Pronunciation of vowels

Greek Learning Greek
21

Thanks for looking at that sentence in Allen. I really just wanted you to validate what I was reading into it, not so much that the arguments re open or closed are that important. Allen is a curious book, with a combination of conciseness and sometimes difficult style.

I have nothing against pet as a reasonable English starting point. I agree that I really don’t hear the Greek epsilon as being lower than the e in pet.

With θεός could be that the epsilon is raised a little due to the following omicron.

One thing that comes into this is the tenseness of the facial muscles. The e in pet is maybe a little lax by comparison with Greek or French and that may influence how it sounds.

I think that omicron is a more difficult sound for English speakers than epsilon. It is midway between high- and low-mid and requires lip rounding and tenseness. It requires a bit of practice and I actually learnt it (or think I learnt it) by the method that I gave by starting with the French nasal -on. Sort of trick that I discovered. That “true-mid” vowel is pretty useful for other languages also.

22

First of all, this is not my query but that of Mitch.

The IPA system has two great advantages: the symbols are standardized, and everyone may find the acoustic value of a symbol in his own language.

Not using the IPA put one in the obligation of using symbols which value is not easy to find and, as Katalogon well underlined, of finding equivalents in various languages which the readers of the same language of the writer does not know.

And indeed Allen appears to be using:
— for IPA [e], the close mid front: an e with a point under it and
— for IPA [ε], the open mid front: an e with sort of an inverted comma under it
— for IPA [e̞], the mid front: by using his chart on page 62 as Katalogon did we may suppose that Allen uses a plain e.

Page 63 he writes:

“There is no reason to think that the sounds represented by these letters were ever other than short mid vowels, front* and back* respectively, i.e. rather like the vowels of English pet and German Gott. 3” The view that they were of a specially close mid quality, i.e. [~], [~], as in French gai, beau, is probably
mistaken (cf. pp. 72, 89 f.).
Note 3 : “The vowel of English pot is decidedly less accurate, being fully open rather than mid.”

In the article about the mid front vowel [e̞] here:

https://en.wikipedia.org/wiki/Mid_front_unrounded_vowel

the example given for the Received Pronunciation is let [le̞t] and there exists no mid front vowel [e̞] in the French phonology.

Then page 64, after a paragraph about the similarities with Latin î, Allen concludes by:
“Thus the E in Greek words of this type may, even in Attic, stand for a specially close variety of [e] which would then be particularly near to the Latin î. 4”

Finally, in The summary of Recommended Pronunciations page 177 one only finds:
ε As in English pet 63f.

Now the problem is that the pronunciation of pet is:

/pet/ in the Cambridge Dictionary:

https://dictionary.cambridge.org/dictionary/english/pet

/pet/ in the Oxford Learner’s Dictionary:

https://www.oxfordlearnersdictionaries.com/definition/english/pet_1?q=pet

but /pɛt/, [pʰɛt], [pʰɛʔt] in Wiktionary:

https://en.wiktionary.org/wiki/pet#Pronunciation


So by not using the IPA, you cannot know precisely what one is saying.

The Cambridge Grammar of Classical Greek has the great advantage of using the IPA system: at least you know what they mean.

It would be very interesting to know the reasons of the choice of the writers of the CGCG of the pronunciation of epsilon as the close mid front vowel IPA [e]. As Mitch has noted, they agree with Threattle (who page 137 define epsilon by the IPA [e]).

Of course I am an absolute layman here and for sure I don’t understand everything but, for what I have read, the reasons of Allen appear to be neither very numerous nor very strong.

You are all very experts on Ancient Greek: how do you analyze the work of Threattle about epsilon being IPA [e]?

23

Here one may find a wonderful IPA vowel chart with audio (meaning with the sounds of all vowels):

https://en.wikipedia.org/wiki/IPA_vowel_chart_with_audio

24

Again, Threatte isn’t trying to make the same sort of arguments as Allen about the precise openness or closeness of the vowel here. I wouldn’t read so much into his choice of IPA symbols. However, Threatte does claim that that Latin i is much more closed than Greek ι, and is therefore closer to ε, which explains its use in transliteration from Latin to Greek. Allen, however, takes this as evidence of the opposite, and that it was Latin i that was peculiarly open. Allen’s arguments for the open value of Latin i are in Vox Latina (see pg. 47-48). I don’t know what Threatte relies for his idea of a more closed Latin i.

25

So by not using the IPA, you cannot know precisely what one is saying.

A general comment. IPA is great, but …

IPA is not going to give you the correct pronunciation of a language. It is just a rough categorization, useful as a standard for discussion.
IPA is an idealization. The symbols do not really correspond to actual sounds.

I would invert your assertion:
By using IPA, you will not necessarily know how to say anything correctly.

even worse,

By using IPA, you may well be misleading yourself.

As Joel said, don’t read too much into the choice of IPA symbols.

26

This is because most British English dictionaries (and English language teaching materials) haven’t changed the way they use IPA. They stick to the the values of mid-Twentieth Century Received Pronunciation, even when standard pronunciation has now changed.

In the case of /e/ and /ε/ this is relatively innocuous - the sounds they represent are pretty similar, after all, and this won’t lead to confusion. It’s a bit stranger when a word like ‘air’ is listed as containing a diphthong, when in standard British pronunciation nowadays it’s a monophthong best represented by /ε:/ - and so, incidentally, an excellent example to use for the pronunciation of ancient Greek η.

RP, by the way, was never the pronunciation used by the majority of British speakers, nor a particularly conservative one - it actually contained some fashionable innovations that did not spread particularly widely in the UK and others which have now fallen out of favour.

