Your Firework Eyes

By: | Post date: 2019-03-22 | Comments: 8 Comments
Posted in categories: Modern Greek, Music

Your Firework Eyes, Τα βεγγαλικά σου μάτια, is a 1995 song (Lyrics: Michalis Bourboulis, Music: Stamos Semsis), first sung by Giorgos Dalaras, and covered the following year by Dimitris Mitropanos.

It is a moving, fragile, beautiful song about the loss of love. And there are some interesting things about how it was put together, that make it so striking. Both lyrically, and musically.

Musically, it is a torch song; it is sorrowful, vulnerable, whispered almost. The more effective in that it’s been sung by two artists who aren’t normally whisperers, who can do steel behind their plaint, and transmute it into something more.

And musically, it goes around and around, obsessively, with the same tune over and over each stanza, sternly sinking down to the tonic in stages, in the relative major (so beloved of Greek song writing), in Phrygian mode, in resigned vi–vii–i. And with no chorus to relieve it. Something unusual for a zeibekiko. But then, this is a very unusual zeibekiko.

It’s astonishing to realise that this is a zeibekiko at all. The zeibekiko, the 9/4 mainstay of bouzouki pop, is realised as a stern, heavy-footed, confident swagger. It can be fatalistic; it certainly gets to be self-important. It doesn’t whisper. It doesn’t sound like this. And that’s the genius of the arrangement, which has been maintained in the covers: it’s a song that undermines its own genre.

I’ve written years ago of another such instance, Markos Vamvakaris’ Είσαι μελαχρινό και νόστιμο, whose notes are the notes of the free-flowing Levantine chromatic lament at the root of rebetiko—but whose ethos is of the jaunty, four-square Peiraeus Sound that followed it. Your Firework Eyes is another such instance of musical alchemy. It would be very easy to sing the notes of the song like an actual swaggering zeibekiko. Noone dares to. This somewhat out-of-tune karaoke recording is the closest I’ve been able to find:

It’s not just about the music, though. The lyrics are doing a lot of work here:

I lit all the lights. I put on a show.
When love dies, it knows no resurrection.

Your firework eyes shine like phosphorus,
like ships passing through the Bosphorus at night.

You switched off the lights and left, you became invisible.
Mist that the wind took away, in an automated town.

Your firework eyes are a bonfire
and loneliness drips like rain onto the floor.

I am trapped now in your perfume, in your name,
and in your eyes, yes, your cold firework eyes.

Your firework eyes shine like phosphorus,
like ships passing through the Bosphorus at night.

Disjointed images of loss, of sorrow, of cold. With a lot going on that’s culturally specific to Greek.

Like the mention of the Bosphorus. Phosphorus and Bosphorus are the rhyming words in the original: Τα βεγγαλικά σου μάτια φέγγουν σαν το φώσφορο / σαν νυχτερινά καράβια που περνούν το Βόσπορο. In English, that sounds too marked to be anything but silly. (I’ve received a guffaw about it that is in retrospect painful. There’s some personal associations going on here.) That is why I’ve had to dodge the rhyme in English.

But it’s not silly in Greek. First, because they aren’t Greek words that stick out in English; they are just Greek words in Greek. Second, because the rhyme in Greek isn’t that rich: it’s [ˈvosporo] rhyming with [ˈfosforo]. And third, because ships in the Bosphorus is a painfully rich image in Greek. The romance and melancholy of Istanbul, yes. But also the pain that comes with thinking of Istanbul: the memory that once, this was our city. And now it is lost to us.

Which is just right for what the song is about.

The second thing about the lyrics is that the metre gets disrupted in the final stanza (before the repeat of the Bosphorus stanza). Up until then, the stanzas were all in trochaic octameter, with a masculine ending (i.e. ending on a strong syllable: ´ – ´ – ´ – ´ – | ´ – ´ – ´ – ´ .) The metre of final stanza falters: it adds a weak syllable at the end:

Είμαι πια εγκλωβισμένος στ’ άρωμά σου στ’ όνομά σου
και στα μάτια ναι στα μάτια τα ψυχρά βεγγαλικά σου

And everything falters with it. The music puts that extra note on an uncomfortable minim, that sounds drawn out too long, deliberately out of place. Dalaras captures the hesitation and awkwardness of the notes beautifully. And the lyric matches it: the faltering repetition of “your”, the syntactically disruptive, rueful recapitulation “yes, your eyes”, the forced piling up of adjectives at the end of the stanza.

There’s a third thing. The images are vivid, but they are disjointed, they don’t really come together as a narrative. There’s a reason for that; and the reason tells you a lot about how lyricists work in Greece—and how composers can make a virtue of it. With a little help from their family.

The composer Stamos Semsis has told the story about how the song came to be written. It explains the disjointedness; and it also explains how the obsessive, single-minded tune took the song over.

When I started collaborating with the lyricist Michalis Bourboulis—someone much older than me, and a great writer—the initial material he had entrusted me with was a package of some 80 pieces. He asked me to read through them, to pick what I liked, and take it from there. Most of them were printed, but a large number were handwritten. In the back of a handwritten sheet, he had printed the quatrains of Firework in random order. They turned up there by accident.

I was married at the time, and my stepson Alexandros was around 14. Despite me being his stepfather, we were good friends, and we got along very well. We had the following routine: I’d work mornings at home, and when he’d come back from school, he’d listen to my songs and tell me what he thought of them. I had started working on that particular piece, and I’d constructed the basic tune for one of the quatrains and a small bridge.

When Alexandros came home from school, I told him about the piece, and I explained that it was a little weird, because the quatrains were out of order. He asked to listen to it, and he went crazy. “Look, Stamos, don’t go complicating the melody like you usually do. It’s so beautiful and simple.” When I asked him what order to put the quatrains in, he told me to leave it with him. And that’s what happened. He took a sheet of paper and put the quatrains in order.

That’s how the piece came to have the form it does. “What you’ve just done is producer work”, I told him, and I asked him whether he wanted credits on the album. “No, no, be serious”, he answered. Twenty-odd years on, Alexandros is working in one of the biggest consulting firms in the world.

Crossover artists in Greek pop: the Malamas–Karras effect

By: | Post date: 2019-03-20 | Comments: 2 Comments
Posted in categories: Modern Greek, Music

I am about to post here on late song renderings by Dimitris Mitropanos, and there’s something about what he did with his late repertoire that was special, but that I couldn’t quite put a name to.

Mitropanos had a decades-long career as a Laiko artist: he worked in the mainstream Greek bouzouki pop tradition, singing songs of love and machismo and disillusionment. Nothing too intellectual.

In his later repertoire, Mitropanos sang Entechno repertoire. He sang Entechno with a firmly Laiko sensibility; and that made it all the richer for it.

Entechno music, “art music”, is a parallel tradition to Laiko; for a time it was emblematic of Greek music. (Theodorakis and Hadjidakis were its main early exponents.) The music is of the same family as Laiko, but tends to be more European than Levantine; it is friendlier to acoustic guitars and/or Western orchestral instruments, without letting go of the bouzouki bedrock; and (possibly the most important difference) the lyrics are consciously poetic. It often drew on established poets early on, and the lyricists who worked in the tradition regarded themselves as poets, and usually wrote like that. At their best, they wrote astonishing, richly and darkly allusive poetry. At their worst, they were obscurantist.

There isn’t a good equivalent in Anglo popular music; it’s like the singer-songwriter tradition exemplified by Bob Dylan times a hundred, staying in the mainstream for decades, setting Auden and Eliot and keeping poets in business.

And getting covered by artists like, say, Prince.

Now, I have ventured hesitantly back on Quora, although I can’t see myself putting in the investment there any more that I used to. The reasons I left there still hold, and the feed looks much more dysfunctional than it used to. But it’s good to have rekindled some friendships. Such as friend to this blog, Evangelos Lolos.

And this article draws on two observations he made, when we briefly discussed Mitropanos.

First: I just said that Mitropanos sang Entechno with a firmly Laiko sensibility. Evangelos put it more simply than that:

they are not entechno songs when he sings them.

There is a simpler name to put to it still, in the English tradition: Mitropanos was a crossover artist. That doesn’t necessarily make it a better name though. Crossover in the Anglo tradition has bad connotations of selling out and dilution, but then the Anglo tradition has some curious preoccupations with authenticity and purism.

The exemplar of this kind of crossing over, Evangelos proposed, was the Malamas–Karras effect. The song Πριγκιπέσα started out in the entechno tradition, recorded by singer-songwriter Sokratis Malamas in 2000:

With a lovely, singer-songwritery story behind how he came to write it:

I wrote it as a birthday present for a woman, because I had no money to buy her anything. She was cooking lentils, which is all we had left, and I looked at her and thought: “This song is worth singing at this moment, it’s worth getting out there.” I wrote it and played it immediately, without a pause. I burst out laughing when I played it, I thought it so funny. My friend, who had no idea about music, put down her ladle and said, “When did you write that? Do you realise how good it is? Why would you laugh?”

(And yes, Reader, he married her.)

It’s a lovely, melancholy tune, with tinkling bouzouki and guitar. But the guy is a singer-songwriter, and it shows: I don’t think he does the song justice, he doesn’t sing it with anywhere near enough oomph. Still, there was a rocking tune there, and an even more rocking lyric: a lyric which is full of love and wonderment and magic, like good Entechno should; but also pulsing with bad boy machismo and fatalism—the stuff of Laiko, in fact:

I want one thing, I do another.
How can I make you understand.
I thought, the years are going by,
I’ll go straight.

But it does you no favours
to try and change your nature.
No point in keeping score.
No point in forcing yourself to be good.

The wind blows outside,
but in my heart,
in this house,
my princess,
your light and the light
dance around us.
The world is unbelievable
and so is our nature.

I want one thing, I do another.
That’s how I’ve ended up here.
Errors, missteps, and passions,
they’ve set me straight.

At dawn in the street,
I cast a line.
I catch myself;
I lose my mind.

As this blog post from 2009 blog put it,

When I first heard the record, I liked the uniform style he maintained on all the songs, but none of them stood out above the rest. (That particular song, I found somewhat confusing.) I found it good, in fact, but non-commercial. It never occurred to me that Pringipessa would become a massive crossover hit, admired by singers ranging from Haris Alexiou to Vasilis Karras, and that it would end up with 1,000,000 hits on YouTube! What I had momentarily missed was the approachability (λαϊκότητα) of Malamas’ songs, which would hit a vein in just two verses; and people wanted to keep listening to them over and over, like vain spells against time: it starts with “I want one thing, I do another, how can I make you understand”, anticipating defeat, and it ends in that conclusion full of bitterness and bewilderment: “The world is unbelievable, and so is our nature.”

That’s what Malamas did with it; this is what Karras does with it:

I love when a good, stern Laiko interpreter like Mitropanos turns Entechno into austere gold. I don’t think Karras can turn anything into austere gold, but it’s clearly not Malamas’ sensibility any more. It is a sensibility more in tune with the initial machismo, but it can’t do anything with the wonderment of the chorus.

Glykeria, on the other hand, turns everything to gold, though I think she too rattles the chorus off.

Peggy Zina does it a little more justice:

I don’t think Mitropanos ever did Pringipessa. I’ll come to what he did turn to gold in a future post.

Updated posts on “What did your language sound like 1,000/500 years ago?”

By: | Post date: 2019-03-16 | Comments: No Comments
Posted in categories: Mediaeval Greek, Modern Greek

I have consolidated my old Quora posts and, and just had it published in Greek on Nikos Sarantakos’ blog:

Albert/Bedwere/Nicholas: Imaginum Vocabularium Latinum in Ancient Greek

By: | Post date: 2019-03-02 | Comments: 3 Comments
Posted in categories: Ancient Greek, Modern Greek

In the Textkit Greek and Latin Forum, Bedwere has been translating Sigrid Albert’s Imaginum Vocabularium Latinum into Ancient Greek over the past year (as Λεξικὸν Ἑλληνικόν).

Albert’s dictionary is a Duden-style illustrated dictionary, where concepts are organised into thematic groups, and pictures of the concepts are accompanied by Latin glosses. In the (extensive) back of the dictionary, there are indexes of all Latin words used, and of their equivalents in German, French, English, Italian, and Spanish. The intriguing part of the dictionary is that the concepts are modern. (Or at least, modern-ish: the dictionary was written in the 90s, which means its stationery and technology section has dated badly: when was the last time anyone dealt with a typewriter ink ribbon?) So the Latin in the dictionary is substantially Neo-Latin.

