DOI article:
Lewis, J. M.: Notes on Attic vocalism
DOI Page / Citation link:https://doi.org/10.11588/diglit.8561#0283
268
J. McKEEN LEWIS,
result being an e with a more or less perceptible vanish, but not a
diphthongal sound, this being impossible to a short vowel. Thus, in
this case at least, ei must stand for a simple sound which had long been
written f, the change being easily accounted for by the increasingly
monophthongal character of the genuine ei; so that e, e, a now meant
much the same vowel-sound, — a narrow f-souhd with a slight vanish.1
The change undergone by the diphthong consisted in the comparative
suppression of its second element.2 The same thing will be shown
to have happened in the case of ov. Examples3 of ei for e before
vowels are : Seiwn-ai, 119 (about 340) ; tSpuVciws twice, 168 (330) ;
Kina'w', ibid.;4 u&o'xpeiu), 578 (after 340) ;5 davrov, 115, p. 13
(after 350); aiav, 14, p. 11 (3S7) ; ILW9et[as] (originally with
diphthongal u, but cf. Il\m6rj%, etc., ibid.), or IIA.(u(9«as, for -diws, 570 ;
etc. These spellings are discussed by Herwerden,11 who remarks
that they are not found in inscriptions later than the early part of the
third century B.C. The cause of this is obvious : et was after that
time no longer qualitatively equivalent to e, but had begun to feel the
influence of itacism.7
1 It is remarkable that no sooner had graphical distinctions between the various
e- and i-vowels begun to be thought necessary by the Athenians than the real
distinctions began to be obliterated by phonetic decay.
2 The simplification of the diphthongs in Latin was nearly complete at the
date of the earliest literary monuments we possess in that language. That e could
have taken the contrary course, and been diphthongized, is against all analogies
presented by the history of phonetic decay. It has been supposed by Blass and
others that the itacism of ei resulted from an increasing preponderance of its
z'-element. This, however, seems quite incompatible with the confusion of ei and
*> *> vf- L
3 All of a good time (between 387 and 329). Nearly all will be found in
Meisterhans, Gr. pp. 21 seqq. Those here cited are from C. I. A., II. I.
4 As gen. pi. of Kmeus. We should, of course, expect Kinmy, but the rule is
not absolute even in the best time. Cf. Meisterh., Gr. pp. 56 seq.
6 a^ioxpeio', Kohler. But this is perfectly analogous to the other cases of et
written for e, and the accent should therefore not be changed. So in II. S72,
KoWvrdes must be read, not KoKAvreTes. In Aeiaiydpov, 553 (circa Eucl. ann.),
perhaps the earliest instance of this inaccuracy, we have et written for an Ionic e.
6 Lapidum de Dial. Alt. Testimonia, pp. 10, II.
7 The forms li/eiyi<ri for iveyitri, etc., frequent between 373 and 332 (v. Meisterh.,
p. 89), apparently on the analogy of Ion. ive'iny, show that the same confusion
could exist before a double consonant when, as before vowels, quantity could not
be affected by the interchange.
J. McKEEN LEWIS,
result being an e with a more or less perceptible vanish, but not a
diphthongal sound, this being impossible to a short vowel. Thus, in
this case at least, ei must stand for a simple sound which had long been
written f, the change being easily accounted for by the increasingly
monophthongal character of the genuine ei; so that e, e, a now meant
much the same vowel-sound, — a narrow f-souhd with a slight vanish.1
The change undergone by the diphthong consisted in the comparative
suppression of its second element.2 The same thing will be shown
to have happened in the case of ov. Examples3 of ei for e before
vowels are : Seiwn-ai, 119 (about 340) ; tSpuVciws twice, 168 (330) ;
Kina'w', ibid.;4 u&o'xpeiu), 578 (after 340) ;5 davrov, 115, p. 13
(after 350); aiav, 14, p. 11 (3S7) ; ILW9et[as] (originally with
diphthongal u, but cf. Il\m6rj%, etc., ibid.), or IIA.(u(9«as, for -diws, 570 ;
etc. These spellings are discussed by Herwerden,11 who remarks
that they are not found in inscriptions later than the early part of the
third century B.C. The cause of this is obvious : et was after that
time no longer qualitatively equivalent to e, but had begun to feel the
influence of itacism.7
1 It is remarkable that no sooner had graphical distinctions between the various
e- and i-vowels begun to be thought necessary by the Athenians than the real
distinctions began to be obliterated by phonetic decay.
2 The simplification of the diphthongs in Latin was nearly complete at the
date of the earliest literary monuments we possess in that language. That e could
have taken the contrary course, and been diphthongized, is against all analogies
presented by the history of phonetic decay. It has been supposed by Blass and
others that the itacism of ei resulted from an increasing preponderance of its
z'-element. This, however, seems quite incompatible with the confusion of ei and
*> *> vf- L
3 All of a good time (between 387 and 329). Nearly all will be found in
Meisterhans, Gr. pp. 21 seqq. Those here cited are from C. I. A., II. I.
4 As gen. pi. of Kmeus. We should, of course, expect Kinmy, but the rule is
not absolute even in the best time. Cf. Meisterh., Gr. pp. 56 seq.
6 a^ioxpeio', Kohler. But this is perfectly analogous to the other cases of et
written for e, and the accent should therefore not be changed. So in II. S72,
KoWvrdes must be read, not KoKAvreTes. In Aeiaiydpov, 553 (circa Eucl. ann.),
perhaps the earliest instance of this inaccuracy, we have et written for an Ionic e.
6 Lapidum de Dial. Alt. Testimonia, pp. 10, II.
7 The forms li/eiyi<ri for iveyitri, etc., frequent between 373 and 332 (v. Meisterh.,
p. 89), apparently on the analogy of Ion. ive'iny, show that the same confusion
could exist before a double consonant when, as before vowels, quantity could not
be affected by the interchange.