Greek -eza ethnonyms

By: | Post date: 2019-05-09 | Comments: 6 Comments
Posted in categories: Linguistics, Modern Greek

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.


  • sarhanis says:

    In my case, also possibly exacerbated by my mum being from Cephallonia and her Greek presumably more influenced by Italian.

  • Sarhanis says:

    Very excellent!

    I had always wondered why I naturally want to say αγγλέζος and αγγλέζικα with the influence of my parents growing up in Australia, yet after a few weird looks from visits to Greece have had to mentally correct myself to say άγγλος and αγγλικά.

  • crilk says:

    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.

    Anywhere I can read more about this?

  • Evangelos Lolos says:

    Very informative Nick, I’m glad you followed up your Quora post with this more detailed analysis.

    A few comments from me:
    – Iaponas isn’t that archaic, is it? People could pause with ɣiaponezos because of its negative stereotypes from the past (e.g cheap electronics, odd tourists etc).
    – Araviða has two problems: a) it was used by Antzela Dimitriou and as you probably know she is considered to be wrong even when she’s right (“Φυσικά και κάνω λάθη. Κανείς δεν είναι άσφαλτος…”) and b) the word means a short, lightweight rifle and many men may have encountered this term during their military service.
    – I’m weird, so I would vote for Aravi’ni, which rhymes with the female form for Saracen. Also, is it just me, or the terms for Saudi Arabian women feel less odd that just for Arabians?

    • The problem with the negative stereotypes of ɣiaponezos is that any ethnic term can pick up those stereotypes (although yes, -ezos is usually not given the benefit of the doubt). Turkala “Turkish woman” sounds wrong and negative (not least because its feminine suffix -ala is a one-off, so it’s as marked as it gets); but it really is the only vernacular term for “Turkish woman” there is, and no learnèd term has displaced it.

      Poor, poor Angela Dimitriou. If a rifle is called an “Arab woman” in the military (with the learnèd form, no less), wouldn’t that also mean that kanaðeza would be avoided as well, because it refers to an army truck?

      sauðaraviða will indeed sound less odd than just araviða, because we expect learnèd coinages in recent country names like Saudi Arabia, more than we expect them in old broader terms like Arab.

Leave a Reply

  • Subscribe to Blog via Email

  • June 2024
    M T W T F S S
%d bloggers like this: