For instance, the verb ἐλευθερῶ augments to ἠλευθέρουν in the past, whereas the verb ἔχω augments to εἶχον (not ἦχον as might have been predicted).
1 Answer
You'll basically have to memorise them, yes, though there are patterns. Both the η- in ἠλευθέρουν and the ει- in εἶχον represent a contraction of ε + ε, but the former is much older than the latter.
The ε + ε > η contraction dates to a time (prehistoric, as far as Attic is concerned) when η was still meaningfully the long version of ε, which is also when most verb forms were established and therefore the most common outcome of an augmented verb starting with ε-.
At some point before the Classical period (in Attic), η was lowered to [ɛː] and ε was raised to [e] and that connection was severed, with a new long vowel ει [eː] now serving as the long version of ε (a similar thing happened for ο/ω/ου). Any contraction of ε + ε that happened after this happened will yield ει instead of η. You'll probably see this in verbs coined late (no examples come to mind just now), but it's also very common in verbs that used to start with a digamma (ϝ [w]) or a σ followed by a vowel, both of which also disappeared in Attic relatively late:
- ἕπομαι 'follow' → εἱπόμην (from PIE *sekʷ-)
- ἔχω 'have' → εἶχον (from PIE *seǵʰ-)
- ἑστιάω 'host' → εἱστίων (earlier ϝεστια-)
- ἐθίζω 'grow accustomed' → εἴθιζον (earlier ϝετιδ-)
- cf. also ὠθέω 'push' → ἐώθουν rather than ˣὤθουν (earlier ϝωθ-)
A disappearing initial ϝ or (especially) σ will often leave a rough breathing, but not, as you can see, all the time. Still, if you do see a verb start with rough breathing, that's a good clue.
Sometimes you'll also see these "new" contractions show up by analogy in verbs where they shouldn't:
- ἐάω 'permit' → εἴων for no etymological reason
Or conversely, the old contractions in verbs where you'd expect new ones:
- ἐργάζομαι 'work' → ἠργαζόμην (earlier ϝεργαδ-); expected εἰργαζόμην also attested
There are other irregularities involving the augment as well. Some verbs get a double one:
- ἁλίσκομαι 'be seized' → aor. ἑᾱ́λων (next to expected impf. ἡλισκόμην)
- ὁράω 'see' → ἑώρων
- βούλομαι 'want' → ἠβουλόμην (probably analogy with (ἐ)θέλω → ἤθελον)
- μέλλω 'be about to' → ἤμελλον
And, because it's Attic, sometimes you get quantitative metathesis:
- ἑορτάζω 'celebrate' → ἑώρταζον (< *ἡόρταζον/εἱόρταζον)
So the short of it is that yes, you'll have to study your principal parts.
-
1Excellent answer. A couple of notes: ἐάω seems to have no clear etymology, so its ει may be etymological for all we know; I don't think there are late-coined verbs that augment as ει (the analogy of η was probably stronger), but I don't know for sure; and it's worth pointing out for learners that even knowing your principal parts won't necessarily get you these forms, since the imperfect usually isn't listed as a principal part. (BTW "η was lowered to [ɛː] and ε was raised to [e]" -- surely just one or the other?)– TKRCommented Apr 28, 2021 at 21:57
-
@TKR I've always been told PIE *e was probably [ɛ] and *ē [eː], like in Latin and Dutch/German, but in retrospect that's probably mostly because my university is Dutch-speaking. Commented Apr 28, 2021 at 22:31
-
@Cairnarvon reading anything at all into the phonetics of PIE is tricky. It's not even 100% clear that e was a front vowel rather than central (although in most languages it seems to have become a front vowel pretty early), so minor distinctions of height are pretty much unreconstructible– TristanCommented Apr 29, 2021 at 12:38
-
Please excuse my ignorance, I don't know the first thing about Ancient Greek, I just happen to speak modern Greek. Was
η
really the long version ofε
and notι
?– terdonCommented Apr 29, 2021 at 12:46 -
@terdon Definitely; the long version of ι /i/ was, well, ῑ /iː/ (not traditionally marked in writing). A total of eleven distinct Classical vowels and diphthongs, including η, ended up collapsing into /i/, which is one reason why modern Greek has so few different vowels now, but that's a process that wouldn't start until some time after the Classical period. Commented Apr 29, 2021 at 13:25