Ancient Greek Questions

[Jetboy]

So, I've been taking an Ancient Greek course this year, and the language is terrific. The multitude of active phonological rules, the utility of the article, the shifting meaning of prepositions based on what case they take, the middle voice, and many of the other things we've covered have all been great fun, and we haven't even gotten to any non-present tenses yet (though I think they're coming up soon). Having stumbled upon a summary of the language's historical phonology, a lot of the seeming irregularities make much more sense, but I still have some questions, mostly about Ancient Greek's diachronics:

Firstly, the /aː/ > /eː/ shift that occurred in Attic seems to have left /aː/ around in plenty of places besides the oft-named exceptions of /e i r/; for example, why didn't <βοᾷ> become <βοῇ> in the same manner that, say, <καλῇ> had? Similarly, why do the 1st declension plurals retain /aː/? This applies especially the accusative plural, which is identical to the genitive singular in nouns with stems ending in /e i r/.

Secondly, the declensions of some third declension vowel-stems are rather puzzling; while <ναῦς> is a thing of beauty that makes perfect sense without appearing to, the happy buzz from figuring out its paradigm quickly faded upon seeing those of <πόλις> and <βασιλεύς>, which seem to thwart any attempts to comprehend them (especially the latter). Recalling that, like <ναῦς> and <βοῦς>, <βασιλεύς> had historically had a final digamma to its stem only compounded this confusion; why were its accusatives not <-εῦν> and <-εῦς> like those two? And why didn't the accusative singular contract to <η>, as did the nominative and accusative plurals of <ἄστυ>? Similar questions appear about <πόλις>; I know the stem was historically /pɔleːj/, but this doesn't seem to do anything to explain the presence of /i/ in the nominative and accusative singular, and in fact raises further questions about the length of the /e/ in the datives. And finally, where does the final /y/ in <ἄστυ> come from?

Any help understanding these enigmas would be much appreciated; so far my searches have been to no avail.

[Basilius]

Wow. Good questions, really :)

Similarly, why do the 1st declension plurals retain /aː/? This applies especially the accusative plural, which is identical to the genitive singular in nouns with stems ending in /e i r/.

This one is easiest: /aːs/ in acc. pl. comes from /a(ː)ns/; other dialects have different reflexes, e. g. /ans/ and /ais/.

Similar questions appear about <πόλις>; I know the stem was historically /pɔleːj/, but this doesn't seem to do anything to explain the presence of /i/ in the nominative and accusative singular, and in fact raises further questions about the length of the /e/ in the datives.

Just to show how it could be still more complex than it looks: in some dialects, πόλις is declined as if the stem ended in digamma (dat. sg. /poliwi/ etc.). This was probably due to an analogy, but you can imagine how such early analogy could complicate the matters at later stages.

And finally, where does the final /y/ in <ἄστυ> come from?

I am not sure I understand this one; why not from PIE /u/?

(Sorry, I remember I've seen the answers to the rest of your questions, but, not relying on my memory, I need to consult some books so I don't trick you, unwillingly.)

Also, grammars and manuals with focus on literary Attic tend to give an oversimplified picture, IME; I suggest trying something on Ancient Greek dialectology - you may forget much of the detail like I did, but you'll have a more adequate understanding of the whole :)

[vec]

I knew the answers to all of this a couple of years back, but I've forgotten all I learnt about it, basically.

[alice]

The best thing about Ancient Greek is being able to say "lesbian" with a straight face Although at least one book seems to prefer "Lesbic" for some strange reason.

[mapking27]

Couzu the Orincaryian sun god doesn't like when you worship the Ancient Greek gods instead of him. It makes him angry, very angry

[Jetboy]

This one is easiest: /aːs/ in acc. pl. comes from /a(ː)ns/; other dialects have different reflexes, e. g. /ans/ and /ais/.

Oh, dang it, I feel quite the fool now. That /n/'s responsible for the diphthongs of the accusative plurals of <ναῦς> and <βοῦς>, no?

And finally, where does the final /y/ in <ἄστυ> come from?