The dictionary is not only of its time, but of its place. It was written by a German, for German students of Latin, and it shows: several of the concepts involved are heavily rooted in German culture, and a few of them need research for an unknowing outsider to make sense of. (Why is there a puck being used in curling? Oh, that’s Bavarian curling. And what precisely is a Konzern anyway?)

But both of these make it a fascinating undertaking, to see how Latin has been pummelled into place to cope with modern concepts. And the same holds for the dictionary in its Ancient Greek clothing. As Bedwere’s blurb puts it:

χαῖρε, ὦ φίλε ἀναγνῶστα. σύγε τούτῳ τῷ λεξικῷ χρώμενος, οὐ μόνον γράψεις τε καὶ λαλήσεις περὶ τὰ καθ’ ἡμέραν τῶν ἀρχαίων Ἑλλήνων ἀλλὰ καὶ περὶ τὰ τῶν νῦν ἀνθρώπων. αὕτη γάρ ἐστιν ἡ εἰς τὴν Ἑλληνκὴν γλῶσσαν μετάφρασις ἐκείνου τοῦ εἰκόνων Ῥωμαϊκοῦ λεξικοῦ

Halfway through Bedwere’s work, I got involved in the project, suggesting corrections and emendations to his translation, informed by my perspective as a Modern Greek speaker. As you can well imagine, the project had added fascination for me, not in terms of Ancient Greek, but in terms of Modern Greek. I made a point of seeking out any and all instances where Puristic Greek had already come up with its own Hellenic-based renderings, and minimising novel coinages as much as possible.

Which means that, for the purposes of this work, I embraced linguistic Purism, and did a lot of researching of older sources (including some quality time spent with the Iliou Encyclopaedia, Νεώτερον Εγκυκλοπαιδικόν Λεξικόν Ηλίου, 1945–1960). And I found it a lot of fun!

There were a lot of discoveries along the way:

  • The occasional faux ami with Ancient Greek (this is still, after all, meant to be a picture dictionary of Ancient Greek, not Puristic Modern Greek.) For example, see my startled discovery that, whereas Modern Greek differentiates drying something of excess moisture (στεγνώνω: plates, hair, clothes) vs drying something up (ξεραίνω: fruit, rusks, mummies), Ancient Greek used ξηραίνω for both. As a result, I was adamant that a hair dryer had to be a στεγνωτήρ, and ξηρός with relation to hair only made sense as dry, flaky hair as opposed to dried, not wet hair. Not so: the semantics of Greek have in fact changed over the last two thousand years. (It took a couple of passages tucked away in Aristotle for me to work that out.)
  • Old dictionaries of Greek (pre-1850), which have become widely available thanks to Google Books, are very valuable for working out how Greek used to deal with Modern concepts before the influx of French, and indeed even before the influx of Puristic coinages. Theocharopoulos (1834), for example, or Daviers (1844). And I continue to have a lot of time for Hepites (1912). (I have even more time for Dehèque (1825), because it captures a lot of early vernacular Greek, and I found it very useful in my time at the TLG; but Dehèque did not turn out to be as useful for this particular exercise.)
  • I was taken aback by how modern some of the concepts in the dictionary really were. There were equivalents of buttons in antiquity, for example, but the clear distinction between buttons, pins, and brooches is quite recent, and one that even 19th century vernacular dictionaries struggled with (θηλυκωτήριον: “something you insert into a plug [“female”]). The notion of nightclothes was meaningless in Ancient Greece, which is why the closest terms available are pretty much just blankets. And so on.
  • Purism ran out of steam in Greek, and French loans, took over, by the early 20th century: certainly by the 1930s. The early Puristic rendering of zipper in the 1910s as τορμοσυνάπτης “peg-linker” for example was likely coined too late to prevail against French φερμουάρ, and it is quite forgotten now. I could find no trace of a Puristic rendering of “fashion model”, a concept which would have been popularised in Greece by the 1930s: only the French μανεκέν and the Italian μοντέλο. And indeed, even “roulette” had no rendering but the French ρουλέτα and the Italian ρολίνα as early as the 1860s.
  • There is nonetheless a large body of Puristic coinages that have stuck from before the 1930s, and there is a smaller number of Puristic coinages that are succeeding and taking root to this day; γενόσημος for “generic”, for example, or λογισμικό for “software”. This of course varies by domain—sometimes seemingly randomly: there is a full Puristic terminology for soccer (although some English words have persisted in usual usage), but the terminology for tennis is substantially English. (That is not *that* random: tennis has a long history in Greece, but it was not mainstream until fairly recently, let alone reported on in the press; so the pressure for Greek-based vocabulary was simply not there, the way it was for soccer. There has never been a Hellenic term for “tennis racket”, for example: it has always been ρακέτα.)
  • And of course, there is more purism in official usage than in colloquial usage, to this day. This becomes really obvious with automotive terminology: the local garage and the Ministry of Transport have completely different vocabularies: e.g. πεντάλι vs ποδομοχλός/ποδόπληκτρον for “pedal”.
    • Whenever the Glorious Ancient Ancestors are discussed, there is an immediate and unsurprising recoiling from modern loanwords. For example a hairdresser’s cape is always a μπέρτα < French berthe; but the discussion for schools of an ancient sculpture depicting a hairdresser could not use anything but the ancient-looking περιώμιον.
    • There are ebbs and flows in fashion in language, and there has been some movement back towards more archaic usage in Greek in the past decade or so, as a reaction to the “victory” of Demotic in the 70s and 80s. I have had a waiter offer me an ἀπόσταγμα for “spirits”; I don’t think that would have been possible twenty years ago.
    • The parallel legal texts translated for the European Union have been a rich source of Puristic coinages (prominently figuring in online search engines like and
    • Online shopping catalogues are a boon for purism as well: the individual items on sale often use loanwords in their descriptions, but the categories of items in the sidebar (which are, after all, formal ontologies devised by boffins) tend to use Puristic terms, or at least more abstract terms. That was particularly noticeable on (“Scrooge”, i.e. “Thrifty”). This is not a great example (I couldn’t find the one that made me sit up and take note), but to give a poor example: the category is Διανομείς Καρτών “Card Distributors” (which the site feels obligated to translate in English (!) as Card Shufflers, but the individual items are ανακατευτήρας “shuffler”, σαμπό < French chabot, μοιραστής “sharer”.
    • A somewhat unexpected source of purisms has been the description of stock photos online. The usual term for “hair roller” for example is μπικουτί < French bigoudi; but the Puristic βοστρυχωτής shows up instead in photo search engines like Again, I strongly suspect the tag words for these sites are coming from formal ontologies translated by boffins, rather than colloquial live translation.
  • A lot of Puristic coinages were of course awkward calques from French; and in the context of trying to use Modern Ancient Greek, some of them are just too awkward to be palatable. So κόμμα for “political party” is just a calque of partie, and “fraction” really would not make immediate sense to someone who didn’t know the etymology of partie. The Classical phratry “sub-tribe” would at least make more sense as a social sub-grouping (although its modern use to mean “faction” is itself anachronistic.)
  • Similarly, some attempts to reimport Ancient Greek terms into Puristic would be just too loose to work: καταιονάω has been used for “to shower”, but its original meaning in Hippocrates is “to foment, i.e. to bathe with warm lotions”.
  • Liddell-Scott consciously purged itself of Mediaeval words with each successive edition, where they were obvious (attested in Christian theologians). Where they were less obvious, or more relevant to Classics (in scholiasts and mediaeval dictionaries), they were kept: user beware. That applies for example, notoriously, to στοίχημα in the modern sense of “wager”; it also applies to late-attested derivations like ἀποχαιρετίζω “to farewell”; ἀλφάδιον “carpenter’s square; modern: water level” (so called because they were A-shaped) is also clearly mediaeval; and the meaning “divorce” for διαζύγιον is no earlier than Arethas in the 9th century.
  • There has been a little bit of work on Modern coinages in Ancient Greek done outside the context of Puristic and Modern Greek: the Akropolis World News, for example, or the Ancient Greek Wikipedia. These attempts are welcome, but not infallible: ἀσθεν-ούχ-ημα instead of ἀσθεν-όχ-ημα for “ambulance”, for instance, is incorrect. I am biased towards coinages by Greeks, but those coinages are mostly morphologically reliable.
  • … Not always though! I put my foot in it when I introduced myself to the θαμῶνες “regulars” of Textkit: that is a modern derivation from θαμά “often”—and an obvious calque of French habitué; but it is also a derivation impossible in Ancient Greek. ἐλατήριον for “spring” is another such modern error: in Ancient Greek it is a purgative (“that which drives out”), and the modern sense “spring” has been affected by the etymologically related ἐλατός “ductile”.
  • The Centre of Research into Technical Terms and Neologisms of the Academy of Athens (unsurprisingly) was a last bastion of Purism, and they were still suggesting Hellenic coinages for technical terms until the late 2000s. I got a lot of value out of their Athens Olympics volume of sporting terms. (As should be clear, they had a lot of coining work to do with tennis, and not much with soccer. In fact, with some tennis terminology, they just balked: there is no Hellenic neologism proposed to counter σερβίς “service” or ρακέτα “racket”.) The academy’s French- and Icelandic-style work on puristic coinages has attracted derision (see this this newspaper review of the Olympics volume), and the centre has given up on proposing Puristic alternatives in the past few years, now that they are under new management: they simply can’t keep up with the influx of terminology, and they aren’t being taken seriously as an authority for terminology, so they have now switched to descriptive rather than prescriptive work.
  • By the way (as you may have guessed): unlike the people who actually live in Greece, I find these Puristic coinages charming and enjoyable, and I am saddened at Greek giving up and borrowing English words wholesale. But its their language, they get to make it impure and parasitic. (And of course, it’s not like this kind of thing hasn’t happened before. Like, say, with French a century ago. Or Turkish a century before that.)
  • There are a few nice anecdotes to be had. Greek has stuck with referring to bronze medals as copper medals (χάλκινα), because they were copper in the Athens Olympics of 1896, and damned if they were going to pay any attention to the switch to bronze in 1900. On the development of Greek conventions for telling time with minutes, I will post in a future article…
  • I knew this, but others may not: The Perseus copy of LSJ has systematically mis-stressed words when it filled in the prefixes of lemmata. (The LSJ source text would give forms derived from a headword by just their suffix, e.g. λῐθολογ-έω “build with unworked stones”… -ος “one who picks out stones for building”.) Perseus often got the completions wrong (e.g. λιθόλογος, not the correct λιθολόγος), and unfortunately sometimes, like in that instance, you need to know the meaning of the word to know where the accent should go. (Other times, the accentuation they came up with is impossible.) The TLG copy of the LSJ spent a lot of time correcting these; unfortunately I’ve lost access to it, but the rest of you have not.

