The wrong vowel in a Modern Greek compound

This post is about an obscene compound of Modern Greek, made up on slang.gr, and how it clearly violates a rule of compounding, by including what looks like a piece of inflection in the first half of the compound. The follow-up post is on how verb–verb and verb–noun compounds work in Modern Greek, and why the coiner of the compound had no choice but to break the rules.

Again, it's not that I'm making a point of posting about obscenities; but obscenities do tend to have interesting linguistic stuff go on with them.

The compound is: γαμαοδέρνουλας.

If you know Modern Greek, have a stare at it. It's not quite right, is it? If you know Ancient Greek, pretend instead it's βινεδέραξ. Which isn't quite right either.

Story beneath the fold.

So, the background to the story is, Language Hat was kind enough to post on my repost of the slang.gr taxonomy of linguistic attitudes. The apotheosis of that taxonomy was "slang.gr FTW". The original had something similar to FTW, Το slang.gr γαμεί, "slang.gr fucks!" Language Hat was amused by this semantic extension of γαμώ.

As I followed up in comments there, this use of γαμεί is a truncation of the old expression γαμά και δέρνει, "he fucks and bashes". (I'll come to the alternation between γαμεί and γαμά in a bit). The sense of the expression I knew was "he is in command of a situation"; the allusion is to the use of penetrative sex and violence as the privilege of the powerful. There is regrettably no shortage of instances where this has happened; the explication I've seen (in Zahos' slang dictionary Το Λεξικό της Πιάτσας) was to tradesmen, and how they used to enforce obedience on their apprentices.

slang.gr does not list this meaning under its entry for γαμά και δέρνει; it gives instead the related meaning "(s)he exercises great skill". If the derivation from masters and apprentices is correct, this extension makes senses; but it is plausible even without it. The original sense is however recoverable in the noun someone has derived from it there, γαμαοδέρνουλας, a sarcastic term for a he-man, "Lord Master Fuck-And-Bash". γαμαοδέρνουλας is derived from the verb γαμάω (Ancient γαμέω), the verb δέρνω (Ancient δέρω), and the augmentative suffix -ουλας, which echoes μπαμπούλας "bogeyman".

Modern Greek can form compounds starting with a verb, and even compounds of two verbs; that's an oddity in Indo-European which will get a separate post. But such compounds still follow the rules of Greek morphology: stem₁ + /o/ + stem₂ + inflection. Having the inflection on the end is normal in compounding: the stems are combined to form a new stem, and that new stem is what gets declined and conjugated as a unit.

• άνθρωπ-ος "human", μορφ-ή "form", ανθρωπ-ό-μορφ-ος "human-shaped".
• έ-χασ-α "I lost", δίκ-η "trial", χασ-ο-δίκ-ης "lose-trial, incompetent lawyer".

The linking -o- is only dropped in Modern Greek when the second term starts with /a/, following the Modern hierarchy for vowel deletion:

• παλι-ός "old", άνθρωπ-ος "human", παλι-άνθρωπ-ος "ol' human = scoundrel"
• έ-σπασ-α "I broke", αρχίδι "testicle", σπασ-αρχίδ-ης "ball-buster"

Now, Ancient Greek verb roots could end in vowels: /aɡapa-/ "love", /ɡame-/ "marry". The inflections following those roots in the present tense typically started with vowels (the thematic vowel): /-ɔː -eis -ei/.

That meant that a vowel went next to another vowel in those proto-Greek verbs: /aɡapa-ɔː/, /ɡame-eis/. But most vowel–vowel combinations in proto-Greek ended up coalescing into a single contracted vowel. You can tell it's contracted, because contracted vowels got circumflexes· /aɡapɔ̂ː/, /ɡameîs/ (ἀγαπῶ, γαμεῖς). So the proto-Greek (and Homeric) ἀγαπάω "I love" (column 1) was contracted by Attic into ἀγαπῶ. The way contraction works, this meant that the thematic vowel, which alternates between /e/ and /o/, turned into an alternation of /aː/ and /ɔː/.

Proto-Greek	Attic	Modern Pronunciation
aɡapá-ɔː	aɡap-ɔ̂ː	aɣap-ˈo
aɡapá-eis	aɡap-âːis	aɣap-ˈa
aɡapá-ei	aɡap-âːi	aɣap-ˈa
aɡapá-omen	aɡap-ɔ̂ːmen	aɣap-ˈomen
aɡapá-ete	aɡap-âːite	aɣap-ˈate
aɡapá-onsi	aɡap-ɔ̂ːsi	aɣap-ˈosi

In fact contraction meant that there are few verbs in Ancient Greek whose stems ended in a vowel intact next to the thematic vowel. Stems ending in /a/, /e/ and /o/ all contracted, and stems ending in /ɛː/ and /ɔː/ were absent from Attic. That left /i/ and /y/, and the Modern Greek vernacular has got rid of many of even those exceptions by analogy: λύω "loosen", the standard example verb, is now λύνω.