I am not sure I understand this one; why not from PIE /u/?

Sorry, I was a bit unclear there; why was there a /u/ there in the first place when the stem ends in /e/?

@Mapking: Kindly be a bit more relevant.

[vec]

The best thing about Ancient Greek is being able to say "lesbian" with a straight face Although at least one book seems to prefer "Lesbic" for some strange reason.

What's so funny about Lesbians or lesbians?

[Basilius]

This one is easiest: /aːs/ in acc. pl. comes from /a(ː)ns/; other dialects have different reflexes, e. g. /ans/ and /ais/.

Oh, dang it, I feel quite the fool now. That /n/'s responsible for the diphthongs of the accusative plurals of <ναῦς> and <βοῦς>, no?

Hmmm... no, I think these have been rebuilt by analogy (in Attic).

And finally, where does the final /y/ in <ἄστυ> come from?

I am not sure I understand this one; why not from PIE /u/?

Sorry, I was a bit unclear there; why was there a /u/ there in the first place when the stem ends in /e/?

No, it was -u alternating with -ew: dat. sg. ἄστει <- astewi, nom. pl. ἄστη <- astewa; I forget where gen. sg. ἄστεως got its ω from - it looks like quantity metathesis, but I don't see why....

What's so funny about Lesbians or lesbians?

It was them who had acc. pl. in /ais/ (and /ois/ in o-stems), that is, identical with dat. pl. That must have sounded ludicrous to the rest of Greeks. I am certain Bricka meant that and not their having something to do with Lezz Boss.

[Radius Solis]

The best thing about Ancient Greek is being able to say "lesbian" with a straight face Although at least one book seems to prefer "Lesbic" for some strange reason.

I'm not sure quite what your intention was, but it sounds as though you are saying you couldn't otherwise mention lesbians with a straight face. Even if that wasn't your intended meaning, it still looks like it enough that lesbians could reasonably interpret it as a comment hostile to them.

[Jetboy]

Hmmm... no, I think these have been rebuilt by analogy (in Attic).

Oh, I'd been guessing that the original form was a syllabic nasal followed by /s/, which would have explained everything except βασιλεύς.

No, it was -u alternating with -ew: dat. sg. ἄστει <- astewi, nom. pl. ἄστη <- astewa; I forget where gen. sg. ἄστεως got its ω from - it looks like quantity metathesis, but I don't see why....

But if there was a ϝ, why doesn't the dative plural pattern with βασιλεύς?

[Basilius]

Hmmm... no, I think these have been rebuilt by analogy (in Attic).

Oh, I'd been guessing that the original form was a syllabic nasal followed by /s/, which would have explained everything except βασιλεύς.

No, a syllabic nasal would produce an /a/, so e. g. for ναῦς the ending would be -η(ϝ)ας (IIRC indeed attested outside Attic), then (probably) -εᾱς.

No, it was -u alternating with -ew: dat. sg. ἄστει <- astewi, nom. pl. ἄστη <- astewa; I forget where gen. sg. ἄστεως got its ω from - it looks like quantity metathesis, but I don't see why....

But if there was a ϝ, why doesn't the dative plural pattern with βασιλεύς?

In the βασιλεύς type the stem ended in -ηϝ; note the vowel length (obtained via quantity metathesis) in gen. sg, βασιλέως ( <- ηος), acc. sg. βασιλέᾱ ( <- ηα), acc. pl. βασιλέᾱς ( <- ηας). In nom. sg. and dat. pl. (βασιλεῦσι(ν)) the diphthong shortened very early before a coda consonant, and in voc. sg. it was short already in PIE.

For dat. pl. ἄστεσι(ν) I can't think of anything but analogy (note that e. g. πόλεσι(ν) isn't phonetically regular either).

EDIT: in fact, all the datives in vowel + σι(ν) must have been rebuilt by analogy at some point (because of the intervocalic -σ- which would be regularly dropped; the pretext for the analogy was offered by the s-stems).

