Subscribe to Blog via Email
I have had an updated version of my old Quora post Which formerly Ottoman-occupied peoples understand “s–tir” today? published in Greek on Nikos Sarantakos’ blog, as Η υστεροφημία του σιχτίρ, “The legacy of sixtir”.
I have had an updated version of my old Quora post Does the Greek word for watermelon, karpouzi, come from Ancient Greek? published in Greek on Nikos Sarantakos’ blog, as Από πού βγαίνει το καρπούζι;, “Where does karpouzi come from?”.
Updated post on “Why do English-speaking people not prefer to say natrium, silisium, kalium, and use other Latin names of elements instead?”
I have had an updated version of my old Quora posts Why do English-speaking people not prefer to say natrium, silisium, kalium, and use other Latin names of elements instead? and Where do the distinctive Greek names for chemical elements come from? published in Greek on Nikos Sarantakos’ blog, as Τα ελληνόψυχα χημικά στοιχεία, “Greek-souled chemical elements”.
I have had an updated version of my old Quora post What are some interesting examples of Ancient Greek vernacular? published in Greek on Nikos Sarantakos’ blog, as Ο Κρίμωνας και το παιδί του Βαθυκλή, “Crimon and Bathycles’ son”.
My original post on the clumsy coinages of terms for paraphilias was a bit of careless venting on Quora, and did not bother researching the creation of the words too deeply. It was, as commenters at Nikos Sarantakos’ blog correctly identified, some xavales, “goofing off”. (Sarantakos did say when I sent him the post, “I think you are being a little harsh on John Money.”) And it was not the kind of thing I should be saying with my linguist hat on: with my linguist hat on, autassinophilia was not intrinsically worse than automobile as a hybrid form (though, by invoking the Asāsiyyūn, it was a more bizarre form.) Eproctophilia is, in fact, intrinsically worse; but it’s not the first dumb coinage that has become a real word in language, and it won’t be the last.
When I updated the post for Sarantakos (which he titled “Fifty shades of paraphilia”), I thought I should put a somewhat less goof-off addendum, about what the clumsy coinages were demonstrating:
For centuries, the scientific terminology of Western science had its footing on classical languages, because classical languages were the languages of learning. In the 20th century, that footing dropped away: classical languages are not particularly known outside of classical philology (and often enough are barely known even there.) Consequently, newer Western scientific terminology does not pay particular attention to Latin and Greek. The terminology of IT in particular is playful, simple, and Anglo-Saxon; and that results in clumsy solutions when we make those terms learnèd again in Greek. Bubblesort is much more spontaneous than ταξινόμηση φυσαλίδας “classification by gas globule”.
Not knowing Latin and Greek is a sign of the times. It’s the poor command of Latin and Greek that shows how marginal classical languages have become in scientific terminology. There are plenty of experts who want to come up with impressive Latin and Greek terms, because impressive Latin and Greek terms are still an institution in the sciences; but they don’t do it right, they come up with broken forms, and noone bothers to correct them. (At least, noone within the English-speaking world.)
So speciesism was coined in 1973 by a psychologist who was not particularly interested in Latin morphology, and so treated species as an uninflected form. As I analysed in ουγκανιά and αντισπισισμός, at least Romance languages corrected the word when they borrowed it (espécisme, specismo, especismo ). (Greek also changed it to spisismos, when it doesn’t nativise it as iðismos; but I believe that was just haplology.)
The case of coulrophobia, the phobia of clowns, is much sillier. Noone knows where the first person to launch it online in 1995 came up with it from, and after it was launched, noone took the time to check it. It might be a distortion of Modern Greek klooun < clown. It might have something to do with kōlobathron “stilts”, as is often claimed, because someone dug up that word, and thought that the closest Ancient Greek ever got to clowns was stilt-walkers. (That corrected etymology has made it back to Greece: Τι είναι η κωλοβαθριστοφοβία; Μήπως πάσχεις και εσύ και δεν το ξέρεις; “What is kolovathristofovia? Might you be suffering from it without even knowing abut it?”) But the opinion of etymonline rings truest: “Coulrophobia looks suspiciously like the sort of thing idle pseudo-intellectuals invent on the internet and which every smarty-pants takes up thereafter.”