Our work concluded in January, and you can see the results:

The dictionary does not include an English–Greek glossary, or the images (although the Textkit forum thread includes most of them); that work could be done by someone else, but there will be some difficulties using the dictionary without them. To get the most value of the dictionary for now, you should obtain a copy of Albert’s Latin original, and use them in parallel.

I recommend the linguistically curious go through my discussion on Textkit: it’s somewhat dry in that it follows the dictionary page by page, but there are some pleasant surprises to be had in there.

Gloriana, as refracted by Alkaios

By: | Post date: 2019-02-21 | Comments: 1 Comment
Posted in categories: Culture, Modern Greek

Akis Alkaios was one of the great Greek lyricists of the past fifty years, in a culture which valued and cultivated the great lyricist. In his biggest hits, With a Canoe and Rosa, he was darkly allusive, yet still successfully universal and moving—like his great contemporary Manos Eleftheriou. (Alkaios had to insist against the record company on “the land of the Visigoths” being mentioned in a zeimbekiko pop song.) I’ve hyperlinked the translations by “Ross”, which are remarkably good by the standards of, the Greek lyrics database.

Those lyrics date from the endpoints of Alkaios’ mature period, 1982 and 1996. As Wikipedia notes, his youthful period was marked by leftist protest songs:

With his record Embargo (1982) the lyricist immediately marked out an identity apart from his politically engaged contemporaries, in that he also wrote as a citizen of the world, expressing the desire for world freedom with a theoretical Marxist grounding.

And in that period, it goes on to say, he was liberally influenced by Mayakovsky, Brecht, and Wolf Biermann.

As I was perusing to find decent translations (like those by “Ross”), I happened upon a poem by Alkaios from his protest period: it was published in his 1983 collection of poems also titled Embargo.

