CH. VII. Correspondence of Latin with modern Italian. sound. The well-known passage of Plato'-oi pèv apxaιóτατοι ἱμέραν τὴν ἡμέραν ἐκάλουν, οἱ δὲ ἡμέραν, οἱ δέ νῦν nuépav-probably means that Plato had heard some Greeks use the form iμépa: any one who is conversant with Plato's references to oi apxaîou will not attach any deeper meaning to the phrase than a polite refusal to investigate a question any further: here at least it is perfectly certain that the original vowel was a and not i: and e was of course used in ordinary Greek before ʼn was established: but it undoubtedly would appear to be the natural conclusion from this passage, that the first vowel of the word just before Plato's time wavered between € and ɩ, and that ʼn afterwards expressed the sound more exactly. On the other hand, we do not know what Greeks they were who said iμépa: we have no other evidence of the sound having occurred anywhere. Mr Geldart rightly rejects any evidence drawn from Cratinus' sheep which said (in our texts) ẞn, ẞn: because in Cratinus' own spelling they must have said Bé, Bé. On the whole I do not think that any slight variations of the n to in classical times, are sufficient to overbalance the distinct connection of the sound with a. Semitic transcriptions of n by i need not prove anything more than the absence of the ŋsound in those languages. Plutarch, transcribing Latin words, gives κάρηρε, μαιώρης, ῥῆγας, &c.: and therefore I should think that it was probably either è, or (ě) pronounced long. A further piece of evidence with respect to e and o is gathered by Mr Roby from transliteration : o represents both o and u in Latin, e represents both e and i in Latin which would seem to place o and e in Attic nearer to u and i than o and e are in Latin. The sounds of e and o in Latin are deduced by Prof. Munro from the correspondence of modern Italian. In that language close e and o represent ē and ō in Latin (and also and u, which is natural enough, see Table of Sounds in Chap. IV.): and open e and o represent ě, ae, 1 Crat. 418 c, quoted above. and ŏ, au. Prof. Munro would give the open sound to the short vowels, and the close sound to the long vowels, distinguishing however between vowels naturally long and long by position. Mr Roby1 makes some just deductions from the force of this rule: but, allowing for mistakes and variations, I think it on the whole the best we can get. The analogy of ae and au is curious, because ae was a single sound, and au was not. But they were each open sounds, though in a different way. The rule is in accordance with that mentioned concerning English i and u that the open sound is regularly short, and the close sound long. There is however a discrepancy between the theories here given of Greek and Latin pronunciation of e and o, which at first sight seems fatal to their truth: n corresponds nearly to ě: and e to ē. But it must be remembered that we are speaking solely of the quality of the sounds, not of their quantity: and in each quality it is possible to have both long and short sounds. It ought however to be stated that when the sounds of one language had to be represented in the other, ĕ and ŏ were regularly represented by e and o, and ē and ō by ŋ and w; in Latin also the quantity of the Greek vowel was generally maintained: the more obvious difference of quantity in such cases must have caused the subtler difference of quality to be ignored. Words which are not borrowed, but which correspond in the two languages, agree in length in far the greatest number of cases, but not always: e.g. Onp=fera, πаρ = iecur3. I do not assign much weight to the statement of the grammarians (quoted by Roby) that “ŏ primis labris exprimitur, ō intra palatum sonat:" which, if received, would contradict the theory given above: because they were separated by too long a time to make the statement trustworthy; and it is unsatisfactory in itself: the share of the palate in producing the more open sound of o is much less perceptible, η CH. VII. 1 Grammar, p. lxv. 2 At p. 89. 3 Roby, p. 73. CH. VII. and hence this o might not unnaturally be thought to be sounded by the lips alone: in reality both organs must be employed in each sound. Mr Roby does not appear to draw any distinction between open and close sounds of e and o in Latin. I think the evidence of Italian usage, though not convincing, yet sufficiently strong to justify me in arranging the Latin sounds accordingly. Subjoined is a possible table of the single vowel-sounds of both languages. The idea is borrowed from Mr Roby1: whose table should be compared with this. I differ from him sometimes, but never without the feeling that he is more likely to be right than I am. 2. Greek Diphthongs. CH. VII. 2. AI=al, ει, οι ; ευ, ου. The substitution of a, e, o for original a led of course to a corresponding increase in the number of diphthongs, in Latin originally as well as in Greek: but the Latin had suffered almost the whole of them to fall into disuse before | AU=av, the classical period of its literature. In Greece the number of the symbols for the diphthongs was still further increased in classical times by the introduction of ʼn and w. Thus the language possessed in the room of the original ai, au, āi, āu no less than ten symbols, ai, ei, oi, αν, αν, ου, ηι, ωι, ην, ων, besides the rather rare v. The diphthong iu is only found in the Teutonic family. There seems no reason to doubt that these were all at first what their name implies, double sounds; in which the transition from the first to the second sound was distinctly audible. It is probable from the nature of the case that two sounds should be sounded as two, and probable also from their origin. When it was not a dynamic modification of a simple vowel intensifying the idea which that vowel conveyed, a diphthong arose, either from the coalition of two distinct vowels by the loss of an intermediate consonant, e.g. λéye(u)ev: or secondly, from a spirant being resolved into a vowel in accordance with laws of consonantal substitution to be mentioned in their place, e.g. ἀνδρεῖος from ανδρε-ψo-s, λόγοιο from Moyo-syo (where the σ has left no trace of itself): or thirdly, from the prolongation of the original vowel-sound to compensate for the loss of a following consonant; thus when v was lost in povoa the first form must have been μούσα, as is shewn by the Doric μώσα, while the new vowel was weakened to v in Attic μovoa, to in Aeolic μοῖσα. These new diphthongs often remained double sounds in Lesbian later than any other form of Greek Diphthongs were originally "double sounds." CH. VII. History of the change of dissimilar diphthongs in Greek. し speech: thus we find Çota, Axiλλétos', &c. where the ‹ is It is difficult, perhaps impossible, to trace the history of the Greek diphthongs, and fix the time when they ceased to be double sounds—each sound presumably the same as when it occurred separately; neither can we do more than guess with more or less of probability at the new single sound of each. There can be little doubt that the corruption of the diphthong must have been little later in time than the causes which produced it. Two vowels following immediately upon each other are commonly troublesome to pronounce; the most simple (and probably the oldest) combinations of language shew us vowel and consonant occurring alternately': when a consonant fell out and two vowels met, there must have been at once a tendency to subordinate one vowel to the other so that one of the two should become a "glide," i. e. a sound too short to be called a perfect vowel, because it never 1 Theok. XXIX. 5 and 34. See Ahrens, 1. 105. : 2 See Leo Meyer, Vergl. Gram. 1. 285, where numerous examples of Greek and Latin diphthongs are given, from which I have borrowed largely in this section. |