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

274

J. McKEEN LEWIS,

for the continued representation of ov, 6 by the simple 0 when these-
sounds had by supposition become = u.1 Blass infers this change in
the o by its frequent transliteration with Latin u. It is most uncom-
mon to find it confounded with ov in inscriptions of the Roman
period, and the vowel has at this day in Greece the same sound which,
so far as we can gather, it had at the time of Pericles and at that of
Homer. We find in contemporary Greek and Latin inscriptions a
continual interchange of Latin o, u, and Greek ov; if Latin u is
found for o, so is late Greek ov (u), used to represent Latin o {e.g.
<f>ov\\u<\o<; = folliculus). The Latin -us for -os doubtless helped the
confusion of the two vowels; the change is often attributable to the
working of a special analogy (as in paenitla = <f>aivo\^) which may
equally change any other vowel to u (as crapula = KpaiiraXr), purpura
= Trop<f>vpa). The examples collected by Saalfeld2 make clear the
Roman preference for u, whether to stand for an original ov, a, e, o,
u>, v, or even to separate consonants whose collocation was unwonted
in Latin." But where analogy or some more obscure causes of per-
version are not at work, the Latin o is the recognized equivalent of
Greek o. The Latin u is transliterated by Greek o from a different
cause; this is discussed with much good sense by Dittenberger
{Hermes, VI. p. 281), who observes that no similarity of sound is
indicated by the transliteration, as the Greeks of the Roman period,
having no short 11 represented in their alphabet, were obliged to-
resort either to o or ov, the one in violation of vowel-quality, the
other of vowel-quantity. And, in any case, the vagaries of ignorant
lapicides at a time when the instincts of language were in a univer-
sal decline, and the changed and changing relations of the Roman
phonetic system to that of the Greeks were an added source of con-
fusion and misrepresentation, are not of such authority as to justify
us in supposing a temporary divergence from the normal pronuncia-
tion of a vowel, standing in no relation to the general progress of

1 Cf. Blass, Ausspr. des Gr., p. 31, etc. Nothing can be inferred from such rare
orthographical caprices as Adftuvovs, nIkiuvovs.

2 Lautgesetze d. Gr. Lthnw'drter im Latein., pp. 74 seqq.

3 In the case of iirioTo\T\ the process of analogy betrays itself, epistola
becoming epistula. The history of "EKa!3rj in Latin is similar, and instances of the
kind can be easily multiplied by any one familiar with early Latin literature and
epigraphy.