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

NOTES ON ATTIC VOCALISM.

26S

there are yet some indications of this even in the later phases of the
dialect. I am inclined to regard the now authentic forms t«s, tiOw,
tcis, iriOas, etc., as a proof that aboriginal rj held a slightly different
place in the vowel-series from Ionic 17, which was at no time confused
with e (et). The form hrifhfs slipped into irffluK in the new Attic
orthography as naturally as /3ao-iAr;s, on the lips of later Athenians,
became /3ao-iAeis, or as reOrjKa. was changed, at a still later time,
to rWuKa. The laws of analogy, than which no* formative principle
is more marked in the development of the Attic dialect, would as
surely have produced Torcts, lorei, had the Ionic 77 of To-t^i been
identical in sound with the vowel heard in ridrftu, "17/u, /3am\rjs, etc.1
These indications, together with the facts that Attic was a direct out-
come of Ionic, and that the distinction was so marked in the latter
dialect as to call, in some localities, for graphical representation, leave
little doubt that it continued to exist at least into the fourth century
B.C. To define exactly the pronunciation of the Ionic 7/ would, of
course, be impossible. We may assume with much plausibility that
it lay between the a of English bad and the German a. The other
rj, which in nearly all cases is grammatically .related to e and belongs
purely to the ^-group of vowels, was not essentially different from the
long Italian ? as pronounced at the present time. It should be kept in
mind that whatever may have been its origin, an rj was always an open
vowel:2 in other words, no vanishing or /-sound was heard after it,

1 Though these changes were all occasioned by analogy, they could not have
occurred but for a close resemblance between the vowel-sounds interchanged.
Thus bara, XPU<*V< o-iSripa, were shaped after the corresponding uncontracted
inflections; bmUu suggested iriBeis just as iro'Aeiy suggested /WiAeis, and as
tha (Meyer, Gr. 71) supplied a reason for writing TtBcMa. This principle could
not, however, effect violent phonetic changes; it worked by stealth, not by force,
and practised its deception only with such nearly equivalent sounds as d and
Ionic 17, or 6 and aboriginal t).

2 The sound produced by the lengthening of e and contraction of es was at first
written universally E: the designation El appears first among the Ionians, the
Corcyreans, and Locrians; while the mass of the Dorians wrote E, and afterwards H.
That the vowel was sounded differently in Doric and Ionic appears not to have
been proved. The same may be said of Old-Doric a for ov. (So also Doric riji6v
for tov/j.6v, etc.) The difference was, perhaps, only an orthographic one, and the
Dorians kept, for the most part, the spellings with tj, a, until Ionian influence
caused them gradually to disappear. This seems to me a simpler explanation