Ἡλληνιστεύκοντος

The reason why I picked Markos Vamvakaris’ song Ο Ισοβίτης for my ruminations on hiatus is its last verse, with its startling macaronic juxtaposition:

όπως τον Έκτορα ο Αχιλλεύς τον έσουρνε στο κάρο
Like Achilles dragging Hector in his cart

The clash isn’t just thematic of course, it’s also linguistic: Hector and Achilles are solemnly invoked in Puristic phonology; Achilles desecration of Hector’s corpse—τον έσουρνε στο κάρο, “dragging in his cart”—lands in a colloquial thud. It’s the kind of register switcheroo that Greek—and I imagine other diglossic languages—can exploit for artistic effect; it’s one of the main things lost in translation in Cavafy, for example.

So Homer had Achilles bind Hector to his δίφρος, his chariot-box; the word is obscure enough that Greeks now speak of his ἅρμα, the more generic word for chariot. Yet the hárma too is displaced in the song by a a vehicle rather more familiar to the 1930’s streetscape: κάρο “cart”. It’s a Romance word, and oddly enough, Alexander Pope chose the same word in translating the same passage: Proud on his car the insulting victor stood.

And when Achilles drags Hector, it’s not using the Ancient and Puristic verb έσυρε /ˈesire/, or even in the colloquial έσερνε /ˈeserne/; Markos uses the Athenian slang form έσουρνε /ˈesurne/.

Ironically, the sin of /ˈesurne/ against the standard language is that it actually preserves the upsilon of Ancient Greek (ἔσυρε /ésyre/.) It’s no good being an archaic dialect like Athenian, if none of the other dialects have stayed archaic—and if spelling pronunciation prevents people from realising that it’s archaic: that just gets your dialect called weird. (Hence Theodosius Zygomalas’ mistaken verdict in the 16th century, that Athenian was tragically the most corrupt dialect of Greek.)

On the other hand, the Homeric heroes are unrepentantly textbook in their pronunciation. The vernacular would demand Έχτορας /extoras/ instead of Έκτορας /ektoras/ for Ancient Ἕκτωρ /hektɔːr/. You will still on occasion hear κτ in learnàd loans pronounced as [xt], but it is decidedly out of fashion to write it so. But Alexander Pallis’ translation of the Iliad was written to follow Psichari’s ideal of a pure vernacular phonology; his Hector is indeed written down as Ekhtoras, but that’s not the translation Markos read in school, nor indeed the translation read in school now. Hector keeps his non-vernacular /kt/ in the song; and in the song, the non-vernacular /kt/ is jarring.

The modern ear is more shocked to hear the Achilles in his Puristic garb, as Αχιλλεύς /axiˈlefs/, a spelling pronunciation of Ancient Ἀχιλλεύς /akʰilleús/. To the modern ear, the only legitimate Demotic form is Αχιλλέας /axiˈle.as/, which has switched its third declension for the surviving first declension. Anyone still using the third declension (and unprononouncable) Akhilefs now is deemed ideologically suspect. Surely Markos should have known better than to use such a retrograde form.

But of course Markos in 1935 would have known no such thing. After all, there is nothing vernacular about the hiatus in /axiˈle.as/. It’s not the form he would have got at school; and it’s not the form Achilles would have, had it in fact survived as a vernacular name.

In fact, Αχιλλέας is a compromise form: it is a reconstruction of a mediaeval pronunciation, after the word switched declension, but before the vernacular’s grubby [j] got to it. So too βασιλεύς “king” /basileús/ became βασιλέας /vasiˈle.as/ in the Middle Ages, abandoning the now unpronouncable /vasiˈlefs/; but the modern vernacular form is βασιλιάς /vasiˈljas/, following the i > j /_V rule. Achilles should similarly have ended up as Αχιλλιάς /axiˈljas/ in the vernacular.

Noone has dared devise so self-consciously Demotic a form of Achilles… except, unsurprisingly, for Pallis. Hard though it may be for contemporary Greek speakers to credit, Pallis does in fact use Αχιλλιάς. But not only is Pallis an extreme of Psicharist phonology among modern writers; even Pallis was reluctant to clothe Achilles in that much synizesis. In Iliad I, he has 16 /axiˈle.as/, hiatus and all, and only two vernacular /axiˈljas/:

Μούσα, τραγουδά το θυμό του ξακουστού Αχιλέα, I 1
…
παρά άσ’ την μιάς και δόθηκε στον Αχιλιά απ’ τους άντρες I 276

So Pallis only dared break Achilles’ hiatus one time out of ten, and the subsequent vernacular standard didn’t even dare that much. Vamvakaris was not going to deliver a more Psicharist version of Greek than Psichari’s pupils. But nor was Vamvakaris going to come up with the post-Psicharists’ compromise form Akhileas—a form as artificial as the spelling pronunciation Akhilefs is, even if Greek speakers no longer realise it.