Of course, as that article also points out, the word is not an official term in psychology, though the phenomenon is acknowledged there; and if it were an official term, the requisite plausibility check might have happened…
… Maybe, but Money’s acrobatics, and the unbelievable sloppiness of Aggrawal’s catalogue of paraphilias (in an otherwise serious monograph from the reputable publisher Taylor & Francis) shows that it probably wouldn’t have. Sloppiness, misinterpretation and solecisms in classicising terminology now go unnoticed, because classicising terminology is no longer a living process with many participants actively checking it. It is a dry convention, with little attention paid to it. And when we need to borrow its results back into Greek, we end up patching them up as best we can.
As several commenters chimed in (Diver of Sinks first),
Bad news for those of us who boast about the words we have given the Dumb Franks: we’ve already offered anything there was to offer (ό,τι δώσαμε δώσαμε).
There was a bit of discussion on that between me and commenter Neo Kid: yes, coinage based on Graeco-Latin roots is still going on. But it no longer particularly cares how true to Greek and Latin the results are; much less so than in the past. The Graeco-Latinity of the coinages is a much more empty gesture now. And commenter Nikiplos made the justified (if painful) observation that the Asian and African future pioneers of science are going to have a lot less patience for Graeco-Latin coinages than their European colleagues.
I was also justly asked by Neo Kid whether my expression of revulsion was a scholarly opinion or merely a matter of personal taste. It could only, of course, be a matter of personal taste, because as a linguist, all I get to say is, “shit happens”. That’s why I’d linked to Professor Higgins in the original post: Professor Higgins might well have said that Liza Doolittle should have been “taken out and hung for the cold-blooded murder of the English tongue”; but no linguist in the past century would issue those kinds of aesthetic (and classist) judgements ex cathedra.
Though that’s not the kind of lack of privilege in evidence with these coinages. And I’m not going to forgive eproctophilia soon. If you’re going to go to the bother of making up a Graeco-Latin word for “fart fetish”, at least take the time to look up an actual Greek or Latin word for “fart”, instead of coming up with “out-of.(Latin)-anus.(Greek)-philia”.
The crowd at Sarantakos’, I have to say, is tougher than on this obscure blog, or even on Quora. (And I’ve been blocking commenters much more on my return to it!) Commenter Ioanna was critical of the unserious part of my post, and the criticism is not unwarranted. I reproduce it here.
Dr Nick Nicholas, who is acknowledged to be an excellent linguist, should not have rushed to conclusions about the competence in Greek of a scholar, whose brilliant career he is completely unfamiliar with. The late John Money was one of the most preemiment American sexologists, professor for years at John Hopkins and Baltimore Hospital. Professor Money had Greek learning, had boundless admiration for the Greek language, and was in no way only familiar with it via dictionaries, as Dr Nicholas maliciously implies.
I would have expected the esteemed Dr Nicholas to have at least leafed through some of Prof Money’s books before writing this article. It is embarrassing for a linguist with Dr Nicholas’ reputation to have used Wikipedia to look up the meanings of kinds of paraphilias, and ignored John Money’s masterpiece Gay, Straight, and In-Between: The Sexology of Erotic Orientation (1988), where all the words are explained in detail. If he can’t afford to buy the book, he can say so and I’ll mail it to him.
For instance, this is what Money writes about erotophonophilia, which so offended the linguistic sensitivities of Dr Nicholas, clearly acting as a snob:
erotophonophilia: a paraphilia of the sacrificial/expiatory type in which sexuoerotic arousal and facilitation or attainment of orgasm are responsive to and contingent on stage-managing and carrying out the murder of an unsuspecting sexual partner (from Greek, eros, love + phonein, to murder + -philia). The erotophonophile’s orgasm coincides with the expiration of the partner. The reciprocal paraphilic condition is autassassinophilia. Syn. lust murder.
Is that anything to do with Wikipedia’s definition?
I’ll take my lumps. Erotophonophilia is clearly defined by Money as a murder fetish and not a death fetish, the way Wikipedia said it was. And using Wikipedia was sloppy; yes, it was deliberately sloppy, because I was writing a flippant post, but it was impugning a scholar in doing so. (Although I still don’t think autassassinophilia, let alone olfactophilia, are evidence of boundless admiration.) I also can’t confirm that eproctophilia was Money’s doing (although olfactophilia is on his 1988 list). But if he didn’t want to be judged by classicist snobs, he shouldn’t have used olfactophilia instead of smell fetish to begin with. The snobbishness is inherent in Graeco-Latin coinages.
Money’s reputation won’t suffer that much by me grousing about autassassinophilia, even if it was a cheap shot. There are far more serious controversies to his research anyway.
I have had an updated version of my old Quora post What is it called when you get aroused by watching people die? published in Greek on Nikos Sarantakos’ blog, as Οι πενήντα αποχρώσεις της παραφιλίας, “Fifty shades of paraphilia”.
I’m sure someone somewhere has already written about this. In fact, I’m sure multiple people have. But I enjoy reinventing the wheel.
(And getting inspiration from Quora, as well as Nikos Sarantakos’ blog, for articles here. My thanks to Evangelos Lolos for prodding me on this.)
Modern Greek used to have a lot of ethnonyms ending in -ezos, -eza, a suffix taken from Italian -ese, -esa. Those suffixes originally came as part of ethnonyms that were borrowed wholesale from Italian: englezos, ɣenovezos, frantsezos, olanðezos, kinezos “English, Genoese, French, Dutch, Chinese”.
With the advent of Puristic, many of those ethnonyms were replaced with Classical terms (anɡlos, ɣalos “Angle, Gaul”), and others just had the -ezos suffix stripped (olanðos).
The current status of the -ezos is therefore a messy mix:
- There are some ethnonyms for which the -ezos suffix is unavailable: frantsezos is long forgotten, and no form like ɣalezos has ever developed.
- There are some ethnonyms for which the -ezos suffix is the only available form: maltezos, kinezos, filipinezos “Maltese, Chinese, Filipino”. (Puristic did come up with malteos, sinas, but those variants did not stick for whatever reason)
- For a large number of Western ethnonyms, the -ezos suffix was retained, as a colloquial variant. So the normal forms are anglos suiðos olanðos, norviɣos, polonos, kanaðos “English, Swedish, Dutch, Norwegian, Polish, Canadian”, but the variants enɡlezos, suiðezos, olanðezos, norviɣezos, polonezos, kanaðezos are possible. They are however quite marked now: they sound folksy, potentially derogatory, and they are seldom used. In fact, the suffix is so marked now, that people even pause on instances like maltezos, where it is the only form in widespread use, or ɣiaponezos “Japanese”, where the alternative form i.aponas should be unacceptably archaic.
The oddity is with the female counterparts to those -ezos forms. While the masculines in -ezos are now marginal, the feminines in -eza are still in widespread use. So engleza, suiðeza, olanðeza, norviɣeza, poloneza, kanaðeza are nowhere near as restricted in usage as their masculine counterparts, and can turn up productively in unexpected quarters. (kanaðeza “Canadian chick”, for example, is the term for a small truck in military slang, because the first such vehicles in the Greek army were WWII surplus Canadian Military Pattern trucks. The term persists though they haven’t been used for decades.)
In fact, the -eza is being applied innovatively to ethnonyms where it has no historical warrant. “Albanian” for example was traditionally arvanitis, arvanitisa (now restricted to the Christian ethnic Albanians living in Greece), replaced by the learnèd alvanos, alvani. There has never been a form alvanezos, but the feminine alvaneza has emerged recently.
So a suffix that now sounds vaguely pejorative in the masculine flourishes in the feminine. Is the explanation “because Greeks are sexist”? Greeks are sexists, sure. But there’s a little bit more going on linguistically.
First, the new learnèd forms were learned by people by exposure, as much as by paradigm. The learnèd masculine forms featured in the press, because the press talked about newsmaking foreign men. The learnèd feminine forms did not feature anywhere near as often, because the press didn’t talk anywhere near as often about newsmaking foreign women. That’s sexism, but it’s the sexism of the entire world blocking women from making news, as well as blocking the papers from reporting it if they did. So there was less opportunity for the learnèd forms to displace the vernacular -eza forms, than there was for their male counterparts.
Second, the new learnèd masculines used the bog-standard inflectional ending -os, with zero-derivation; they posed no particular morphological challenge in Modern Greek. But some new learnèd feminines used the Ancient Greek ending -is, which is alien to Demotic; Demotic nativised it as -iða, and it does turn up in some vernacular forms (nɛːrɛːis “Nereid” > neraiða “fairy”), but there was some discomfiture with it, and it isnʼt always taken up. Generations of Greeks now have had no problem nativising platɔːn “Plato” to platonas and tʰɛːseus “Theseus” to θiseas, but feminine proper names like artemis, alkistis “Artemis, Alcestis” have resisted nativisation to the regular artemiða, alkistiða—to the extent that speakers would rather keep them indeclinable.