I am culturally Greek. I am culturally Anglo. I am not both at any given instant in time. Which is why I did a double take, when I saw a poem about Gloriana, as seen through the lens of Greek left populism. Elizabeth; or, Epithalamium, 1600 AD. This link is more stable than that on

Οι ρεβεράντζες οι αυλικές πολύ μ’ αρέσουν
Το βιργινάλι η ιερακοτροφία
Κι οι μενεστρέλλοι σαν στα πόδια μου θα πέσουν
Μα πιο πολύ αγαπάω την Αγγλία

Το νιτερέσα-μου φυλάνε οι φτωχοί
Σα να ‘τανε δικά- τους νιτερέσα
Μοιάζει η Αγγλία με ολόγιομο πουγγί
Μα πιο πολύ αγαπάω το που ‘χει μέσα

Language play is one of the things that gets sacrificed in translation; and language play—specifically, register play—is one of the things Alkaios excelled at. The final verse of With a Canoe is wrenching, with is comparison of the singer’s lovelorn body with “a cheap shooting range, / where foreign conscripts train, cursing”. It is all the more wrenching because the previous line speaks of Attica as a “pallid quarry”, violently juxtaposing an ancient Greek and a Turkish word (φαιό νταμάρι).

That gets lost here: the sneer of the low (Italian) word for “interests”, νιτερέσα, the awkward folksiness of the syntax in the final line and its bathetic rhyme, “what it contains” (το που ‘χει μέσα). What gets lost even more irreparably is how the Elizabethan cultural references sound in Greek. Curtsies, virginals, minstrels: these are familiar in English to generations bred on Shakespeare. In Greek, ρεβεράντζες, βιργινάλι, μενεστρέλλοι are utterly exotic and alien; he might as well be writing about the marvels of the Safavid court. That, you can’t communicate in English.

I’ll translate it anyway:

I love the curtsies of my court,
and hawking, and the virginal,
and minstrels playing for my sport.
And I love England best of all.

The paupers hold my interests close,
close as their own, each passing minute.
England is like a purse of groats;
And more than it, I love what’s in it.

Updated post on the language of Syros

By: | Post date: 2019-02-13 | Comments: No Comments
Posted in categories: Modern Greek

I have expanded my old Quora post, and just had it published in Greek on Nikos Sarantakos’ blog:

The expurgated and unexpurgated online versions of the earliest dictionary of Macedonian Slavonic

By: | Post date: 2019-01-29 | Comments: 1 Comment
Posted in categories: Other Languages

In the leadup and midst and the aftermath of the Prespa Agreement, the Macedonia naming dispute has flared up again within Greece, and it’s never been terrain I’ve been enthusiastic about wading in to. I guess I’m on the side of those pro, being an διεθνικιστής “internationalist” = “anti-nationalist”, as my non-“internationalist” friend George Baloglou has smirked at me. (I’d prefer to refine it as an open, civic nationalist; I wouldn’t be moved to tears by Giannis Antetokounmpo’s assertion of love for Greece in the face of Greek racism, if I was a complete anti-nationalist.)

But there was a bit of misleading going on in a Facebook thread I waded into, that I’d like to correct.

There’s a language in the language that I often refer to in English as Macedonian, as indeed just about everyone in English does, and that I sometimes refer to in English as Macedonian Slavonic.

There is a discourse that objects to the name, and part of the armament of that discourse is to deny that it is a distinct language—in particular, a language distinct from the closely related language, Bulgarian.

Now, the earliest record of the Macedonian (Slavonic) language is an anonymous glossary written in the late 16th century. The glossary reflects the dialect of the Kastoria/Kostur region, and one of its phrases is oit koja strana da pojdime vo Bogasko “how do we get to Bogasko”; Vogatsiko (Macedonian: Bogatsko) is a village 15 km away from Kastoria. (Vaillant adds that “the village is now completely Greek, as Georgios Hatzikyriakes was happy to note in his Σκέψεις καὶ ἐντυπώσεις ἐκ περιοδείας ἀνὰ τὴν Μακεδονίαν [Athens 1906, pp. 60–61], a ‘topographical, historical, archaeological’, and not least patriotic survey of Macedonia.” In one of those ironies of fate that the Balkans abound in, it was the ancestral village of the Greek nationalist and politician Ion Dragoumis, who was one of the major pro-Greek agitators in the Macedonian Struggle—or as the Bulgarian Wikipedia terms it, Greek Armed Propaganda In Macedonia.)

(Why yes, Dragoumis is related to Dragomir.)

The grammar was published in 1958, and its title page is used in online discourse as a refutation of those who would say that Macedonian is not a real language:

Un Lexique Macédonien du XVIe Siècle. Par Ciro Giannelli avec la collaboration de André Vaillant. Paris: Institut D’Études Slaves de l’Université de Paris. 1958

“See? There’s a 1600s dictionary calling it Macedonian!”

Well… no. There’s a 1600s dictionary of a language variant we now call Macedonian. The dictionary does not call it Macedonian, nor was it likely to have, and it does not take the position that the contemporary international community takes on Macedonian. That does not argue against what the language should be called; it just does not argue for it, either.

I would not have waded in publicly, except that the link posted to the dictionary, and the version most readily found online, is the version on and It has “helpfully” taken out the foreword and commentary, and left just the lexicon. It has added a summary in English of the foreword and commentary, and it is a very very brief summary, that does not do the edition justice.

(For example: Giannelli, the scholar who found the manuscript and who wrote the foreword, speculated that the compiler of the glossary might have been a native speaker, and that speculation is reproduced in the English summary. The actual linguist involved in the edition was Vaillant, who wrote the linguistic commentary, and his conclusion was that there were too many grammatical errors for the compiler to have been a native speaker: he was likeliest a Greek cleric who worked in the area, and was curious about the language spoken there.)

Rather more offensive to me, the copy chops out the first sentence of the dictionary. The sentence where the compiler gives a name to the language he is recording.

Over the years, I have come across resources from the site. If you go looking for a bias in the site, you might notice that there’s a bit more Bulgarian presence than you might expect from a Pro-Macedonian website. But the site puts up its sources unexpurgated, and that is to its credit.

Those sources include the Giannelli–Vaillant dictionary.

Now, what Wikipedia says about Macedonian ethnic identity—which reflects on the nomenclature around the language—is fairly uncontroversial, at least in most circles:

The concept of a “Macedonian” ethnicity, distinct from their Orthodox Balkan neighbours, is seen to be a comparatively newly emergent one. The earliest manifestations of incipient Macedonian identity emerged during the second half of the 19th century among limited circles of Slavic intellectuals, predominantly outside the region of Macedonia.