[Jetboy]

Ah, that would explain ἄστυ for the most part, then. I was thinking that the accusative plural was like the accusative singular (which was a syllabic nasal, no?); in nouns with vowel stems, the nasal would simply have stopped being syllabic, but between consonants, it would have become /a/, which would have been deleted by following /s/ in the plural; e.g., /nawn̩s/ /naʊn̩s/ /naʊns/ /naus/.

[Xephyr]

The best thing about Ancient Greek is being able to say "lesbian" with a straight face Although at least one book seems to prefer "Lesbic" for some strange reason.

I'm not sure quite what your intention was, but it sounds as though you are saying you couldn't otherwise mention lesbians with a straight face. Even if that wasn't your intended meaning, it still looks like it enough that lesbians could reasonably interpret it as a comment hostile to them.

lol, you said "lesbians"

[alice]

The best thing about Ancient Greek is being able to say "lesbian" with a straight face Although at least one book seems to prefer "Lesbic" for some strange reason.

What's so funny about Lesbians or lesbians?

Nothing at all, and I want to make it clear that I did not in any way mean to give any offence. This was probably not the best way to remark on the fact that it can be a bit disconcerting, especially if you aren't expecting it, to come across the word "Lesbian" (always capitalised) in an otherwise specialised academic context.

I'm getting way off-topic here, but a more extreme example of this is when textbooks on comparitive or historical linguistics throw in a few dirty words in other languages here and there to keep you awake, for example "We can also see the evolution of /ai/ to /u/ in FooBarese fuk "fuck", cognate with BarFooian faike.

[Jetboy]

So, more questions:
Firstly, what sound does <υι> represent? For some reason I feel it's something along the lines of /ɥ/, but that's just a guess.
Second, why the <ζ> in the nominative of Ζεύς, Διός? It seems to have something to do with palatalization of the /d/, but what triggers that in the nominative and not the other cases?
Finally, why does the sigma elide in the futures of nasal and liquid stems, verbs ending in -ιζω, etc., but not in contract verbs?

Also, thanks Basilius, I think I get βασιλεύς now!

[Drydic]

Second, why the <ζ> in the nominative of Ζεύς, Διός? It seems to have something to do with palatalization of the /d/, but what triggers that in the nominative and not the other cases?

It's an extremely common word, and they are far more likely to retain irregularities. Latin has a similar variation in DEUS (though without the initial consonant variation, which is I guess your main point but work with me here), and it wouldn't suprise me if Sanskrit/Avestan did as well (my computer just died, and I don't have all my drives/files hooked back in/found yet so I can't check).

[Sleinad Flar]

Second, why the <ζ> in the nominative of Ζεύς, Διός? It seems to have something to do with palatalization of the /d/, but what triggers that in the nominative and not the other cases?

It's an extremely common word, and they are far more likely to retain irregularities. Latin has a similar variation in DEUS (though without the initial consonant variation, which is I guess your main point but work with me here), and it wouldn't suprise me if Sanskrit/Avestan did as well (my computer just died, and I don't have all my drives/files hooked back in/found yet so I can't check).

Latin deus is from the thematic noun *deiwos however, which is also the source for Germanic tiwaz and Sanskrit devah. The Greek word is from the athematic root noun *djé:us, acc sg *djé:m, gen sg *diwós (this indeed has reflexes in Sanskrit dyáus, as well as Latin Ju-piter) The sound changes from PIE to Greek (*dj >Ζ but *di > * Δι) are perfectly regular AFAIK.

[Jetboy]

Ah, so it was irregular even in PIE. Well, that it is an unfortunately un-Greek-involving answer.

[Jetboy]

Some more questions, mostly about verbs:

Firstly, why is it that the aorist and future tend to share a stem, from which the present differs (obviously excluding sigmatic 1st aorists and the corresponding futures), for example, γίγνομαι versus γνήσομαι & ἐγενόμην, ἀποθνῄσκω versus ἀποθανέομαι & ἀπέθανον, or μανθάνω versus μαθήσομαι & ἔμαθον. It seems rather strange for the present stem to be less basic than the future and aorist.