When there was no alternative, because the original vernacular form was banned, people shrugged and used the -iða form: eliniða “Greek woman” was obviously going to be hard to resist, and ɣaliða, ɣermaniða, rosiða, ispaniða “French woman, German woman, Russian woman, Spanish woman” were the only forms available after the elimination of the vernacular, Italian frantseza, alemana, rusa, spaniola. But if the old -ezos form survived at all, people seem to have jumped on it, as something more comfortable than the -iða form: thus poloneza for poloniða, enɡleza for anɡliða, suiðeza for suiðiða.
The zero-derivational feminines that also turned up as learnèd forms, like irlanði as the feminine counterpart of irlanðos “Irish”, clearly sounded odd to people, that they weren’t comfortable with; presumably, as marked feminine forms, they were too close for comfort. But I’m not sure it was just the zero-derivation that was the issue; there are plenty of zero-derivational feminines in the vernacular: eliniða replaced romja “Roman woman”, “Bulgarian woman” is vulɣara.
Nikos Sarantakos’ article on the issue (which I stumbled on after almost finishing this post) points out a more plausible reason why people hesitate over formal feminines in general.
Feminine ethnonyms are often difficult, and most of the time we find them in two or three forms, precisely because they have not been settled by usage [NN: my first point above on exposure], whereas one form dominates in the masculine. For example, we have alvanos “male Albanian”, but for a female Albanian we have alvani, alvaniða, alvaneza; the same goes for a female Ukranian [ukrani, ukraniða, ukraneza]. The -eza suffix is considered sometimes overly colloquial, and sometimes dismissive; but how else can you call an Irish woman or Dutch woman? irlanði of course in formal language, but also irlanðeza, just as you can’t do away with olanðeza, despite the existence of olanði; the same goes with suiðeza.
Feminines in general in Greek present an embarrassment of riches, and when novel feminines emerge, like “female doctor” or “female member of parliament”, there is considerable uncertainty as to what to use: the vernacular feminine endings often sound dismissive, because of past sexism, and there’s lots of them; the formal endings often sound awkward, because they’re learnèd (or because they use seemingly masculine endings, which can make sense for -os in Ancient Greek, but not in the vernacular).
In the instance of feminine ethnonyms, the formal language offers you a choice of zero-derivational -i, which may be a bit too adjectivey (because morphological markedness, too, is sexist); or -iða, which sits poorly in the vernacular.
That confusion can lead to mere avoidance. The article in fact is premised on the peculiar fact that Greeks aren’t sure what the feminine form of aravas “Arab” should be, at all; and the feminine of “Arab” is plagued by even more uncertainty:
- The original feminine is arapina, feminine of arapis. But that vernacular form no longer means “Arab”; because Greeks conflated the Sudanese Ottomans slaves they encountered with Arabs, its meaning has changed to “black person”: unmarked two centuries ago (“moor”), pejorative now. And arapina only survives in an orientalising song by Tsitsanis from 1946, that would have given Edward Said a heart attack. (Αραπίνες λάγνες ερωτιάρες, “Lustful erotic Arab women”. It really does sound so much worse when you translate it, and realise what he’s actually saying, as opposed to just singing along to the nice tune.) So aravas has been revived as the Demoticisation of the classical form araps, gen. arabos. But that revival did not bring a suitable feminine along with it.
- aravina is the analogy from arapina; but the -ina suffix is too colloquial for a learnèd stem
- aravisa is the most plausible vernacular form, but there isn’t enough precedent for -isa in learnèd ethnonyms either (whereas it is extremely common in demonyms within Greece)
- araviða is the expected learnèd form. But because noone actually got around to formally proposing it, it gets greeted by peals of laughter. As Sarantakos’ commenter Pepe found in high school.
- The situation is so bad that the first edition of Babiniotis’ dictionary, normally a den of conservatism, actually capitulates and gives ɣineka aravas “Arab woman”. Sounds fine in English. Is utterly humiliating in Greek, since aravas is masculine. (The second edition allows aravisa.)
A similar confusion on where to accent the genitive plural for feminine nouns has led Greeks to just avoid them completely, as my friend Dion Mertyris found in his thesis. For that matter, the controversy around how to interpret Esperanto tense in compound verbs led Esperantists to simply abandon the passive, which can only be expressed as a compound tense.
Faced with that conundrum of learnèd forms, Greeks are likely to run away too. At least they have somewhere to run away to: -eza has the advantage of familiarity and productivity, even if it is visibly informal. (And yes, sexism helps people get over the inconsistency between masculines and feminines; but that isn’t the only factor involved.) So people are happy to lean on it. That familiarity has in fact helped it expand to instances like alvaneza, ukraneza, where the corresponding masculines alvanezos, ukranezos have never existed.
