CH. VII. Loss of letters in the final syllable. certainly in inscriptions, as in the grand line on the tomb of the young L. Cornelius Scipio', Quoiei uită defecit, non honos, honore. In locatives we find i in domi2, and others in Plautus: mihi, &c. could have the final vowel short or long down to the Augustan age; compare nisi and quasi. Imperative dissyllables early shortened the last vowel, as was but natural, to the brevity of command; rogă, iubě, maně, &c. are frequent in Plautus: also other (not imperative) forms, as dari, dedi, because of their shortness and frequency: o also in the first person is short in eo and volo-dissyllables again, and found in Plautus, but in the Silver, age the tendency had affected longer verbs as well. We now come lastly to the absolute loss of the vowel, either when it stands actually last, or when it is followed only by a weakly-sounded consonant, that is, practically, by none at all—the result, like the loss of quantity already considered, of the tendency in Latin to throw back the accent as far as possible from the end of the word, subject to the rule of the length of the penultima. First under this head comes the loss of original o, or later u, in the nominatives, such as ager (os), puer(os), &c., a numerous class; as famul for famul (os) used by Lucretius after Ennius, Ossa dedit terrae proin ac famul infimus esset. 3 It is not easy to determine in these cases whether the vowel or the s went first: we should rather have expected thes but there are no traces of the vowel surviving: on the contrary, s is found alone in words like Campans*, but this seems almost unique. But the vowel i has certainly fallen out and left the s in nouns like Arpina (ti)s, where the t after the loss of the vowel would seem to have quattuor:" unless we follow Fleckeisen and transpose Sol and Luna, for 1 Mommsen, Corpus, n. 34. 3 III. 1035. 4 Plaut. Trin. 545, quoted by Corssen, II. 591. assimilated itself to the s; so in men (ti)s, fron (di)s, and very many others: in orb(i)s, &c. where the preceding consonant is not a dental, it keeps its place unchanged. Where a liquid precedes, the liquid maintains its ground, and the s is lost, e.g. uigil(is), uomer(is), pedester(is), and very many others; where however the accent falling on the antepenultima sometimes drove out the e of the next syllable, and produced the other form, as pedestris. The same principle seems to have produced out of uelis (i.e. si uelis) the conjunctive uel'. I followed by no consonant fell away regularly in neuter nominatives, such as animal(i), lacunar (i), cochlear(i), piper(i), lac(ti): though Corssen mentions forms in -e, as lacunare, existing side by side with these, as was quite natural; he quotes sale (i.e. sal) from Ennius. Similarly in many adverbs the i has been lost, as tot(i), ut(i), post(i), &c. For tot and quot Corssen compares the Sanskrit tati and kati, and calls ti a "demonstrative particle:" but tati seems to be rather an old locative form produced by adding i to the pronominal base tat. The i was lost in very old times from the verbal terminations, as regis(i), regit(i), regont(i); also from regebām(i). E was lost in imperatives of the third or old conjugation, just as a and were shortened in the first and second: : e.g. in dic(e), fac(e): but the full forms are common in Plautus: this loss therefore was a late one. Many little words in common use have lost their final e, as neu(e), originally ne uelis, hic(e), &c.; nec (i.e. nequĩ or neque), qui-n(e), si-n(e), &c. The fuller forms, hice, hae-ce, the nom. plur. hisce, &c. are sometimes still to be seen in Plautus. This e, which was in these cases weakened from i, must have been so slight a sound, and so little inconvenient at the end of a word, that it is lost less frequently than we might have expected. 1 Corssen, II. 60, ed. 1. Dr Wagner (Academy, July 11, 1870) prefers Prof. Key's explanation that uel uol, an obsolete imperative. It seems to me that in si-ue and ne-ue the first part of the word suits better with the explanation given in the text. CH. VII. CH. VIII. Superiority of the Greek over the Latin thus far manifested. CHAPTER VIII. CONSONANTAL CHANGE. IN the last chapter I have mentioned, I think, the most important variations of the Greek and of the Latin vowelsystem from that of their common Graeco-Italian ancestors, and from the simple vocalism of the earliest historic period of our race. We have seen in these variations the strength of the Greek and the weakness of the Italian. We have seen how the Greek could adhere in the main to the simple distinction of scales in the original vowel-system, and yet could avail itself with remarkable success of any expansion of that system. The Greek vocalism shews the greatest observance of rule, combined with the greatest individuality: and thus coincides with the highest development