At 6:59 pm -0500 28/3/02, J Y S Czhang wrote:
[snip]
>    Talking of Greek... do to havin' few books that have current IPA (if at
>all) on Greek, I am not too clear on what exactly was/is the vowel
>inventories of Greek in the various phases of its development (i.e.,
>Classical Greek, Hellenistic Greek/_koine_, Byzantine Greek, etc.). My
>websurfing got me conflicting and erroneous transcriptions &/or didn't have
>IPA.

Oh dear, with various interruptions, it taken three days to do this -
probably been answered already by now  ;)

Well, classical Attic (i.e. Athens, 5th cent. BC) had five short, and seven
long vowels

(a) Short vowels
    ------------
The original Greek system of short vowels was just the 'classical 5', i.e.
        Phonemes           Graphemes*
     /i/       /u/      {I}         {Y}
       /e/  /o/           {E}    {O}
         /a/                 {A}

(There is no implication that /e/ and /o/ were high mid, rather than low
mid; we just have no way of knowing.  They were probably mid-mid :)

*lower case symbols were developed later.

That remained the system in the Doric dialects right up till the
Hellenistic period when Doric began to give way to the internation Koine
Greek.

In the Ionian dialects, including clasical Attic, /u/ had shifted to the
front, but remained rounded, thus:

     /i//y/
        /e/    /o/
           /a/

The later Koine was essentially Attic Greek, modified of so that purely
Athenian peculiarities gave way to more common Greek forms (e.g. Attic
_thalatta_ (sea) was _thalassa_ in Koine), and the classical Attic short
vowels remained in Koine and well into the Roman period.

Sometime during the Roman period /y/ started to get unrounded and this
process appears to have become complete by the 4th cent AD; this would, of
course, have given just four short vowels.  But another process had also
been going on in the Koine: the ancient pitch accent had been giving way to
a stress accent and, with this development, the distinction between long
and short vowels ceased to be phonemic, so the old /u:/ now simply became
/u/, so the Byzantine and modern Greek system is simply:
    /i/       /u/
       /e/  /o/
         /a/
with no phonemic distinction in length.

(b) long vowels
    -----------
  The early history is not quite as simple as the short ones, but there
appears to have developed in most dialects a seven vowel system thus:
               Phonemes            Graphemes
          /i:/          /u:/       {I}             {Y}
            /e:/       /o:/          {EI}       {OY}
              /E:/   /O:/               {H}  {‡}
                 /a:/                     {A}

The vowels /i(:)/, /u(:)/ and /a(:)/ do not seem to have different
significantly in quality between the short and long versions.

But the mid vowels developed long mid-high & mid-low versions (as Middle
English did).  The long vowels inherited from ProtoIndoEuropean appear as
low-mid vowels in Greek; furthermore, PIE /a:/ survived in ancient as /a:/
in all environments only in the Doric dialects.  In Ionian dialects it
normally became /E:/, thus merging with PIE long-e.  In Attic, however,
while it normally became /E:/, the older /a:/ was retained if it was
preceded by /e/, /i/ or /r/.

But Greek developed a secondary lengthening through certain types of vowel
change, e.g. Vns >> V:s, thus /ans/ >> /a:s/, /ens/ >> /e:s/.  In these
developments the mid vowels developed long high-mid versions.  Thus we have
the 7 long vowels above.

In the western alphabets, {E} was often used to repesent /e/, /e:/ and
/E:/, likewise {O} was used for three phonemes.  The Ionians adopted the
use of {H} = /E:/ and {‡} (omega) = /O:/.  Also diphthongs /ei/ and /ou/
had merged with /e:/ and /o:/ at an early date and use was made of this in
Athens to provide distinct representation for the two long versions of each
of the mid vowels, as shown above.

 In 5th cent. Athens the /u:/ shifted to the front, /o:/ was raised to /u:/
and /O:/ to /o:/ thus:
           /i://y:/         /u:/
            /e:/       /o:/
              /E:/
                 /a:/


By the time of the Koine, the two front mid vowels had also been raised, so
that /e:/ fell together with /i:/, and /E:/ became /e:/, thus:
           /i://y:/          /u:/
            /e:/       /o:/
                 /a:/
This was what the Romans knew of Greek; but during the Roman period, /e:/
was eventually raised to /i:/ and /y:/ was unrounded, becoming /i:/ also
and with the loss of phonemic length, they fell together with the short
vowels giving the Byzantine & modern system shown above.

(c) Diphthongs ending in /i/ [j]
    ----------------------------
   Ancient Greek had two sets, one with long vowel + /i/ and and one with
short vowel + /i/, thus:
  /E:i/ /A:i/ /O:/                {HI} {AI} {‡I}
    *   /ai/  /oi/  /ui/=[uj]          {AI} {OI} {YI}

* original /ei/ has become /e:/ by the Classical period.
Also it likely that /ai/ and /oi/ has already begun to shift towards /e/
and /y/ at least in word final position as early as the 5th cent. BC, since
the two diphthongs are nearly always reckoned as short in that position.
Certainly during the Koine, /ai/ gave way to /e(:)/ and /oi/ became /y(:)/,
the latter eventually becoming unrounded and joining /i/ by the Byzantine
period.

[ui] >> [yi] in Classical Attic and had probably fallen together with [y:]
by the time of the Koine, thereafter it developed just like [y:] >> [i(:)]
of Byzantine & modern Greek.

As for the "long diphthongs", by the 2nd cent. BC, they had lost the final
/i/ and become identical with the developments of the corresponding long
vowels.  It became the habit write the iota _beneath_ the vowel (the
so-called "iota subscript") when lower case forms developed, and this is
the modern practice when writing ancient Greek (and Katharevousa).

(d) Diphthongs ending in /u/ [w]
    ----------------------------
   Ancient Greek also had two sets of these:
   /E:u/ /A:u/ /O:u/             {HY} {AY} {‡Y}
   /eu/  /au/   *                {EY} {AY}

* /ou/ had become /o:/ by the Classical period and, indeed, shifted to /u:/
in Athens before the 4th cent. BC.  It has remained [u(:)] ever since.

Also, we can discount /O:u/; it remained for a while in Ioanian Greek, but
was rare in Attic Greek and had gone by the time of the Koine.

When /u/ to /y/ in Attic Greek, the second element of these diphthongs
remained [w] and appear to have kept this prononciation until the end of
the BC era; sometime, however, between the 1st & 4th cent AD [w] >> [v],
with assimilation to [f] before voiceless consonants.  This was the
Byzantine and is still the modern pronunciation, i.e.
 Classical Attic    Byzantine & modern
 ----------------   ------------------
   /E:u/                /iv/
   /A:u/                /av/
   /au/                 /av/
   /eu/                 /ev/

-------------------------------------------------------------

A useful book is "Vox Graeca" by Sidney Allen.  I don't agree with his
speculations on ancient Greek stress, but it is pretty sound on the actual
pronunciation of vowels & consonants.

Ray.



======================
   XRICTOC ANESTH
======================