Learning Ancient Greek Vs Koine Greek ?

If I had the opportunity to begin again, I would likely choose Homeric Greek as my starting point. Regarding pronunciation, I tend to prefer modern Greek to Erasmian. Obviously there's plenty of room for good argument on both of these points, but considering my amateurish knowledge of these topics, I'll choose to run away before I'm made too much a fool.

Which dialect to start with depends on your goals. If you want to be able to read broadly, starting with either Attic or Homeric is probably best.

For Attic Greek, there's a book called Vox Graeca by W. Sidney Allen that inspired a shift to something sometimes called 'restored' classical pronunciation that I believe has gotten a lot of traction in universities (not so much in seminaries). There are some really interesting audio recordings available based on it. Stephen Daitz has some recordings published by Bolchazy-Carducci that include the use of pitch accents, rather than stress accents. The last time I dug into this issue I noticed that Daitz treated Omicron differently than Allen, and that surprised me. The Cambridge series JACT offers a CD called Speaking Greek that is also based on Allen's work. If I'm remembering right, it's handling of Omicron/Omega was more in line with what I read in Allen, but it uses stress accents instead of pitch accents (I think there are a couple short examples of pitch accents, but they aren't used throughout). In comparing the CDs, Speaking Greek benefits greatly from having many different voices, and in my subjective opinion, the voices sound more mellifluous. But it's just one CD, so you're dealing with excerpts, where Daitz has CDs available for whole works, including entire plays, both of Homer's epics, an anthology of poetry and entire orations. If you're particularly interested in the oral performance of poetry or drama, he's well worth checking out. Mark Miner also has some very dramatic 'restored classical' recordings available online (again, with some differences in Omega/Omicron, different in his own way from both Allen and Daitz, but very fun, lively readings).

(I wrote a post on Omicron/Omega that might be of interest: http://community.logos.com/forums/p/52319/383828.aspx#383828 )

When I first encountered the restored pronunciation, it took a lot of work to change the way I said the Eta from being identical to the way I said the Epsilon-Iota diphthong to something more open, like the 'a' in the English word 'hat' (for some reason the other changes were very easy for me to adopt). But when I subsequently learned years later that my pronunciation of Omicron and Omega wasn't actually in line with Allen/JACT OR Daitz (though it did match Miner's recordings), I lacked the energy to try to make another major change to the way I say the vowels. At a certain point, the way I'd done it for years just sounded 'right' to me and I could tell it would be a painful switch.

There is a lot of debate about exactly when the individual changes from something like the restored pronunciation to modern Greek happened. By the time of the NT, I think everyone agrees that pitch accents were long gone. There were a lot of shifts in vowel sounds, but I'm sure there's debate about which shifts would have already taken place by the middle of the Hellenistic period. So there's probably some anachronism introduced by reading the NT with a restored Attic pronunciation, even if you drop the pitch accents. I know Randal Buth has come up with a reconstruction of Hellenistic/Koine pronunciation.

I learned Erasmian pronunciation in seminary, but switched over to Restored Koine a few months ago. A comparison of the pronunciation schemes (and a defense of Restored Koine) follows in this document prepared by Randall Buth.

http://www.biblicallanguagecenter.com/koine-greek-pronunciation/

I wouldn't let pronunciation be the determining factor as to whether you should learn Attic or Koine. The switchover from Erasmian to Restored Koine was actually pretty easy. I pretty much had it down after a week of reading Greek, but I can still read it in Erasmian also.

As far as US versus UK pronunciation, in Erasmian, the omicron would be pronounced /ɔ/ in IPA. However, since Americans don't use the /ɔ/ sound, it often comes out /a/. Unfortunately, this makes the omicron and the alpha sounds indistinguishable. Jonathan Pennington's New Testament Greek Vocabulary has this issue.

However, since Americans don't use the /ɔ/ sound, it often comes out /a/. Unfortunately, this makes the omicron and the alpha sounds indistinguishable.

Buth's work is interesting--when I have more time I will probably dive in (whenever that might be). But with regard to the comment above, an observation.

First, lots of different letters have the same sound in many different languages. The allure of Erasmian, of course, is that it makes each letter have its own sound, or nearly so. This is an attractive concept, but I'm not sure how much one can expect it to be followed in actual historical practice. Rarely, if at all, I would think. The tendency of humans is to go lazy in most pursuits, and language follows that trend. Modern Greek is a perfect example, where the original 10-12 vowel sounds have been reduced down to about 4-5 (actual numbers may vary), to the sounds that are the least taxing [8-)] to pronounce, thus mostly short sounds remain.

However, since Americans don't use the /ɔ/ sound, it often comes out /a/. Unfortunately, this makes the omicron and the alpha sounds indistinguishable.

Buth's work is interesting--when I have more time I will probably dive in (whenever that might be). But with regard to the comment above, an observation.

First, lots of different letters have the same sound in many different languages. The allure of Erasmian, of course, is that it makes each letter have its own sound, or nearly so. This is an attractive concept, but I'm not sure how much one can expect it to be followed in actual historical practice. Rarely, if at all, I would think. The tendency of humans is to go lazy in most pursuits, and language follows that trend. Modern Greek is a perfect example, where the original 10-12 vowel sounds have been reduced down to about 4-5 (actual numbers may vary), to the sounds that are the least taxing  to pronounce, thus mostly short sounds remain.

Very true, David Paul. I am in no way in opposition to the idea that different (combinations of) letters can have the same sound. However, regarding the alpha and the omicron, I'm in agreement with Buth that they should not be pronounced the same. The reason is that Buth's research shows that, historically, alpha and omicron WERE NOT pronounced the same, and that omicron and omega WERE pronounced the same.

One major advantage of Buth's Restored Koine is that letter combinations that were pronounced identically in the Koine period (ο and ω, ει and ι, αι and ε, υ and οι) are pronounced identically. Textual variants of these can be explained as dictation errors.

Erasmian, on the other hand, has letter combinations that are pronounced identically that weren't ever pronounced identically (ει and η, υ and ου, sometimes α and ο).