Papers of the American School of Classical Studies at Athens (4.1885-1886)

266

J. McKEEN LEWIS,

except, of course, where this is added, as in the diphthong r;t. This is
demonstrated by the entire absence of any confusion between 77 and
171 in inscriptions of good date." The character of the r; is con-
veniently illustrated by the well-known verse : 6 0" rjh'dios ii<s n-pofid-
tiov fir) (ir) Xiytav /3aoY£et (Cratinus), on which the E. M. has the

gloss : [3rj ■ to fxijj.rjTLKoi' T77S tuiv irpofiaTuiV cfxiiryj'; • OUYl fial \eyerai

'ArrtKais. The form /3at is the natural rendering of the same sound
in later times, when at had usurped the pronunciation originally
belonging to 77. Thus also before the beginning of the fourth cen-
tury, the Boeotians, with whom the process of vowel-degeneration
was more than a century in advance of that of the other dialects,
borrowed the new vowel-sign 77 from their Ionian neighbors to serve
in place of at, which in Boeotia had already ceased to be a diphthong.
The at, as is well known, has continued to the present day to be
sounded in Greece as it was at that time in the most corrupt of the
dialects.

Two of the <?-sounds long continued to be represented by E after
77 began to be used as a vowel-sign in Attica, — namely, c, and
the long vowel arising from its " compensative " lengthening, or from
the contraction of cc. This will be called, for convenience, e or
long £. That it had a narrower sound than 77, and was in fact a
closed or quasi-diphthongal vowel, is indicated by its graphical con-
fusion, soon after the year of Eucleides, with the diphthong ei. It
was felt to be (as its functions show) qualitatively equivalent to ?,
and was thus during a long period written with the same symbol.
From this it may be inferred that ? had also a sharp or closed sound,
rather like the e of Italian venti than like the short e of our own
language.

We are thus able to distinguish, in the Ionic dialects, four or five
<f-vowels, with three degrees of divergence from a : a broad 77 arising
from a (as in 01*177, vucqcriij), perhaps its corresponding short (as in
aWemv, veak) ; an aboriginal and narrower 77 (as in [3ifi\r)Ka, d-roV}?) ;

than to suppose (with G. Meyer, Gr. 69) an actual phonetic modification of an
open ?-vowel to i first consummated in Ionic and afterwards in Doric. Had such
a thing occurred in these dialects (as it certainly did in Thessalian), it should
have affected all the ^-sounds of this class, and resulted in forms like xpe'^'M"' for
Xprffiuvs, ttfeive for t077Yf, etc.