27

There’s a great video on vowels and IPA here: https://youtu.be/FdldD0-kEcc?si=dlhH88V7XoSvn5s7

28

Katalogon, I’m afraid I do disagree on this. On the contrary, by using the IPA system you may refer to multiple examples, and in your own language which you’ll always master far better than any other foreign language however familiar you may be with.

If I look for the [e] sound in different websites and different books I’ll find été, clé, beauté, blé, nager, etc.
If I look for the [ε] sound I’ll find fête, bête, mettre, paître, etc.

Even if there are local variations, for certain you’ll know what sound was meant by the author.

But if you don’t use such a system, you have no other choice than to rely on one or only few examples in your language or, even, foreign languages.

For the omicron sound, Allen has to fall back on the German Gott because:

“The vowel of English pot is decidedly less accurate, being fully open rather than mid.”

For the close mid front sound, he has to fall back on the French gai, which sound, as we saw here, may easily be misinterpreted.

Here Allen says one place that:

“The view that they were of a specially close mid quality as in French gai, beau, is probably mistaken.”

But in the summary of recommendations he gives the word bet as a model and the students will then pronounce omicron as they do bet..

But the Cambridge, the Oxford’s Learner’s and the Longman dictionaries give the pronunciation of bet with a close mid front [e] (and by the way, none of you seems to agree on the pronunciation of the others), a close mid front that precisely Allen seems to reject!

Moreover we may ask what precisely is “a specially close mid quality”.

Why, on Earth (by Zeus rather) to masochistically delight in complications and difficulties?

Cordial and friendly joke here: One may as well choose so crazy, foolish, and surrealist system of units and measures as the imperial system which divides units by 12, 4 or 6… Okay guys, I’ll leave.

29

In the case of /e/ and /ε/ this is relatively innocuous

In English perhaps MattK but not for sure in French and in other languages…

30

Exactly right. A foreign learner needs to know that when the dictionary says /e/ it means the English sound, which is actually /ε/. But once they know that they hopefully shouldn’t get too confused…

31

OP here, thanks everyone for an illuminating discussion on several fronts :slight_smile:

So reading thru this thread and what I think I can understand so far from trying to read the relevant sections of Vox Graeca 3rd ed. it seems to me that:

  • CGCG says to pronounce ε like IPA [e] i.e. a short close-mid front unrounded vowel while Allen says that during classical times ε was probably a short open-mid front unrounded vowel i.e. like IPA [ε]


  • CGCG says to pronounce ο like IPA [o] i.e. a short close-mid back rounded vowel while Allen says that ο was probably a short open-mid back rounded vowel i.e. like IPA [ɔ].


  • CGCG and Allen mostly agree on the pronuniations of η, ω, ει, and ου.

Is that correct? :slight_smile:

And @jeidsath you said that “Threatte’s IPA system in the introduction isn’t meant to be a guide to vowel quality. He had to pick a convention for talking about how the vowels changes over time, and that is what he uses.” Is there somewhere where Threatte makes this point explicitely? Because as I’ve noted before, the vowel table in CGCG sec. 1.15 seems to me to match what Threatte says on page XXVIII of vol 1.

Thanks!

32

Open-mid was my belief about what Allen was saying, but katalogon convinced me that Allen was really arguing for mid.

33

You’re right about the CGCG, but not Allen. Allen argues that ε was a mid vowel, not an open-mid vowel. This is clear if you look at the chart on p.62, where his ε is halfway between his ει (close-mid) and η (open-mid) in terms of openness. For η as open-mid see his p.69ff.

34

Again, Allen argued that Omicron was mid, not open-mid. As Joel noted earlier, there is a slight difference in where CGCG and Allen place it on the front-back axis.

35

@MattK - I don’t understand the vocalizaton symbols Allen uses in his table on page 62 as they don’t match those on the Doulos SIL chart at https://www.internationalphoneticassociation.org/sites/default/files/IPA_Doulos_2015.pdf. Is there an older IPA chart that Allen is drawing his pronunciations from?

36

I don’t know where he got his symbols from, unfortunately. On p. 11 he gives his reasons for not using IPA - it wasn’t standard, in his view, at the time he wrote his book. You can figure out what Allen’s symbols mean in IPA terms by looking at his chart. A dot under the letter seems to mean ‘closer’ (i.e. close-mid instead of mid) and the backwards comma (I’m sure it has a proper name!) ‘more open’ (so open-mid instead of mid)

37

I also tried to find out where these symbols came from.

Mattk is right:

Allen appears to be using:
— for IPA [e], the close mid front: an e with a point under it,
— for IPA [ε], the open mid front: an e with sort of an inverted comma under it
— for IPA [e̞], the mid front: by using his chart on page 62 as Katalogon did we may suppose that Allen uses a plain e.

Hence the interest of using a system… okay, I repeat myself.

38

I just watched one of these videos, and it’s fantastic. Thanks for sharing.

39

Going by CGCG’s vowel triangle on pg. 8, they mean for their ε and ο to be more open than ει, and in fact they position them right across from the mid.

40

@MattK, Curius - Just dug up https://en.wikipedia.org/wiki/Diacritic and here’s what it says:

“In academic notation of Old Latin, ẹ̄ (e with underdot and macron) represents the long vowel, probably /eː/, that developed from the early Old Latin diphthong ei. This vowel usually became ī in Classical Latin.”

The above is what Allen has for the pronunciation of ει on page 62 of his book. I’m not sure about the hook symbol under e for η and under o for ω.