Though noone to date has come up with araveza. That -eza suffix, being Italian in origin, is associated with the West too strongly, even if it also gets applied to the Far East (Chinese, Japanese); it can be applied to the Near West (alvaneza), but there seems to be a block to applying it to the Near East.
Updated post on “If you were allowed to add a symbol to unicode, what symbol would it be, and what would it mean?”
I have had an updated version of my old Quora post If you were allowed to add a symbol to unicode, what symbol would it be, and what would it mean? published in Greek on Nikos Sarantakos’ blog, as Πώς χαντάκωσα τα Παμφυλιακά, “How I ruined Pamphylian”—referring to how I’m responsible for the psi-like Pamphylian sampi not getting a distinct codeset from the regular alphabetic sampi.
Many posters were non-plussed by the highly technical discussion of Ancient Greek orthography and Unicode; a few made it through to the end. My especial thanks to commenter Efi-Efi, for chiming that she too lamented (θrino) the fate of the pitchfork (in Cretan dialect: θrinaki).
I’ve finally taken the time to read Günther Henrich’s 1976 thesis on the spread of the -o vocative and -o genitive in Greek.
My blog series has been something like 15 pp written off the cuff, with minimal research. Henrich’s is 270 pp of meticulous historical and dialectal research. He has orders of magnitude more detail than I provided, and his tracing of developments is much more fine-grained than I could have done. (Of course. After all, I spent a couple of evenings on this, and he spent several years.)
It’s an impressive piece of work, all the more so because it was done in 1976; and I’m always impressed with the kind of work people used to do with index cards instead of computer corpora. (My own PhD was at the tail end of that.) I haven’t been well disposed the recent Bible Code-like attempts by Henrich to read cryptogram signatures into the works of the Cretan Renaissance; but the quality of research in this work has more than made up for that.
And I’ll allow myself the gratification that he hasn’t contradicted me. The developments I posited, like the neuter for Manolio(s) “Manuel”, were confirmed. The claims of analogical factors that I’ve made are all there in his thesis: formality, penult stress, proper names, all given with quantitative justification. (Of those original third declension nouns like ɡerɔːn that have switched to the second declension, and which are the core of the o-vocative, 3/4 are penult-stressed, and 2/3 are proper names.)
He also a quirky-looking derivation for the o-vocative (and its related o-genitive), which seemed implausible to me at first, but which has grown on me:
|Early Mediaeval (ca. 1000): Indeclinable||ɣeron||ɣeron||ɣeron||ɣeron|
|Later Mediaeval; Cretan||ɣeros||ɣero(n)||ɣero(n)||ɣero(n)|
In other words, the old third declension in -ɔːn did not switch wholesale to the second declension (-os, -u, -o, -e); it moved one case at a time, and the vocative (subject to less analogical pressure than the genitive, as a more rare case) is the contemporary battleground. That means that Cretan, with its o-genitive corresponding to the o-vocative, is archaic rather than innovative, which is consistent with the o-genitive being more common in Mediaeval Greek. (That also makes more sense because it’s harder to think of the vocative or accusative influencing the genitive, which is what would be required if the o-genitive were innovative.)
Because Henrich had a lot more old data to work from than I did, and because he’s taken more of an old school historical linguistics approach, he spends a lot more time on how the phenomenon got started than how it spread. Not only does he analyse how the original nom. ɡerɔːn, gen. ɡerontos, voc. ɡeron gave rise to the o-vocative; he also introduces a lot of pathways for the o-vocative to have been continually imported into the language, mainly through borrowings (proper names from every Balkan people there ever was, and both proper names and common nouns from Venetian/Italian).
It is a subtle game to establish the relative importance of the factors that bring about language change: the initial reanalysis or borrowing, that brings the change into being, versus the subsequent analogy that generalises it through the language. My own opinion is that Henrich trusts etymology too much: I’m not convinced all those Albanian and Bulgarian names were imported complete with an o-vocative from the very beginning, and without the powerful analogy of proper names already in place, I don’t think the loanwords would have lined up so promptly.
And not all of Henrich’s etymological accounts have the same explanatory power. For example, he thinks that the o-vocative is particularly popular after velars, because it avoids the switch between palatal and velar, which makes the morphology less transparent (that’s how he explains [kakurɣos ~ kakurʝe] “criminal”, [likos ~ lice] “wolf”); but those allophonic switches happen in verbs and nouns constantly, and native speakers are not aware of them at all.