Of course, Akhileas was artificial and a compromise, but the post-Psicharists weren’t the only ones cowed by awe before Homeric names, and forced to Mediaeval compromises. Such a compromise can be seen in the name of the warrior Androutsos, a couple of generations before Psichari. By the start of the 19th century, Greek national awakening meant that Greeks started taking Classical first names, to assert that they too were Hellenes. And so Androutsos, born in Ithaca while has family was on the run from Roumeli, was called Odysseus.

Odysseus’ name is just as hostile to the vernacular as Akhillefs [axilefs]. The spelling pronunciation of the original name, /oðiˈsefs/, was unpronouncable and undeclinable. None dared come up with a Pallis-like reduction, which would rhyme with /vasiˈljas/. The contemporary standard has done the same: the Odysseus king of Ithaca is /oðiˈse.as/, and Odysseus Androutsos of Ithaca is recorded in Greek textbooks as /oðiˈse.as/.

But Androutsos’ contemporaries did at least vernacularise his name a smidgeon more than people now do. One of the many vernacular rules ossified thouɡh Puristic influx is that unaccented initial /o/ is dropped; so ὀλίγος /olígos/ “few” is now λίγος /ˈliɣos/, and ὡρολόγιον /hɔːrolóɡion/ “clock” is now ρολόι /roˈloj/. His comrades called Odysseus Δυσσέας /ðiˈse.as/, ‘Dysseas.

These are the shoals of Homeric proper names in Modern Greek; and this is why Markos used a more archaic version of Achilles then contemporary Greeks are comfortable with—not that /ektoras/ and /axile.as/, the versions they are comfortable with, are any more true to the vernacular phonology. And this is how Markos’ /ektora o axilefs/ sounds jarringly pedantic in the song, just as the mention of Hector and Achilles should.

It would be the cherry on the cake if his pronunciation of Hector and Achilles would also follow the pedantic hiatus of Puristic, and set up the phonological force field around them that the legal terms ισοβίτης and έφεση already carry. But in his scansion, Markos elides three syllables across word boundaries, in a most un-Puristic fashion:

όπως τον Έκτορα ο Αχιλλεύς τον έσουρνε στο κάρο
opws| ton e|ktora o ax|ilefs || ton e|surne| to ka|ro

Which means /ektora o axilefs/ is supposed to be pronounced [ektorw axilefs], in order to scan.

Well, you tell me what you hear; 2:53:

I head [ek.to.ra.o.a.xi.lefs], with each syllable distinct. What I hear is that, for all that his versifying has Hector cozy up to Achilles in synizesis, when he comes to singing the names, he balks. Even if he is singing in the voice of a murderous street thug, his street thug has been to school, and can’t shake the shade of Puristic, any more than Pallis could.

Which would be a nice example to end on, except that I have been listening to all of Markos’ early songs, and his singing does not support the conclusion. It turns out that Markos has trouble doing synizesis with a final -o, whether the words are colloquial or learnèd. The year before Ο Ισοβίτης, Markos recorded Στα σίδερα με βάλανε, “They’ve locked me in chains.” Yes, once again, the song’s subject is in prison for murdering a rival lover. Again, the metre is iambic heptameter, and again, there is a synizesis:

Φωτιά μεγάλη μʼ άναψες βρε άπιστη γυναίκα.
Μόλις θα βγω απ’ τα σίδερα θα σφάξω κι άλλους δέκα
fotja| meɣal|i m a|napses || vre a|pisti| ɣinek|a
molis| θa vɣo ap| ta si|ðera || θa sfa|kso kj a|lus ðe|ka
You’ve lit a great fire under me, you faithless woman
As soon as I get out of chains, I’ll kill another ten men

/θa vɣo ap/, “I’ll get out of”, scans as a single foot, two syllables: [θa vɣw ap]; and θα βγω απ’ is as vernacular a phrase as you can get. I hear Markos singing three syllables (1:18)—very distinctly:

So no, the hiatus in Markos’ singing of /ektora o axilefs/ is not indicative of anything.

Oh, did you notice the hiatus in /vre apisti/? With the vernacular-as-dirt vocative particle βρε “hey you!” I’d hate to think that disproves my entire argument; I’ll take the preponderance of hiatus around learnèd words as a statistical argument.