of Greek character to a surprising degree: and since the consonantal system shews the same principle, though developed in a less degree and in different ways, we are justified in believing that the character of a nation can be impressed on its language, so far as that language is the result of pure internal development, and has not been compounded of many foreign elements. The Italian, on the contrary, both confounded that distinction of the three main vowels which is essential for the clear expression of distinct radical ideas, and also subjected itself to a rule which kept ever increasing in stringency-the tendency to uniform monotonous weakening. So far then we have seen the Greek at its best, the Latin at its worst. Now we shall see the better side of the Latin compared strength of the Latin with the Greek, shewn in its greater tenacity of consonantal sound. No doubt the Latin not unfrequently substituted a weaker for a stronger consonant, as well as the Greater conso nants. Greek or indeed any other language: every language has its own peculiar weakenings of this kind; they are the most obvious marks of distinction between one language and another. But the greater strength of the Latin consonants is shewn in their comparative freedom from assimilation, which in many Greek verbs obscures the radical form. Thus in opaσow we have the same root and the same suffix (ya) as in the Latin farc-io: but the k of the root is lost in Greek from the assimilating effect of the y, which in Latin was simply resolved into the cognate vowel, and exercised no power over the stronger consonant: indeed the k is hardly recoverable in Greek because it has regularly sunk to y, as in e-ppay-ov; just as it sank to bairg in Gothic, that language which of all the Indo-European' family comes nearest to the Greek in the richness of its vowel-system: the original k is to be discovered in the less spiritual Latin and Lithuanian. Generally speaking, however, the original form is recoverable in Greek from some of the tenses which are formed directly from the root: the Greeks felt too keenly the necessity of clearness to suffer the consonants to be absolutely obliterated; they are the necessary framework of language, the body which is needed for the soul; yet the soul may be vigorous though many bodily members are weak or even lost. It is curious too how the innate Greek love of symmetry is recognisable even in the weakenings of its consonants: they are nearly always regular, not often isolated: there is a system to be found in almost all of them: while the Latin looks uneven in the midst of its regularity; its loss especially of consonants in groups is arbitrary, and not reducible to rule and even its less corrupted vowel forms have a more irregular appearance than those of the Greek. Nothing can look more regular than ἕζομαι, στίζω, σχίζω, μúsw, &c.; but this regularity leaves us quite uncertain whether the root ends in a guttural or a dental; while there is no such uncertainty about the very unsymmetrical forms which correspond to them in Latin, sedeo, stinguo, CH. VIII. CH. VIII. scindo, mugio. This will appear more clearly when we have seen the difference in the changes of the two languages. Change of hards to softs-not very comтоп. Change of K toy and to g. I. SUBSTITUTION. 1. Change of hard (unaspirated) letters to soft. This change is the simplest process of substitution. It is not very common in either language, and less so in Latin than in Greek: in both languages it is sporadic only, never affecting the whole, even of any class of words. Thus & passes into y in ȧphyw from apk, which is unἀρήγω altered in aρkios. Perhaps the change may arise from the softening influence of the two vowels and of p. 'I have already mentioned the Greek eppayov by the side of farcio; but the σo of opάoow shews that √ppaê must have been the original form; for we should have had φράζω from a φραγ. Indeed the Latin has commonly preserved for us the original letter which the Greek has weakened. Thus in Greek we have πnyvvut, and even in Latin paciscor and pagus: but pac-iscor shews that the oldest form of this common root, to fix—whence to build, or to covenant-was PAK, and not PAG, as we should have rather supposed from the frequency of the g; and τáσσαλos (i.e. Tak-ya-λo-s) tells the same tale. Compare μíoyw and misceo1. In Latin gloria is from √/klu, the Indo-European KRU; the first step is the noun clouos (compare λéƑos), which with the suffix ya becomes the secondary noun clouosia; and this, by the loss of u and the change of s to r, is cloria; after which the l is probably responsible for the g, and the change therefore is rather one of assimilation2. But there is simple substitution in viginti by the side of 1 A full list of all the gutturals thus changed in Greek is given in the Gr. Et. 485-487. 2 See Krit. Beitr. 53. |