[…] Yet, throughout the Middle Ages and up until the early 20th century the Slavic population majority in the region of Macedonia were more commonly referred to (both by themselves and outsiders) as Bulgarians. However, in pre-nationalist times, terms such as “Bulgarian” did not possess a strict ethno-nationalistic meaning, rather, they were loose, often interchangeable terms which could simultaneously denote regional habitation, allegiance to a particular empire, religious orientation, membership in certain social groups.

So you’ll have already guessed what someone in the region in the 1580s would have referred to Macedonian Slavonic as. And indeed, the first sentence of the dictionary, which the copy chops out, is:

Ἀρχ(ὴ) ἐν Βουλγαρίοις ῥιμάτου, εἰς κινῆ γλότα ἐρχομένη
Beginning of words in Bulgarian, coming [= translated] into the common language [= Greek]

Again: that does not argue against us now calling the language Macedonian. It does not argue for it, either.

For what it’s worth, the question of Bulgarian vs Macedonian as the name of the language does not occupy the attention of either Giannelli or Vaillant for a second. As far as they were concerned, back in 1958, it was the language of the Kastoria region (the Kostursko, as Vaillant adds); and the way for Italian and French scholars to refer to the langauge of the Kostursko in 1958 was as Macedonian.

Karamanlidika orthography

By: | Post date: 2019-01-26 | Comments: 6 Comments
Posted in categories: Modern Greek

I blogged about Phanariot in the last post, but what I actually wanted to talk about was something far more tangential.

Phanariot, as we discussed, was filled to the brim with Turkish loanwords. Phanariot was still Greek, and it was still written in Greek script. That included the Turkish loanwords in the Greek.

But the Phanariots using Turkish words pronounced them in accurate Turkish; and they were concerned to write Turkish words in Greek script with phonetic accuracy. So they employed the conventions of Karamanlidika.

The Karamanlides were a Turkish-speaking Greek Orthodox people living in Karaman and Cappadocia; being Christians, they were subject to the 1924 population exchanges, and were resettled in Greece. Whether they were originally Karaman Turks, or Turkicised Cappadocian Greeks, or, much more plausibly, both, is hard to know, and not terribly relevant here anyway.

What is relevant is that script followed creed in the Ottoman world; so if a Greek Orthodox Christian spoke Turkish, and was literate, they would read Turkish not in the Arabic script, like their Muslim colinguals, but in Greek script. As an extension, Karamanlidika was the name given to Turkish written in Greek script — whether it was written by or for Karamanlides, or by the Greek-speaking Phanariots, who sprinkled their Greek so generously with Turkish.

With the opening up of Greek academe to the Ottoman past, there has been much research into Karamanlidika in recent years, and Evangelia Balta of the National Hellenic Research Foundation has been the main researcher active. Her website appears to be down currently, and the most accessible resource online for Karamanlidika is a somewhat unexpected source.

The Karamanlides moved to Greece. Their speaking of Turkish was frowned upon in the new country, and was not something they passed down. They did pass down other aspects of their culture, though, like their cuisine. One of those they passed their culinary heritage down to was Fanis Theodoropoulos, who has opened a Karamanli restaurant in downtown Athens: Τα Καραμανλίδικα του Φάνη, “Fanis’ Karamanli Foods”. And with the web address, Fanis has felt it proper to include a blog on the restaurant website, covering not just the latest news and offerings of the restaurant, but also the culture and language of the Karamanlides and the Cappadocians. Including their script.

Thus, Samples of books printed in Karamanlidika, from the 16th to the 19th century. The 16th century is a reference to Martin Crusius printing a Karamanlidika text in his 1584 Turcograecia, the first Western study of Modern Greek. Other than that, the samples are almost all 19th century, with one text from the late 18th.

Now, there are several phonemes of Turkish that cause difficulty transcribed into Greek: none of <c ç h ı ş ö ü> /dʒ tʃ h ɯ ʃ ø y/ are present in Standard Greek. But as it turns out, <b d g> also pose a long-standing difficulty for transliteration in Modern Greek. Modern Greek has the phones, but not necessarily the phonemes: in many dialects of Greek, and in the Standard Modern Greek of older speakers, they occur only prenasalised, as reflexes of prenasalised stops. In other dialects, and in younger Standard Modern Greek, the prenasalisation drops off: <μπ ντ γκ> Ancient /mp nt ŋk/ > Older Standard Modern Greek /mb nd ŋɡ/ > Younger Standard Modern Greek /b d ɡ/.

Which means, sure, if you want to transcribe /b d ɡ/, you’ll use <μπ ντ γκ>. But given the history of Greek, and the variation in pronunciation, you’ll also use them to transcribe /mb nd ŋɡ/. And, indeed, /mp nt ŋk/.

The pronunciation of all of these has fallen together in Modern Greek, and Greek-speakers are not necessarily aware of it. Dante, for example, was transliterated as Δάντης, <ðantis>; but the Modern pronunciations of the name are /ðandis/ and /ðadis/. Greek loves using the German loanword lumpen to refer to “trash culture”, from the Marxist notion of the Lumpenproletariat, the underclass. But they’ve only ever seen it as λούμπεν, and they know less German than they do English, so they’ve pronounced it as they read it. Hence when time came for a Greek satirical website to adopt the word as its name, it inevitably ended up as

What this means is that if you’re trying to transcribe /b d ɡ/ in Greek, <μπ ντ γκ> might have ended up the only choice in contemporary Greek, especially given how integral /(m)b (n)d (ŋ)ɡ/ are to the language. But if you are not writing Greek at all, especially back in the 18th and 19th century, you don’t have as strong a motivation to lean on the Hellenic <μπ ντ γκ>. Especially when you’re already having to deal with /dʒ tʃ h ɯ ʃ/ anyway.

There is not a unified Karamanlidika transcription, but there are two main conventions one can see in the editions: one with diacritics, the other without. (The one diacritic it does admit is the diacritic 19th century Greek already admitted for Demotic: ι̮ for [j].) The transcription without diacritics fails to make some distinctions, like that between /mb/ and /b/ (both μπ), or between /ʃ/ and /s/ (both σ); it does however take advantage of the fact that Turkish has no /ð/, to transcribe /d, dʒ/ as δ, δζ. And at least by the 19th century, to judge from the samples, Karamanlidika had settled on using a spare Greek grapheme for /i/, η, to transcribe /ɯ/, in both the diacritic-based and the non-diacritic transliteration.