Which would be more convincing, had I actually done any statisics. THUD. Too honest for my own good, there…

We saw a couple of posts ago the rebetiko musician Markos Vamvakaris in the 1930s, being more subject to the phonology of Puristic than Greeks might now expect of a singer extolling the underworld. Such an expectation says more about the romantic notions fomented by centuries of diglossia, than it does about the linguistic realities of 1930’s Peiraeus. But your humble correspondent, too, is subject to romantic notions, and your humble correspondent, too, was surprised.

There’s another song Vamvakaris did, in 1935, that runs along the same lines. It has even more Puristic in it, and in that song too the Puristic results in hiatus. The hiatus is so strong, that Vamvakaris the singer ends up doing more hiatus than Vamvakaris the lyricist. And the trigger for the extra hiatus is a namecheck of antiquity. But there’s a lot to say about the song—not all of it linguistic. So much to say in fact, that I’m going to hold over the lyricist/singer clash till the next post.

The song in question is not the 1937 Ήμουνα μάγκας μια φορά, “Once I was a mangas“, whose namechecking is thick and obvious:

Once I was a spiv, though with an aristocrat’s vein
Now I’ll be a scholar, like wise Socrates.
I’d be Paris, and steal away Helen
leaving Menelaus with his heart crushed.
I’d like to be Heracles when I first saw you,
to chop your head of like the Hydra.
What else do you want me to do to make you love me?
With a mind like that, you’ll be wanting to get with Xerxes.

No, the song in question is from a couple of years earlier. Its injection of antiquity works, because it’s not laid on thick: it’s linguistically out of place—yet strikingly appropriate to its thuggish context. The song is Ο ισοβίτης, “The Lifer”:

I’ll allow myself my bourgeois indignation at the protagonist; we should remember, of course, that Markos could act, and wasn’t necessarily saying it was his own wife he was planning to immolate—any more than that he had a crush on a teamster in Ο Αραμπατζής, or that he was a housewife abandoned by a drunken husband in Ο γρουσούζης. Still, the mangas that Markos sung about were no feminists.

Of course, when I told a female friend (rather more clued in to gender politics than me), that I was writing about a singer before he stopped singing about wife-beating for more bourgeois topics, she retorted that wife-beating is pretty bourgeois. Touché.

The song has a visceral grimness to it, precisely because of its nonchalant thugishness, and that does make it arresting. Which makes the trick he pulls in the final verse all the more effective. It’s incongruous to invoke the Iliad in a song about some low-life in jail. And yet, this isn’t just lightly worn high school learning: Markos has learnt his Iliad all too well. Achilles’ wrath, which made him desecrate his opponent’s corpse, is condemned by Homer himself as “shameful”, ἀεικέα (Iliad XXII 395). Achilles’ wrath is no more highminded than the lifer’s planned immolation.

Or these reenactments, courtesy of the US secondary education system:


The song has a fair amount of Puristic words in it, as we’d expect of a song not only with a shout out to the Iliad, but also several mentions of the legal system: ισοβίτης “life sentence”, ισοβίτης “lifer”, έφεση “appeal”, ενόρκους “jurymen”—as well as πετρέλαιο “petrol” (“stone-oil”, a learnèd coinage), αιτία “cause”, and ραδιουργία “intrigue”. There’s potentially one more learnèd word, in a passage which isn’t terribly clear.

In the beginning of the final verse, Markos sings that he has μπουζούριασα το /xiti/. Neither word is extant now, and if we ask the internets, we find that μπουζουριάζω means to put someone in jail. slang.gr’s illustration is eloquent enough:

Τι κάνεις ρε στρατόκαυλε, με το μαχαίρι του ράμπο στην πορεία; Θα σε μπουζουριάσουν ρε καραγκιόζη!
What the hell are you doing, you army nutjob, carrying Rambo’s knife in a protest march? They’ll lock you up, you maroon!

Or another instance, from indymedia, with indymedia’s known anti-cop animus:

Κι εγώ θα θελα να κάθομαι και να τα παίρνω, αλλά το να μπουζουριάζω αθώο κόσμο, δεν είναι δουλειά, είναι ντροπή.
Oh I’d love to just sit around and earn money too; but locking up innocent people is not a job, it’s a disgrace.

In the thread at rebetiko.gr discussing this lyric, an etymology is offered from μπουζού “hiding place, jail”, which in turn is said to come from Italian bozzolo “cocoon”.

But of course, Markos’ Lifer isn’t supposed to be in jail for imprisoning his rival, but killing him. It turns out that μπουζουριάζω has a second meaning, “to eat up”. That definition is given in a 1932 song, Το λεξικό του μάγκα, “The mangas’ dictionary”. The song actually predates Greek recordings of rebetiko (those are mandolins on the recording, not bouzoukis), but it describes the lexicon of the same social circle, from the safe vantage point of the musical revue:

(3:16: Το μαχαίρι λέω λάσο και το τρώω μπουζουριάζω: “I call a knife a λάσο, and “eat”—μπουζουριάζω.)

