The retreat of the e-vocative in Modern Greek: the data

This is a story of analogy, based on an article at Nikos Sarantakos’ blog (because the traffic between our two blogs has ever been two-way). Sarantakos’ article in turn cites an older blog post by Giannis Haris, which cites two grammars of Modern Greek.

The story is the retreat of the vocative in Modern Greek. The vocative is a case that strikes many a learner of Classical languages, because most modern European languages don’t have them. The vocative has survived in Modern Greek…

… but it has been restricted. Even in Ancient Greek, the vocative did not show up in all numbers or declensions: a distinct vocative appeared only in the singular, and it did not appear in feminines of the first declension, in neuters of the second declension, or in most subclasses of the third declensions:

• First declension: masc. nom. polítɛːs “citizen”: voc. polîta; fem. nom., voc. kórɛː “maiden”
• Second declension: masc. nom. ántʰrɔːpos “human”, voc. ántʰrɔːpe; neut. nom., neut. paidíon “child”
• Third declension: nom. ɡérɔːn “old man”, voc, ɡéron; nom. sɔːkrátɛːs “Socrates”, voc. sɔ́ːkrates; nom. voc. pʰýlaks “guard”; nom. voc. poimɛ́ːn “shepherd”

The vocative has retreated substantially in Modern Greek.

• In the first declension, masculines use the accusative/genitive as the vocative: nom. politis acc. gen. voc. politi. Puristic Greek did revive the ancient vocative for honorific titles: nom. kaθiɣitis “professor”, voc. kirie kaθiɣita “Herr Professor”. But you won’t find anyone under 70 now using vocatives like that.
• In the vernacular, the survival of the third declension is vestigial: most forms had switched to the first declension. Even where Puristic revived the third declension, it did not bring back with it the distinctive vocatives with the short final vowels.

In fact, the first declension (which in the vernacular expanded to include all five final vowels, in order to deal with loanwords) only has two distinct cases in the singular: nominative vs the rest in the masculine, genitive vs the rest in the feminine.

The second declension masculines, however, hold out. The vocative of anθropos “human” remains anθrope, with the same inflection as in Ancient Greek.

(That’s the masculines: the feminines in the vernacular had switched to acting like first declension nouns: Corinth was nom. acc. voc. korθo. gen. korθos. Puristic has undone this development, and restored some feminines in the second declension, although speakers occasionally stumble on them.)

But the vocative is starting to retreat in Modern Greek: there are nouns where the vocative in -e is replaced by the accusative ending -o, analogous to what happens in the first declension. It is retreating in specific classes of nouns, and the retreat is spreading from subclass to subclass of those nouns. The data, which I’m going to present in this blog post, looks at first somewhat random. But if you identify the classes of nouns, and the categories people think they belong to, the retreat makes a lot more sense. As does the confusion and vacillation of speakers about which vocative to use; because this is an ongoing change in Greek, and there are plenty of grey areas among those subclasses.

(The categories are established in the grammars that have been cited; but the analysis is not particularly parsimonious, and I’m going to try and make it more general.)

So where is the vocative retreating now in Modern Greek?

The -o vocative is mandatory in:

• Penult-stressed nouns
  • A small number of bisyllabic nouns that used to be third declension: ˈɣeros “old man”, ˈðjakos “deacon”. (Ancient ɡérɔːn, diákɔːn). (As commenter Panos in Lowercase points out, when those nouns are compounded they go back to the old vocative: paˈljoɣere “stupid old man”.)
  • (Etymologically truncated) bisyllabic given names: ˈɣjorɣos, ˈnikos, ˈðimos (corresponding to the formal forms ɣeorɣios, nikolaos, ðimitrios) “George, Nick, Dimitri” (Astonishingly, noone picked up on the truncation being the explanatory factor until commenter and fellow Esperantist Angelos, 119 comments down.)
  • The trisyllabic (truncated) name aˈlekos “Alec” (corresponding to the formal form aleksanðros)
  • Given names and common noun diminutives ending in the diminutive suffix -ˈakos: kirˈjakos, anθroˈpakos “Cyriac, contemptible little man”

• Final-stressed nouns
  • Multiple-syllabic final-stressed diminutive given names: ɣjanaˈkos, ðimiˈtros, manoˈljos (corresponding to neutral ɣjanis, ðimitris, manolis: “John, Dimitri, Manuel”)

The -o vocative is optional in:

• A few familiar penult-stressed nouns: the grammars give kapeˈtanios “captain”, kamaˈrotos “shipmate”
• A few more nouns not mentioned in the grammars, although these seem to be much more contentious. Commenter Panos in Lowercase gives ˈɣiftos “gypsy, blacksmith”. Friend of this blog Pepe adds kumˈbaros “god-sibling” as having a rare o-vocative (which Sarantakos rejected). Pepe also adduces the vocatives ˈfilo “buddy!” in contemporary slang, and ˈneo “young man!” in older military slang. A few commenters mentioned kaˈkurɣo “criminal!” from literature or old movies.
• The case of ˈθio “uncle!” turned out particularly controversial: Pepe adduced it, Sarantakos rejected it as slang; Pepe insists that ˈθio is widespread, and commenter Alexis believes it is a Northern Greek form, avoided in the Peloponnese.
• Penult-stressed given names that are already bisyllabic in their formal forms: ˈpavlos, ˈpetros “Peter, Paul” (ˈpetre is much rarer than ˈpavle). Stavros is also not truncated; commenter Alexis considered the vocative ˈstavre to be jocular, but commenter SP reports it as extant, though more used by men than women.
• Although the grammars do not say so, longer penult-stressed given names may take an o-vocative too: commenter avɣuˈstinos “Augustin” reported a friend hesitating on what vocative to use for him, Panos in Lowercase reported Marios, and friend of this blog Diver of Sinks reported Marinos and Poulikos (but not e.g. Rodolfos.) Commenter Leonicos accents his name both after the original Ladino as leoˈnikos, and as the more hellenised leˈonikos; following this trend, his vocative of leoˈnikos is leoˈniko, and his vocative of leˈonikos is leˈonike.

The greatest confusion is around penult-stressed surnames. Sarantakos acknowledges this in the title of his post: Κύριε Σαραντάκο ή κύριε Σαραντάκε; “Mr Sarantako, or Mr Sarantake?” Triantafyllidis 1941 grammar claims that the surname Dimitrakos has a mandatory -o vocative, like the other instances of the -akos suffix, and he also extends that rule to two other diminutive endings on surnames, –ukos and –itsos.

Georgia Katsouda (who I’ve had the pleasure of meeting recently at the Historical Dictionary of Modern Greek in the Academy of Athens) wrote a grammar in 2007, which makes a more recent distinction: surnames that are etymologically transparent—that is, surnames that look like common nouns—use the -e vocative: Mr kaˈmene “burnt”, Mr ðefteˈree “second”. Surnames that are etymologically opaque use the -o vocative: Mr aleˈvizo, Mr veniˈzelo. Sarantakos confirms that the distinction is recent: the politician Eleftherios Venizelos a century ago was addressed in the press as veniˈzele, while the vocative veniˈzelo is used for the contemporary politician Evangelos Venizelos (no relation, though Sarantakos could not resist the temptation to stick his partisan boot in). He also cites a mid-19th century work by Panagiotis Soutsos, using e-vocatives in surnames that would now be unacceptable (Karatase, Soutse, Diake, Giatrake, Kanele).

But Giannis Haris, who Sarantakos is citing, attests widespread confusion with surnames: as both the difference in the two grammars and in the treatment of Venizelos shows, surnames appear to be where the vocative is currently volatile. There are lots of surnames that Haris heard take -o on the TV to his surprise:

• Giakoumatos; but that could be because of the patronymic suffix -atos (by analogy with -akos)
• “Dimos Verykios”, ˈðimo veˈrikio; I agree with Sarantakos that veˈrikio is likely influenced by the preceding vocative ˈðimo, and where there is confusion between grammatical alternatives, context is going to have a much bigger sway than normal.
• Alimonos: the vocative aˈlimono sounds exactly like the interjection aˈlimono “alas!” (which is of course the origin of the surname: there is no common noun aˈlimonos for the surname to sound like, so the -e vocative would not make sense here)
• Karamanos: maybe because it was explained as Kara-Manos, where Manos is a truncated given name (short for Manuel), of the type that takes an -o suffix.
• Marinos: an -o vocative when it is used as a surname; but always an -e vocative when used as a given name. (We’ve just seen that Diver of Sinks disagrees!)

Haris’ concluding comment is a despairing “Arbitrariness and chaos?” Sarantakos’ conclusion (after dispensing with Evangelos Venizelos) is “this indicates a more general, but very gradual tendency towards o-vocatives.”

There is no unitary set of rules that will explain the distribution of vocatives in Greek, because speakers themselves are clearly confused. But as I indicated, the confusion is explicable, and I will attempt at least some of that explication in the next post.