Also, I recently discovered that the "primary" personal endings are actually historically derived from the "secondary" ones by the suffixation of /i/, which explains a lot about the endings, and especially about verbs like εἰμί. However, there are a couple things which I still find rather unclear: firstly, was the present suffix applied only to active, or also to the middle/passive? If the latter, it would handily account for the ι in -ομαι, -εται, etc., but would raise the question of the alteration between the α in the present and the ο in the aorist.
And lastly, how in the world did -ομι become -ω?

[Lleu]

Firstly, why is it that the aorist and future tend to share a stem, from which the present differs (obviously excluding sigmatic 1st aorists and the corresponding futures), for example, γίγνομαι versus γνήσομαι & ἐγενόμην, ἀποθνῄσκω versus ἀποθανέομαι & ἀπέθανον, or μανθάνω versus μαθήσομαι & ἔμαθον. It seems rather strange for the present stem to be less basic than the future and aorist.

Strange though it may seem, that is what goes on — the present stem is often not the root; cf. also ἀγγέλ-λ-ω vs. ἀγγελ-ῶ, ἤγγελ-κ-α. I'm not sure what function present reduplication (τί-θημι, γί-γνομαι, etc.) or nasal infixing (λα-ν-θ-άν-ω, π-υν-θ-αν-ομαι, etc.) had, but presents in -σκω, for example, derive from older iteratives, e.g. (ἀπο)-θνῄ-σκω ("die over and over", emphasizing the incompleteness of the action) from *θαν- "die" (aorist by nature). Some verbs show multiple derivations, e.g. δι-δά-σκω, from *δα-, which IIRC is related to δαίω "divide".

Re the alternation of -α(ι) and -ο in the mediopassive, I'm pretty sure the development into -αι is a later innovation; I believe there are traces of present -οι in Arcado-Cypriot, which retains a number of more conservative characteristics.

As for the development of -ομι into -ω, I feel like it might have been by analogy with the subjunctive, but I might be completely making that up.

[Jetboy]

Excellent, thanks!

Re the alternation of -α(ι) and -ο in the mediopassive, I'm pretty sure the development into -αι is a later innovation; I believe there are traces of present -οι in Arcado-Cypriot, which retains a number of more conservative characteristics.

Hmm, well I guess οι > αι makes more sense than ο > αι. Assuming that the change was analogical (since οι seems unchanged most everywhere else), what was it based off of?

As for the development of -ομι into -ω, I feel like it might have been by analogy with the subjunctive, but I might be completely making that up.

Ah, right, the subjunctive, I completely forgot about that (I haven't learned it yet ). That would make sense; I was thinking analogy, but couldn't imagine where it might come from.

One other question on personal endings– the present second & third singular. According to the book on historical linguistics I'm reading, they at one point both had the form ει, and the sigma in the second was restored on analogy with the second & third aorist. However, Wikipedia, I believe, explains the third present change of *-ετι to ει as analogy based on its relationship to the second present compared to the second & third aorist endings, though I think it may have also mentioned the whole thing being a mystery. Do we have any idea how the whole muddle sorts out, or is it just one of those impenetrable mysteries?

[Lleu]

Excellent, thanks!

You're welcome!

One other question on personal endings– the present second & third singular. According to the book on historical linguistics I'm reading, they at one point both had the form ει, and the sigma in the second was restored on analogy with the second & third aorist. However, Wikipedia, I believe, explains the third present change of *-ετι to ει as analogy based on its relationship to the second present compared to the second & third aorist endings, though I think it may have also mentioned the whole thing being a mystery. Do we have any idea how the whole muddle sorts out, or is it just one of those impenetrable mysteries?

This I do not know the answer to. The only thing I have to add is that Homer shows traces of an archaic third person singular present subjunctive in -ῃσι, with -σι coming from older *-ti.

[Sleinad Flar]

As for the development of -ομι into -ω, I feel like it might have been by analogy with the subjunctive, but I might be completely making that up.

Nope, inherited from PIE again: thematic *-o: vs. athematic *-mi. There seemingly never was a -omi.