Poster κκ in the rebetiko.gr thread has worked out that the second meaning applies in Markos’ song: to “eat someone” (τον έφαγα) is long-standing slang for killing someone, and Markos has made the expression more vivid by substituting a slang word for “eat”.

The problem is who the /xitis/ is that the Lifer has killed. Poster κκ—and just about all copies of the lyric online—assume it’s a χίτης, a Chi-man. The Chi-men were members of the paramilitary organisation “Organisation X” (Chi in Greek, of course); the Chi-men appear to have been rather more enthusiastic fighting communists than Nazis, and are now roundly reviled in Greece.

The catch with having Markos talking about Chi-men is, the Organisation was still ostensibly formed to resist the German occupation, and it was formed in 1941. Barring an undocumented talent for soothsaying, Markos is unlikely to have been singing about dispatching Chi-men in 1935. The only other likely match for /xitis/ is χύτης, a pourer—in particular, someone who works in a foundry (where metal is poured out).

So speculated by Aris in post #4. Now χύτης is a learnèd word, κκ’s response to Aris betrays the anti-diglossic romanticism that haunts all Modern Greeks:

Even if Markos really did mean a foundry worker, I imagine he’d use πασπαλιστής “smearer”, καρούλιας “reeler”, or some street word—not the kind of vocabulary you’d find in the Statistical Classification of Branches of Economic Acitivity or the Ministry of Finance.

… A learnèd word like έφεση “appeal” or ραδιουργία “intrigue”, you mean. True, it’s more of a surprise for how Markos names the victim in the song; but this is hardly the song to expect Markos’ Demotic to be pure and unsullied from officialdom.

There’s another suggestion on a forum that χύτης is short for χυτοσίδερο, “cast (poured) iron”, referring to prison bars. That would bring us back to μπουζούριασα meaning “imprisoned”, but grammatically it doesn’t stand.

But we’re supposed to be talking about hiatus.

Vernacular phonology, we have seen, avoids hiatus; and vernacular metrics reflects vernacular phonology. If your vernacular verse has a vowel next to another vowel, they are supposed to be slurred together into the one syllable; to have a metrical break between two vowels is poor versifying. Verse textbooks inveigh against it, but they inveigh against it because vernacular verse itself—in folksong, in the Cretan Renaissance, in the Heptanesian School—all avoided it. Thus the national anthem of Greece written by Dionysios Solomos, to pull up the first example I could think of, starts:

Σε γνωρίζω από την κόψη
του σπαθιού την τρομερή
σε γνωρίζω από την όψη
που με βια μετράει τη γη.
se ɣno|ˈrizo a|ˈpo tin| ˈkopsi
tu spa|θiˈu tin| trome|ˈri
se ɣno|ˈrizo a|ˈpo tin| ˈopsi
pu me| ˈvia me|ˈtrai ti| ɣi

The metre is trochaic tetrameter: Dumdee Dumdee Dumdee Dumdee, Dumdee Dumdee Dumdee Dum. Eight and seven syllables to the line. If you take all the /i/s in the IPA at face value—with all hiatus—you’re going to have several syllables left over.

For the metre to work, μετράει is two syllables, /me.ˈtraj/ rather than /me.ˈtra.i/: perfectly vernacular, and that’s how the word is still pronounced. βια is also vernacular, reduced to one syllable; the word has now been displaced by the learnàd βία, in two syllables.

(At this point I could get sidetracked by the debate over whether Solomos used the word with its vernacular meaning of “haste” or its learnèd meaning of “violence”—a debate held in Greek parliament no less. I won’t get sidetracked this time, but I think it’s clearly the latter, which would make Solomos deliberately vernacularised the word, and consistently did so to avoid hiatus, as he actually wrote down in his notes on metre. The irony is, Solomos’ native dialect of Zante actually *has* hiatus; so he pronounced both “haste” and “violence” as /vi.a/, but followed standard Greek in versifying both as /vja/. Cornaro before him may have also vernacularised the word βια—though I think Erotokritos II 215 still refers to haste, even if in reference to a lion.)

The slurring of a vowel before another vowel (synezesis) also applies across word boundaries; so γνωρίζω από /ɣnorizo apo/ is pronounced as four syllables, [ɣno.ri.zw a.po].

And that slurring is regular in vernacular verse. Versifiers nowadays offend against it, because the language itself, tempered by Puristic, no longer finds hiatus offensive. But hiatus still sounds wrong in verse (if your ear is suitably trained), because verse as a tradition is aloof from the phonological mess of the spoken tongue.

Now Vamvakaris, it has to be said, was not much of a versifier. He does hiatus without good reason, and he does occasionally add or miss a syllable. So it’s not that we can draw ironclad conclusions from the hiatus in his verse. Still, we can see an informative pattern with respect to learnèd words.

Let’s try to jam his lyric to its metre—politikos stichos, iambic heptameter, the default Greek metre of the past millennium: ˘ ˊ ˘ ˊ ˘ ˊ ˘ ˊ // ˘ ˊ ˘ ˊ ˘ ˊ ˘

sti fi|laki| me kli|sane || isov|ia| ɣja se|na
tetjo| meɣa|lone| kajmo || epo|tises| eme|na
esi| se eti|a tu| kaku || ɣja na| me ti|ranu|ne
i pi|kres ke| ta va|sana || na me| strifo|ɣirnu|ne

tora| θa ka|no e|fesi || mipos| me vɣa|lun o|kso
kakur|ɣa ðo|lofo|nisa || ɣja na| se pe|tsoko|pso
na su| xina| petre|leo || ki iste|ra na| se ka|pso
ke mes| sto kse|ropi|ɣaðo || na pa| na se| peta|kso

efta| fores| iso|via || tote| na me| ðika|sun
ke sti| krema|la t a|naplju || eki| na me| krema|sun
psinis| enor|kus ði|kastes || tus pla|nepse i e|morfja| su
ke me| ðika|zun iso|via || ɣja na| ɣeni i| karðja| su

me ti| raði|urɣi|a su || buzu|rjasa| to xi|ti
ðixos| na θe|lo m e|kanes || na ɣi|no i|sovi|tis
tetja| meɣa|li ekði|kisi || an ti|ne kse|buka|ro
opws| ton e|ktora o ax|ilefs || ton e|surne| to ka|ro

A bit busy; let me explain.

• A foot is delimited by | : there should be two syllables per foot, but for the last.
• || is the caesura, the midverse break; synezesis, contracting two syllables into one across the break, is not allowed in traditional versification.
• Learnèd words are in italics.
• /j/ is used where synezesis has happened within a word, consistent with vernacular phonology. No learnèd words have a /j/.
• Blue is synezesis across a word boundary, consistent with vernacular metrics.
• Orange is an extrametrical syllable not explained by hiatus.
• Red is hiatus within a word.
• Magenta is hiatus across a word boundary.

Markos has hiatus at the caesura; since the break at the caesura is so strong, such hiatus can be ignored, and can’t be held up as a major fault of his versification. The extra syllable in ke me| ðika|zun iso|via, on the other hand, is a pretty basic blunder.

And in μεγάλονε /meɣalone/ “big”, Markos has added an -e to the nu movable, the liaison consonant. The vernacular allows this for pronouns, as the song shows: αν την-ε ξεμπουκάρω “if I unclog her” = “If I get out of jail”. But for nominals like μεγάλονε, it’s a sign the versifier has given up, and is begging for an extra syllable. It’s an old poetic license—instances turn up in Early Modern Greek; but they’re limited to verse. By my aesthetic, that makes it bad verse.

The thing is, if you discount caesuras, hiatus is limited to learnèd words. Admittedly, that’s a circular argument, since hiatus is characteristic of learnèd words anyway. But notice that we have several instances of learnèd words starting with a vowel—etia, efesi, isovia, isovitis—which trigger hiatus with the preceding word, rather than synezesis. The vernacular words don’t do that; in fact, we have a synezesis of three syllables mooshed into one in the vernacular pla|nepse i e|morfja| su [planepsj emorfja su].

Hiatus before a vernacular word happens only with a caesura, which doesn’t count. Hiatus is allowed before a learnèd word, caesura or not: the force field of Puristic can block the word from cosying up to its predecessor—kano| efe|si, ɣino| iso|vitis—just as it blocks hiatus within the word. But such hiatus it is not compulsory: esi se eti|a has hiatus within learnèd /etia/, but not in front of it. meɣa|li ekði|kisi also skips hiatus separating it from the preceding word.

The reason why I picked this song is its startling last verse. In that last verse, the Homeric names don’t trigger hiatus across word boundaries, just like ekðikisi didn’t. Or maybe Vamvakaris did sing a hiatus after all. Then again, maybe that wasn’t hiatus that Vamvakaris was singing.

It’s open to some question what Vamvakaris is actually doing phonetically with the last verse. But I’ve run enough over length in this post, to defer discussion till the next.