Greek & Latin vowels (was: CHAT letter names etc)

On Tuesday, March 2, 2004, at 04:39 PM, Mark J. Reed wrote:

> An interesting tidbit of which I was unaware until yesterday: as in
> Latin, what was originally purely a quantitative difference between
> short and long vowels developed over time into a quantitative
> difference.  And, as in Latin, the resulting two vowels for <o> were [o]
> and [O].
Yep - the different development of the two vowels in the Romancelangs make
that quite clear. In western Romance, short Latin /u/ was [U] and fell
together with long /o/, both being pronounce [o]. Similar developments
took place with the front vowels.

> However, their rôles in Greek were reversed - short <o>
> (mikron) was [o], while long <o> (mega) was [O:].
Probably because, like some rustic English dialects, the long vowels were
drawled and laxer than the tense, short vowels. In some (probably now
obsolete) southern English dialects, long-o (RP [ou_^]]) had become [o@_^]
or [U@_^] and RP [ei_^] was [e@_^] or [I@_^]. Actually in French, altho
length is not phonemic, the /E/ in words like 'père', 'terre', 'hêtre' are
the low and long [E:], while the high /e/ is always short and tense.
Ancient Greek seems to have behaved in a similar way.

The qualitative difference between omicron and omega was probably not
great in the later period. The evidence with E, EI, and H is much clearer.

H (eta) /E:/ resulted in the Ionian dialects from PIE /a:/.  A similar
phenomenon occurs today in the colloquial English of Newport & Cardiff in
south Wales where English /A/ is pronounced [&:] (thats's the CXS [&] =
X-SAMPA [{]).  Eta also, in all dialects that used the letter, represented
the development of PIE /e:/, therefore we assume the Greek sound was
between [a:] and [e:], namely [E:].

Those Greeks who used the western varieties of the alphabet used the
symbol H for /h/, and represented both /e/ and /E:/ by E.

But in early inscriptions from both areas we find E also represented a
sound which later Greek became spelled EI and which we know that by the
Hellenistic period had become /i:/. There is no evidence that it
diphthongized en_route, so to speak. The evidence suggests that by the 5th
cent. BCE both earlier /ei_^/ and "native" /e:/, written E in both east &
west, had merged to give [e:]. The evidence is that in the Classical
period, ancient Greek had two long-e sounds and two long-o sounds, rather
like Middle English with low 'ea' and and low 'ee', and low 'oa' ~ high
'oo'. As E was used in early texts for both /e/ and [e:], we assume /e/
was pronounced [e].

Likewise the long sound spelled O in both western and eastern Greek
alphabets was later spelled OY. Again it seems that both original [ou_^]
and [o:] merged as just plain [o:]. The long-o inherited from PIE was,
however, written by the Ionians with a letter they invented which we call
omega. That they felt the need to do this suggests not only a quantitative
difference but also a qualitative difference, i.e. it was a lower vowel
and, on the clearer evidence of E, EI, H, we assume omega was /O:/ (like
Middle English 'oa').

Thus it's incorrect, in fact, to say that in (early) Classical Greek,
short /o/ and /e/ were high and long /o/ and /e/ were low vowels. The
truth is that, like Middle English, the two short mid vowels, /o/ and /e/,
each had _two_ contrasting long vowels: one high and one low, thus:
           Short    Long
mid-high           OY  'oo'
mid         O
mid-low            Ω (omega) 'oa'
and..
           Short    Long
mid-high           EI  'ee'
mid         E
mid-low            H (eta) 'ea'

There is no evidence that there was any qualitative difference in ancient
Greek between long and short vowels at the 'apexes' of the vocalic
triangle, i.e. the difference between long and short I, A and Y was
essentially one of length only, as in modern French (yes, I know length
isn't phonemic in French), cf.
short [i] - vite, dit
long [i:] - rire, Moïse
short [a] - chat, tache, femme
long [a:] - rare, noire
short [u] - tout, goutte
long [u:] - cour, Douvres

Thus when length ceased to be phonemic in Greek, the long and short sounds
simply fell together which was not the case in Latin except for /a/ ~ /a:/

Latin seems to have a system like modern German where, except for /a/ ~ /a:
/, the short sounds were laxer than the tenser long sounds; so when length
ceased to be phonemic, the qualitative difference did become phonemic, as
we see in the development of the Romancelangs (including Brithenig).

> That just seems very
> odd to me. :)
Then natlangs just are odd   :)

For example, the nice neat arrangement of both the ancient Greek & Middle
English mid vowels didn't remain nor did the front & back vowels develop
the same way, as might be expected.

In Greek, /u/~/u:/ were fronted to /y/~/y/ probably as early as the 5th
cent. BCE in Ionia, Attica & Boiotia; and it seems as though OY didn't
waste too much time moving right up to take the vacant place, becoming [u:
]. The change is likely to have been completed by the 4th cent BCE. The
qualitative difference between /o/ (omicron) & /O:/ (omega) became less
pronounced and by the Byzantine period the two sound had fallen together.

But with the front vowels, while EI followed the upward movement of its
'sister' OY, and had became /i:/ by Hellenistic period, /E:/ (eta) wasn't
going to hang around and make friends with E (epsilon); it got going on
its way up. By the Roman period it had become [e:] and by the time of the
Byzantine period had moved up and joined long I and EI to have the sound
[i:].

It's interesting to note that while in English only 'oo' moved to [u:]~[U]
, 'oa' keeping what call an 'o-sound', both 'ee' and 'ea' (with few
exceptions) both became [i:]

Yep - languages just are odd   :)

Ray
===============================================
http://home.freeuk.com/ray.brown
ray.brown@freeuk.com    (home)
raymond.brown@kingston-college.ac.uk   (work)
===============================================
"A mind which thinks at its own expense will always
interfere with language."         J.G. Hamann, 1760

Reply