… Except when there’s an actual Greek word in Karamanlidika. When the word is Greek and not Turkish, Greek historical orthography is respected, and eta means /i/. So in fact, to read Karamanlidika accurately, you are sort of expected to know something about Greek anyway.

The rulebook of the Cappadocian Educational Fraternity, for example, has the following code switch (Greek words in boldface, and I am guessing at the Turkish via Google Translate):

  • συνιστᾶται … Καππαδοκικὴ Ἐκπαιδευτικὴ Ἀδελφότης, γι̮άνι Καΐσεριε ἐπαρχίασηνην Οὐχουββέτι τεδρισιερί τεσ̇κὶλ ὀλουνμούσ̇δο̇υρ
  • Kapaðokiki Ekpeðeftiki Aðelfotis, yani Kayseriye eparxiasının Uhuvveti tedrisiyeri teşkil olunmuşdur
  • The Cappadocian Educational Fraternity, that is the educational brotherhood of the district of Kayseri, has been created

In a Greek word like Καππαδοκικὴ, eta is /i/; in a Turkish suffix like -σηνην, even if it is attached to the Greek word ἐπαρχία, eta is /ɯ/. Similarly, the list of Bible books—

—contrasts the Gospel of Luke, ΛΟΥΚΑΣΗΝ Lukasın, with the Gospel of John, ΙΩΑΝΝΗΣΙΝ Ioannisin. Eta is /ɯ/ in the Turkish suffix of Lukasın; it is /i/ in the Greek name Ioannis.

But I’m more interested in the diacritics used with Karamanlidika. Karamanlidika with diacritics adopted the overdot as a universal diacritic. So <χ̇ σ̇ π̇ τ̇ τ̇ζ ο̇υ ο̇> are /h ʃ b d dʒ y ø/. We have just seen δ used next to τεσ̇κὶλ: because it could not mean /ð/, it was still used to mean /d/ in some transcriptions, though not others: the Bible book list is titled ΑΧΤ̇Η ΑΤΙΚ ΙΛΕ ΑΧΤ̇Η Τ̇ΖΕΤ̇ΙΤ̇ΙΝ ΦΙΧΡΙΣΤΙ Ahdı Atik ile Ahdı Cedidin fi-Xristi “The Old Testament and the New Testament of Christ”. Similarly, while the Bible books use τ̇ζ for /dʒ/, they do not use τσ̇ for /tʃ/: Turkish has no /ts/ (hence it transliterates Tsipras as Çipras), so the Bible is content to use τζ (/ts ~ dz/ in the 19th century) for /tʃ/: Ο̇ΥΤΖȢ̇ΝΤ̇ΖȢ̇ ΠΑΤ̇ΙΣ̇ΑΧΛΑΡ <Ọutzọunṭzọu Paṭiṣaxlar> Üçüncü Padişahlar “3 Kings” (corresponding to Western 1 Kings). And there does not seem to be any use of γ̇: γγ/γκ have been accepted as /ɡ/, since Turkish has no /ŋɡ/.

The Karamanlidika overdots for /b d/ have made enough of an impression on Greek scholars, that they have been reused rather far from Greek transliterations of Turkish. The plays of the Cretan Renaissance were written in Roman script, using Italian orthographic conventions; Italian of course differentiates /b mp mb/, and just as Phanariot in Greek script preserved the non-Hellenic distinctions of Turkish loanwords, so too did the Italianised Cretan of the Renaissance preserve the non-Hellenic distinctions of Italian loanwords. But those texts still ended up published in Greek script, both at the time and in modern scholarly editions. And in at least one instance (I think it’s Alfred Vincent’s 1980 edition of Fortounatos, edited from the Roman script autograph manuscript), the Karamanlidika dots are used to differentiate the original’s <mp> and <mb>.

In the 18th century, though, there was a different way of writing /ʃ/. This merited a passing comment on p. 33 of Peter Mackridge’s paper on Phanariot:

For [b] [d] [ʃ] [dʒ] some writers use the diacritics that had been developed for use in karamanlidika earlier in the eighteenth century (π̇, δ̇, σ̈́, τ̇ζ), but most make do with the unadorned Greek alphabet: thus [kurdízo] is variously represented as κουρδίζω, κουρντίζω and κουρτίζω.

He adds in footnote: “For [ʃ] the sigma is in fact surmounted by three dots in a triangular pattern, but I am unable to reproduce this here.”


I’ve asked Peter, and he’s sent me a sample from the satirical comedy Το σαγανάκι της τρέλας. (Not “The frying pan of madness”, let alone “The saganaki of madness”: contemporary Greek σαγανάκι “small frying pan”, and any dish prepared in a small frying pan, like fried cheese, is a diminutive of σαγάνι < Turkish sahan “copper dish”. The Turkish word here is the unrelated sağanak: “The storm of madness”.) The comedy is attributed to Rigas Feraios, and was published in Lia Brad Chisacof. 2001. Ρήγας. Ανέκδοτα κείμενα, Athens. the text is published alongside the manuscript, and he has sent me two instances of the novel diacritic in question: pp. 158–159 ϗʹ μεσ̈́αλάδες “& torches” (Turkish meşale), and pp. 176–177 μεσαλοσ̈́άνον (= Greek μεγαλουσιάνον “bigwig”; in that case, the character is being used for [sj] < [ʃ], presumably a dialectal pronunciation within Greek.)

This is what the two words look like in the manuscript:

Peter has told me that “What this scribe (who is no doubt the author) actually writes is not three dots but more like an acute accent flanked on each side by a dot.” Other instances may have been triple dot, but this instance is more like the diaeresis + acute that he resorted to in print. Maybe, though they look like triangular triple dots to me in the manuscript anyway.

And there’s a reason they should. I didn’t give the entire footnote above. He continues:

Christodoulos Christodoulou informs me that this use of three dots follows Arabic practice, šīn [ʃ] being distinguished from sīn [s] in the same way in Arabic script.

That diacritic is from Arabic: it is the distinction between س and ش. There is precedence for mixing diacritics and letters from different scripts; Samaritan for example has combined Samaritan letters (related to Hebrew) with Arabic vowel diacritics. This is the first time I’ve seen it in Greek.

