why does the imperative keep the augment εἰπέ?
Short answer: ει- is not the result of augmentation. The verb stem is ειπ- (cf. epic ἔειπον < ἔϝειπον). So also inf. εἰπεῖν with no augment.
Some more detail: ειπ- is derived from the PIE root *wekw- [kw = labiovelar; don’t know how to do superscripts on Textkit]. The aorist derived from the zero grade with reduplication, wewkw–. For the 2s thematic impv:
*wewkwe (reduplicated zero-grade aorist)
*weykwe (dissimilation of wew–)
ϝειπε (labial reflex of kw before a front vowel by leveling)
εἰπέ
Cf. ἔπος, from the e-grade (< *wekwos)
The stem is (Ϝ)ειπ-. The augmented indicative form ἔ(Ϝ)ειπον is contracted to εἶπον. Non-indicative forms are based on the unaugmented stem εἰπ-. So, εἰπέ, εἰπεῖν, εἴπω, etc.
Cross-posted with phalarkos.
now it is clear, thank you
Could you pease explain more about the grades and the zero grade. I found it in Smythe but it is not very clear to me. I guess this is about vocal change in the radical.
Vowel gradation: this is a topic that calls for a book, far beyond a simple post on Textkit, but I’ll try to sketch a quick and dirty answer. Phalarkos knows much more about this than I, so he may have to correct my answer.
Some primitive Greek verbs show a pattern of an -e- root vowel in the present stem, zero in the 2nd aorist, and an -o- root vowel in the 2nd perfect. A relatively clear example with an -ι- diphthong in the present and perfect and just -ι- in the aorist: λείπω/ἔλιπον/λέλοιπα. These patterns ultimately go back to vowel alternations in Proto-Indo-European (PIE), where they were productive. In many cases, these alternations have been obscured in Greek by various intermediate sound changes that occurred between PIE and Greek.
In the case of εἶπον, a verb that exists in Greek only in the aorist, it’s necessary to reconstruct a presumed PIE form to see how a zero-grade aorist with reduplication in PIE has evolved into Greek through intermediate changes, as phalarkos has shown. (And, incidentally, reduplicated aorists were a feature of some PIE verbs, but very few have survived in Greek, and the reduplication has mostly been obscured by intermediate changes.)
ok got it
Sure, here’s a little more about vocalic grades. [Written before I saw Hylander’s post, which says much the same].
Smyth is outdated on these sorts of things (CGCG is much better). If there is interest from you or others I can provide some bibliographic recommendations for getting into historical linguistics and the early history of Greek. A great place to start is Fortson’s Indo-European Language and Culture: An Introduction.
Consider the relationship between the following forms of πατήρ and ἀπάτωρ (fatherless). Pay attention to the vowel between τ and ρ:
πατέρα ἀπατόρος
πατήρ ἀπάτωρ
πατρί
As you probably noticed, there is variation between short e (πατέρα), short o (ἀπατόρος), long e (πατήρ), long o (ἀπάτωρ), and no vowel (πατρί). These vocalic changes exemplify a fundamental pattern in Greek and its hypothetical, reconstructed ancestor, Proto-Indo-European (PIE). Such vocalic change is called “ablaut” (or “apophony” in older scholarship like Smyth). Ablaut variants are referred to as “grades.” There are five grades: e-grade (aka full grade), o-grade, lengthened e-grade, lengthened o-grade, zero grade. The basic vowel in PIE was e, so that is called the full grade and is the usual lexical form.
There is often ablaut variation in different verb tenses. Consider the following pairs of 1st + 4th principal parts:
πέμπω/πέπομφα, τρέπω/τέτροφα, λείπω/λέλοιπα
In each pair, the 1st principal part is formed from the e-grade. Eg. λείπω, derived from the PIE root *leikw- (the letter kw, a voiceless labiovelar stop, often resulted in π in Greek—see this video for a quick intro to PIE stops: https://www.youtube.com/watch?v=8c5GSmSNuRI&list=PLtZ3bVSEpCtQ-YgD5qTVIn_yssSRVYdZE&index=4; I also recommend the History of English Podcast).
The perfect active indicative (4th principal part) is formed from the o-grade with reduplication (eg λέλοιπα < *leloikw–). Second aorists like ἔλιπον are formed from the zero-grade (< *likw-), i.e. with no e- or o-vowel.
εἶπον is a little more complicated. ἔλιπον is formed directly from the zero-grade of the root (an example of a thematic root aorist). Other aorist tense-stems are formed from the zero-grade but with reduplication (cp ἤγαγον). For εἶπον, the PIE root is *wekw-. The zero-grade is *wkw-. With reduplication: *we-wkw-. In a very ancient phase of Greek (Proto-Greek), the combination *wew- became ϝει (wey-) through a process called dissimilation. Thus wewkw- became weykw-. As with λείπω above, the labiovelar kw- resulted in π in this context. Thus *weykw- > ϝειπ-. Much later (and at different times depending on the dialect), ϝ was lost and we are left with -ειπ, the aorist verb stem seen in εἶπον, εἰπεῖν, εἰπέ, etc.
From the same PIE root *wekw- is derived (ϝ)ἔπος, Latin vox, English voice, epic, evoke, vocabulary, etc.
all this is very interesting, thanx. I guess the Russian говор is clearly related.
I don’t think it’s related, but I know very little about Slavic languages unfortunately. Hylander may know more. I think говор is cognate with Greek γοάω (and γόος).
You’re getting that [the говор derivation] from wiktionary, I assume? https://en.wiktionary.org/wiki/Reconstruction:Proto-Slavic/govorъ
I have heard some complaints about it, but it’s pretty useful for this sort of thing.
Hylander: If you haven’t already seen it, you might be interested in Willi’s The Origins of the Greek Verb. I haven’t finished yet and I’m not an expert in PIE (or pre-PIE!), but I’m enjoying it. The reduplicated aorists are key in his reconstruction. He argues that they were the earliest type of perfective in PIE (even earlier than the root aorist) and that reduplicated imperfectives were secondary formations. I also enjoyed this short summary and discussion of his book with James Clackson: https://www.youtube.com/watch?v=f4ZNplnLdlc&t=2514s
mwh has mentioned this video here before. Willi wrote a book with Probert, I know.
Thanks, phalarkos and Joel. I have the book, and I’ve watched the video with Clackson. I started working through the book, but I felt I needed to shore up what little I know about Proto-Indo-European before taking a deep dive into Pre-Proto-Indo-European, all the way down to the anti-passive. I’m fascinated by efforts to find the regularities underlying the apparent bizarre irregularities of Greek verbal morphology.
the same story with aor of αἱρέω
No, εἷλον is the augmented form; the inf. is ἑλεῖν. From the PIE root *selh1-, cognate with English “sell”(!).