This combination is rare. In Latin, as Leo Meyer observes', it remains unaltered, as in aurea, doceam, &c. 4. €+0=oto=w: as e-οφλον a = ώφλον, φιλέω = φιλώ, The last change-to ev-is very restricted in its area, apparently to the southern coasts of Asia Minor and the adjoining islands. It does not seem to be Aeolic: at least the examples in Alcaeus and Sappho are uncertain, and it is not found in inscriptions: it is certainly not Boeotian. τεοὺς, ἑοῦς, ἵππω occur in Corinna Neither does it belong to the severe Doric; either ∞ (more rarely ov) is found or a variation peculiar to this form of Doric, by which e passes into : e.g. in μαγίομες, ἐμίο, ἐπαινίω. It seems to be Sicilian, but only in the pronouns: it is very common in Theokritus, who however may have got it from Cos. In Ionic (e.g. in Herodotus) it is very common: but it is not found in Attic. In the second change-as in the third of (2), and again in the second of (6)-we see that when oo come together the second sound sinks to v, especially in the Attic Greek. The second vowel has assimilated the first to itself, and then, unable to maintain itself beside its self-created rival, it passes into the weaker v. The third modification, on the contrary, shews an immediate weakening of the second vowel, with no change of the first. Of course it must not be inferred that the other change to ou was unknown in other dialects than the Attic: it is not uncommon even in Doric: thus Καλλικράτους, ἐπιμελουμévns, &c. are Spartan3. CH. VII. CH. VII. 6. 0 +€ = 0 + 0 = w: as ὑπνόεν Recurrence vowel. = ὑπνῶν, =ο+υ=ου: as λόετρον πλοῦτρον. The change of o+e to w belongs to the severer Doric ; the second is the regular weakening, and is found even in Doric as well as w: but when e begins a word it is common, as wλapos2. When o + a meet in two words, we have ávýp in Attic, výp in Doric and Ionic: o+at becomes w, as ὥπολος. η, し When the same vowel occurs twice, the natural result of the same is clearly that the two should coalesce in one long single sound: as is the case in κέρατα, κέραα, κέρα; γένεσε, yévee, yévn; aidóos, aidos. But just as when similar diphthongs had become identical by assimilation, the second vowel was weakened, especially in the Attic, so is it also here. Thus the combination e+e becomes in Doric and Aeolic n, and o to becomes ω: λέγε(μ)εν is λέγην, ἵππο(sy)o is ππw in both—at least in the severer form of the Doric; in the softer Doric e and ov are universal: and there is some irregularity in the Lesbian also. In Attica the two vowels were perhaps in each case sounded as a diphthong, long enough to allow the second vowel to become weakened, to and v respectively: λéуeɩ, iππον. So also when the second vowel-sound is the mere prolongation of the first caused by the loss of a dental, or the dental-spirant s. Thus e is the result in Attic (sometimes in the other dialects) of e+σ, e.g. eiuμí for eo-μɩ (Lesbian eup, severe Doric nuí, softer Doric eiui): of e+v, e.g. εἷς for ἑνς (severe Doric is): of e + ντ, as τυφθείς: of e + νθ, as πείσομαι for πενθ-σομαι. Again, o+8=ου, as πούς: o+dov, ο ντ ου O+VT =ov in doús. These changes might be explained on precisely the same principle as those above; that is, that ἐσμί became ἐεμι, and then εἰμί; the middle form could be supported by the Doric nμí and perhaps by the Aeolic forms to be mentioned below. But it is perhaps simpler, and more in accordance with the views already set forth, 2 Theok. 1. 135. 3 Ahrens, II. 318. 1 Ar. Lys. 143. し to explain this et as the natural result of the voice dwelling upon e1; namely, that when we sound (a), as has been already pointed out, an i-glide arises naturally after it: and similarly a u-glide after (o). These forms are Doric as well as Attic; the softer Doric however sometimes allows the second sound to drop altogether: as ἐρίσδεν for ἐρίσδεμεν, not ἐρίσδην; ἀμπέλος for ἀμπέλους", a form, by the way, which was retained by the Argives and Cretans; and many others. In all these the sense of the original length of the last syllable was kept up by not letting the accent be thrown back. The Lesbian sometimes employs a peculiar change of its own, by which appears instead of v, as o+VT=oi in popéotoi, &c.3. This and parallel forms, πpéποισα, Μοῖσα, &c., are very common in the fragments of Alcaeus and Sappho, and are also found in Pindar. It certainly cannot be assumed in these that the is weakened from the ordinary u, if that u be nothing but a glide: such glides are too short in their pronunciation to be capable of corruption. The therefore must be explained as a glide which sprang up immediately after the o in Aeolic: and so it is parallel to the forms θνάισκω, τάλαις, &c., which have been explained above. The Boeotian in general has o in these cases, like the severe Doric; both when two vowels are contracted and when a consonant is lost. The tendency of this dialect to single vowel sounds is very remarkable: the only exception is the use of e instead of the ordinary ; the tendency toward the particular vowel reminds us both of the Latin, and also of the extraordinary fondness for that sound shewn in modern Greek. In this love for the monophthong the hard Doric stands next to the Boeotian, and the soft Doric next. The Attic has much more of the diphthong, however the second element was sounded. The Lesbian seems to affect glides. Generally the tendency to become し 1 This is Mr Roby's explanation, preface, p. lxvii. 2 Theok. IV. 8, v. 109. 4 At page 227. 3 Theok. XXVIII. 11. CH. VII. CH. VII. monophthongs was greatest in e and ou, which is what we should expect on physiological grounds: next to these come at and av. 3. Latin Diphthongs. These, as I have already said, were once as numerous as those of the old Greeks. But at a very early age they dwindled into simple sounds. Their history has been fully traced by Corssen: the account here given will be little more than an epitome of his results. I will take the six diphthongs, ai, ei, oi, au, eu, ou in order. It would appear from inscriptions' that ai was sometimes retained, sometimes written as ae, from the third century B. C. to the latest times: but it was doubtless in every case pronounced as ae, whilst the older method of spelling was of use to distinguish the genitive singular2 of the first declension, for example, from the nominative plural, which was written with ae. This rule however was probably never universal: we find tabelai datai (nom. pl.) in the Epistola de Bacchanalibus3, B. C. 186, &c. Perhaps the old ai may have been retained longer in the root-part of the word than in suffixes or prepositions: thus we find aides and aidilis in the well-known epitaph on L. Scipio, son of Barbatus, whilst on the Columna Rostrata of Duilius1 we find prae-sens and praeda (i.e. prae-hid-a— the same base as pre-he(n)d-o). In the letter however of the consuls forbidding the Bacchic rites (quoted above), we have aedes, and aiquom. Clearly no fixed rule for the spelling can be given. Corssen however observes that in legal notices from the time of Gracchus to Caesar ae 1 Ausspr. 1. 680, &c. 2 The old termination of this case seems to have been ayas; whence āīs. Then if the s were retained, the a absorbed the i, as in familiās. Corssen however gives examplos of aes, all being female proper names (1.184). See Schleicher, Comp. p. 558. 3 Mommsen, Corpus Inscriptionum, Vol. 1. No. 196. 4 Ib. No. 195. is found universally instead of ai: after which time ai began to reappear: one result, we may suppose, of that restoration of archaic forms of which we have a specimen in Claudius' attempt to remodel the alphabet'. That the pronunciation of the diphthong was not unlike the German ä or (a) is shewn by the fact that it began by degrees to be written as e. It was so pronounced by countrymen2 in the time of Lucilius, as we find from Varro3 who quotes the forms Cecilius and pretor: while educated men preserved something of the double sound. In inscriptions after the Christian era e appears with increasing frequency and an inscription dating 242 A. D. presents the three words Aimilius, Sabinae and Furie. It is quite certain that at that time no distinction would be made between ai and ae: indeed the former had been replaced in the Latin alphabet only by an affectation of archaic forms; it appears at the same date, or even earlier, upon inscriptions instead of an e, which is etymologically correct; but it is possible that even then ae may have been rather nearer to a than e was. Mr Roby thinks that the sound may have been nearer to the (a) in "hat," but pronounced long. It has been already mentioned that both ae and ě correspond to open e in modern Italian. This would agree with Mr Roby's view: but the closeness of the correspondence would point to a sound for ae even nearer to open e, that of (ě) pronounced long, as in Scotch "ell:" (ĕ) short is heard in English "ell.” Not uncommonly ai was also weakened to . This takes place especially in case-endings, as uiis from uia-is, &c.; also in prae when in composition, as priua-tus primus, pri-die, priuignus (for prai-ui-genus), &c.; and generally in compound words, as occido from caedo, iniquos from aequos, inquiro from quaero. The middle form is sometimes retained, e.g. exaestumo, pertaesus regularly. CH. VII. |