Koine gradually merged -αω verbs and -εω verbs. In fact that happened with γαμέω "marry", which in its modern guise "fuck" can be conjugated identically to ἀγαπάω. The merger did not happen in all variants of Greek with the same verbs, which is why it is possible for γαμεί and γαμά to coexist in the Modern language.

The merger generated the Early Modern Greek forms in column 1: the alternation is now between /a/ and /u/. Modern Greek then got rid of the alternation between /a/ and /u/, giving the paradigm in column 2.

Early Modern Greek	Modern Greek
aɣap-ˈo	aɣap-ˈo
aɣap-ˈas	aɣap-ˈas
aɣap-ˈa	aɣap-ˈa
aɣap-ˈume	aɣap-ˈame
aɣap-ˈate	aɣap-ˈate
aɣap-ˈusi	aɣap-ˈane

The thing to note is that since Attic, the inflection on the present tense is not /-ɔː -eis -ei/ > /-o -is -i/, but /-ɔː -âːis -âːi/ > /-ˈo -ˈas -ˈa/. So the -a- is not part of the stem, it is part of the inflection, and it does not belong on compounds starting with such a verb:

• ɣen-ˈo, -ˈas, -ˈa "I give birth", vol-ˈo, -ˈas, -ˈa "I do repeatedly", γεννοβολώ ɣen-o-vol-ˈo "I keep giving birth"
• ɣam-ˈo, -ˈas, -ˈa "I fuck", traɣuði "song", γαμοτράγουδο ɣam-o-traɣuð-o "fuck song, obscene song sung during carnival"

And so we come to γαμαοδέρνουλας, ɣama-o-ðern-ulas. The compound is formed as if the stem of "fuck" is /ɣama-/, not /ɣam-/. That conflicts with compounds like /ɣam-o-traɣuð-o/. Why was this done?

What made this possible is a reanalysis of /-ˈo, -ˈas, -ˈa/ verbs in some variants of Modern Greek, which in the standard is confined to colloquial usage. Unlike normal (uncontracted) verbs, these verbs are accented on the ultima (the final syllable) in the singular. This is odd, and those variants have repaired the discrepancy by adding the normal endings to /-a-/:

Modern Greek		Colloquial Modern Greek
ˈlin-o	aɣap-ˈo	aɣap-ˈa-o
ˈlin-is	aɣap-ˈas	aɣap-ˈa-is
ˈlin-i	aɣap-ˈa	aɣap-ˈa-i
ˈlin-ume	aɣap-ˈame	aɣap-ˈame
ˈlin-ete	aɣap-ˈate	aɣap-ˈate
ˈlin-une	aɣap-ˈane	aɣap-ˈane

This has been called at least once a linguistic atavism: Colloquial Modern Greek has accidentally rediscovered the Homeric forms /aɡapá-ɔː, aɡapá-eis, aɡapá-ei/ (whose modern pronunciation would indeed be /aɣap-ˈa-o, aɣap-ˈa-is, aɣap-ˈa-i/). That's not the reason it's happened of course—someone who says /aɡapá-ɔː/ is unlikely to be stereotyped as an Homeric scholar, and the atavism hasn't extended to the plural */aɣap-ˈaume aɣap-ˈaete aɣap-ˈaune/, whose accent is unexceptionally on the penult (second last syllable).

In conjugation tables, and indeed in people's grammatical intuition, /-ao -ais -ai/ are still treated as inflections; but the whole point of the analogical change was to make /-o -is -i/ the inflections, attaching to a stem ending in /-a-/, so the stress would be more normal. Under that analysis, the stem of /ɣamˈao/ is indeed /ɣama-/. And notice that the expression is cited in slang.gr in the first person as γαμάω και δέρνω, not γαμώ και δέρνω: being colloquial, it is supposed to have the colloquial innovative form of the verb.

So whoever coined a noun γαμαοδέρνουλας, based on the phrase γαμάω και δέρνω, could access an analysis of the verb where the verb stem is /ɣama/ not /ɣam/, and combines as /ɣama-o-ðern-ulas/. But that still looks decidedly odd. /ao/ is an infrequent cluster in Standard Modern Greek in general, routinely occurring only in this colloquial inflection, and in verbs historical ending in /aɡɔː/ (e.g. /pʰaɡɔː/ > /fao/ "that I eat"). Stems ending in /a/ are rarer still: in fact, they shouldn't be there at all; so we are surprised to see one in a stem boundary in a compound. And that's why this form just looks wrong.

So why *did* the coiner of γαμαοδέρνουλας use that odd stem variant, instead of the expected γαμοδέρνουλας? Speakers of Modern Greek can probably guess what I'm guessing the reasons are (γάμος, and echoing the phrase explicitly). But to explain that through, I'm going to discuss the strange history of Greek verb–noun and verb–verb compounds, in the next post.