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...