He also posits -ios > -os as a class of nouns that takes the o-vocative: that accounts for eɣiptios > ɣiftos “gypsy”, ɣeorɣios > ɣiorɣos “George”, and several diminutives like -akos > -akios. But that phonological simplification does not particularly explain why the vocative -ie would switch into -o. In fact the laundry lists of reanalysed and borrowed names that take the o-vocative (names formed from verbs! names formed from adverbs! Albanian diminutives! Slavic diminutives!) have a simpler underlying explanation—which he does mention in passing: novel name forms in general will tend to take the analogically spreading form, since the conservative force of the inherited e-vocative is not present there to push it back. All because the o-vocative is now the unmarked strategy for penult proper names (as he himself concludes, already in 1976: “all penult-stressed proper nouns ending in –os can potentially take an o-vocative.”)
And where the o-vocative pushes into previously unavailable terrain (like common nouns), I think semantic similarly counts for a lot more than phonology in how the analogical spread works. The vocative kakurɣo after all shows up next to the vocative ðolofono “murderer”, which has no velar—but a lot semantically in common with the main contemporary class of common nouns taking the o-vocative (murɣos “uncouth”, bufos “fool”, rufianos “informant”).
On the other hand, I probably underestimated the contribution of loanwords: there certainly appear to be a disproportionate number of Italian words among the common nouns that do use the o-vocative, and the dialects where the o-vocative has prospered most (including extending to antepenults) have been the dialects with the longest Italian contact, Cretan and Heptanesian. The fact that Nicolò was indeclinable in Cretan Greek notary documents up until 1500 (appearing as ŋikolo even in the nominative) shows that the o-vocative was certainly going to be reinforced by the influx of Italian names. I had guessed as much, but I hadn’t realised how much that extended to common nouns as well (like say filiotsos < figliozzo “godchild”). And many of the contemporary common nouns that take an o-vocative can be traced to Italian (kamarotos, kouniaðos, bufos, rufianos, tsarlatanos).
The other overall thing I noticed about our different accounts was that too much detail gets in the way of the big picture. That’s not necessarily a fault in an historical account, of course; but analogy does not work by appeal to etymological detail; it works by squinting the eyes, and blurring all the detail into large, arbitrary generalisations. Sort of like I did with my account. Which, after all, was seeking to explain not what had happened in the 11th or 16th century, but what generalisations explain what has been taking place over the last century.
It’s interesting to see both what hasn’t changed and what has. Renaissance Cretan had anticipated a lot of the developments taking place now in Standard Modern Greek (since the o-vocative was more advanced there); the modern military slang neo “newbie”, for example, turns up in a tragedy around 1600, and Phoebus (which Cretan dramatists had only heard of via Italian) gets an o-vocative, like any Italian name: Φοίμπο fibo. (That’s not the same thing as o-vocatives now spreading to learnèd names; but that’s how it gets started. In fact, a much more modern play, Petros Markaris’ translation of Brecht’s Life of Galileo, makes a point of having scholars in the play use the vocative ɣalilee, while commoners use the vocative ɣalileo.)
On the other hand, Henrich still found very little evidence of o-vocatives spreading to antepenult names, let alone common nouns; that’s something that has changed in the last four decades.
It’s also frustrating to realise the examples you don’t come up with if you’re thinking off the top of your head, instead of doing research. I wouldn’t have made the connection between my name, Nikos, and the Mediaeval name Nikōn; it’s too hard coded in my head that Nikos is short for Nicholas, but that doesn’t mean it can’t have initially been the regular development of Nikon. But I should certainly have realised that the Grim Reaper, Charon < Charos, was a canonical example of the shift of -ɔːn to the second declension (and its o-vocative is attested as early as the late 15th century, in Emmanuel Limenites’ Plague of Rhodes. Which, I notice, was recently published in a popular edition by one Günther Henrich.)
Oh, and Philip Newton? Good catch on the e-vocative in γέρε being associated with adjectival use. Henrich argues that the o-vocative does not stick with adjectives, because there is much more analogical pressure to align adjectives with the well-established first and second declensions, since adjectives decline in all genders and numbers. In fact, it’s why third declension adjectives like askʰɛːmɔːn “misshapen” did not keep an o-vocative when they switched to the second declension: nom.masc.sg asximos, voc.masc.sg asxime “ugly”.