And there’s good reason Peter Mackridge had difficulty rendering a three-dot sigma in his 2017 paper. Unicode has the three-dot Arabic diacritic, U+06DB ARABIC SMALL HIGH THREE DOTS. But combining Arabic diacritics with Greek or Roman script is disastrous: the diacritics are designed for a completely different letter height. Luckily, Unicode does offer a triple dot diacritic compatible with Roman (and Greek): U+1AB4 COMBINING TRIPLE DOT. Unluckily ,the character was added in Unicode in 2014, which means font support for it is still minimal: among the fonts I have installed (and I have a lot), the diacritic only turns up in Google’s Noto fonts and Dehuti, and sigma with triple dot only looks presentable in the latter:

Phanariot: an apology for Schleicherian bias

By: | Post date: 2019-01-25 | Comments: 4 Comments
Posted in categories: Linguistics, Modern Greek

I was recently perusing Peter Mackridge’s paper Some literary representations of spoken Greek before nationalism 1750-1801, and I got sidetracked by an incidental footnote on diacritics use in Karamanlidika in the 18th century.

And now, to unpack.

Peter Mackridge is the emeritus professor of Modern Greek in Oxford. He has written a wealth of papers and books on Modern Greek, and has always been a keen observer, both of how the language works, and of the particularities of its diglossic history.

Latterly Peter has taken an interest in the language of the Phanariots. The study of the history of Modern Greek is a much broader and fuller thing now than it used to be; but the Phanariots have mostly been off-limits until now in any study of Modern Greek; and indeed, little has been said about 18th century Greek in general. As Peter notes in his paper, it is the chunk of the history of Greek least well served by surveys, dictionaries and grammars. (The 16th and 17th century outside of Crete have not been that much better served, but Eleni Karantzola has been active in redressing that imbalance over the past decade.)

There are two straightforward reasons why Phanariot Greek has been something Greek historical linguists have instinctively shrunk away from; and I will admit to have shared that prejudice in my own time. The Phanariots were affluent Constantinopolitan artistocrats under the Ottomans, who occupied high office in the administration of the Ottoman Government, and particularly in the rule of Romania on behalf of the Ottomans. They were urban and urbane, they were multilingual and cosmopolitan, they were loyal Ottomans and unsavoury intriguers; and their Greek was full to the brim of Turkish and Turcisisms.

Almost as full as Greek in Greece now is of English and Anglicisms, in fact.

Modern Greek historical linguistics has had some blind spots it’s needed to get past. That you need to understand Kartvelian languages to work out Pontic, for example. Or that Greek borrowed words from other languages even when it isn’t obvious where they did. Or that there is a lot more Puristic in Modern Standard Greek than the ostensive victors of the diglossia wars would like to think.

And a more pervasive bias than that, one I’ve shared, is a Herderian and Schleicherian view of language change, as tied up with the expression of ethnicity, and as paralleling the evolution of lifeforms. There are sophisticated takes on those views which are still current: historical linguistics continues to have a lot to learn from evolutionary biology, and much of sociolinguistics is about the nexus between language and identity.

There are also unsophisticated takes on those views. Not just Herder’s Blood and Soil nationalist romanticism, or Schleicher’s original notion that there are primitive languages for primitive peoples (or even his subtle variation, that there are overcomplicated languages for primitive peoples). Those have been rejected in polite company; but there are lingering romantic notions in thinking about language change that have outlived them. For example, that rural and oral language is the only true object of study of the historical linguist, and that urban and written language is subject to contaminating, artificial influences, and of secondary interest, if of any interest at all. It’s a naturalistic bias, and it’s a puristic bias. You can see how easily it can turn to cultural purism, with the untutored village folk seen as the only true teachers of the language, and with the learnèd influence on the language derogated, if not disavowed; something that gets in the way of forming an accurate picture of how Standard Modern Greek works to this day.

You can also see why the only mention Triantafyllidis makes of Phanariot in his monumental history of Greek is to raise his eyebrow at how much Turkish there was in their written texts. An untutored Herderian villager would never speak such a farrago, surely.

Any non-Greek linguists sneering at this point would do well to examine their own conscience. The dismissal of written language as not the proper domain of linguistics is a reaction to generations of prescriptivist dunderheads; but it is a biased reaction all the same, and it does not admit the fact that spoken language in literate societies is profoundly influenced by whatever neogrammatically incorrect nonsense takes place in written language. (Nor will fleeing to the Rousseauvian paradise of preliterate societies give you back your pristine language organism: preliterate societies are just as subject to changes in register and genre, and contamination between them.)

There is artifice in human language. There is a lot of artifice. And that is nothing to be ashamed of.

So, the bias against looking at Phanariot is a deep one. It’s informed by comic-book tribal politics: the Phanariots were aristocrats and intriguers, they were the bad guys. It’s informed by nationalism and purism: the Phanariots were collaborators and Turcophiles, they did not speak pure Greek (although, as Peter informs me, some of them were consciously puristic in their Greek, the page after another author seems to be cramming as much Turkish into their text as they can get away with). It’s informed by Herderian Romanticism: the proper object of historical linguistics is to be found among shepherds and peasants, not among dragomans and patriarchs. But it’s also informed by Schleicherian Romanticism: the proper object of historical linguistics is the “natural” evolution of language, and what the Phanariots were doing was anything but natural.

I am glad that in my dissertation, even though I dodged Herder to the extent I could, I did not dodge Schleicher. I was doing a global dialectal survey, and to do it I needed to work on a simplified model of language change, factoring out sociolinguistics as much as I could, rather than work on everything at once. It was a better use of my time to survey all the peasantries of Greece for how they were using the complementiser I studied, than to expend effort diving deep into what the townsfolk of Athens or Leonidio were doing, and how Puristic Greek (or, in the case of Leonidio, how second-language Tsakonian and dying-Tsakonian conspired with hubris) influenced it.

But it was a simplification, A simplification that had me keep seeing Phanariot in the corner, and thinking, nah, that’s not the proper object of my study.

I’m glad Peter is on the case now. And I’m sorry my next contribution to that topic is going to be so tangential to what he is on the case about.

Updated post on the etymology of βύσσινο “sour cherry”

By: | Post date: 2019-01-22 | Comments: 1 Comment
Posted in categories: Uncategorized

I have expanded my old Quora post and just had it published in Greek on Nikos Sarantakos’ blog:

Yes, this does mean I’m coming back. Eventually.
  • Subscribe to Blog via Email

    Join 324 other subscribers

  • May 2023
    M T W T F S S
    « Nov