóziani; for kóza means a she-goat, and kózia a skin in "Bulgarian".
  
• Above the town of Selitsa (Eratyra) there is a region in its mountains called Kóziani, or rather Old Kóziani, where those pursued from Northern Epirus first settled, from the villages there of Kosdiani and Bithikukion. (The latter exists to this day.) Then they came to the region of Paliospita where the Kasmirtzidis Society have their establishment, as do many other settlers in the area, genuine or not.
  
• The Turks called the town Kózana, from Koz ("walnut", or Kákhta in pure Kozani dialect) and Ana "mother", for the multitude of walnuts there.

And here's a map of Përmet, Vithkuq, Kalyvia, Selitsa (now Eratyra, near Askion), Kozani, and Servia:

View Kozani in a larger map
At least one site says that the original Kozdiani was destroyed, which is consistent with Lioufis' wording

So what do we make of all this?

• The Turkish is a folk etymology, and nothing wrong with that.
  
• Kozáni may be named for Kózdiani, but it's quite possible that both are named for the Slavonic for "goat"; Lioufis certainly assumed so.
  
• Kózdiani is in Northern Epirus, i.e. modern Southern Albania; Albanian and Slavonic are both possibilities for etymology, and Albanian and Slavonic would both have been trading words anyway (as appears to be the case with "Sk'rka")
  
• Tzamouria i.e. Çamëria is the name of Southern Albania, although we don't know from this whether we actually have a record of the area around Kozani being called that at the time—which would confirm migration from Southern Albania—or whether this was Lioufis' etymology of modern Tzambra. All of the Tzamouriá > Tzámbra etymology is plausible, except for the stress shift.
  
• I'd like to know more about Lioufis's sources, because the Greek Wikipedia article then adds that the first written reference to Kozani is in a firman of 1528.
  
• I think 1392 is a little early for religious conflict in the Western Balkans—the reference is to τουρκαλβανοί, Muslim Albanians not τούρκοι. But I could be wrong. *shrug*
  
• "The inhabitants to Kalyvia did not reject them, but" they kicked them out anyway.
  
• I'm not at all convinced that the stress shift of Kóziani to Kozáni is a learnèd thing. Sure Kóziani violates ancient accentutation (because Κόζιανη looks like being stressed four syllables from the end—until you realise it's [ko.zja.ni], but Ancient Greek did not have ι as [j]). But I'd have expected a learned form to go with Kozíanon, rather than preserve the feminine gender. And as Don Hat correctly noted, phonologically the shift from Kósdiani to Kozáni does seem a little forced.
  
• So the derivation of Kozáni from Kósdiani has problems. But as it turns out, both Kozáni and Kósdiani seem to have a Slavonic origin anyway, so it's a distinction that doesn't matter.

So, that's what googling tells me. If anyone actually knows any Albanian, Macedonian [there, I said it], or Greek dialect and can contribute, you'll be doing the world a favour. Particularly if they can explain the accentuation of Kozáni. ...Read more