Page images
PDF
EPUB

to e.

been intensified. Thus we have acer from the root ak, which is short in ăcies, &c.; but măcer is still short from mak, măcies; the causal of the same root macero1 is long, but lăcero, similarly formed, is short: perhaps in the case of măcero, the formation of which is denominative rather than causal, the a has been lengthened on the analogy of causals like plac-are from plăc-ere, which is formed quite regularly on the Sanskrit, and probably IndoEuropean principle. Sagax stands by sāgus and sāgire, păc-iscor by pac-s (pax, pāci-s), căueo by caui, plāga and πληγή are Graeco-Italian from plag shewn in ἐπλάγην: by fragilis we have suffragium, by ago, ambāges. From these and other examples which could be given it would be impossible to lay down any rule for Latin use in this scale when the a has been retained and not weakened If we turn to Greek we shall find more clearness. Thus we have from the root ád (Indo-European and Sanskrit SVAD, compare Latin sua (d)vis), adeîv, tada; from γλακ, λακεῖν, λέλακα; from λαθ, ἔλᾶθον, λέλαθα, in the older (Doric) form and in Doric perfects generally. Thus we see the a restricted regularly to the perfect; the presents being otherwise strengthened (ἁνδάνω, λάσκω, λανlávw, &c.). The long a, which sometimes appears in the present of these verbs, e. g. γκρᾶγ, κράζω, κεκράγα, is phonetic, not dynamic; κрá∞ = κрay-y-w. The Ionic 7, the weakened form of ā, is similarly used as the rule for the perfect, not the present; for cases like TTσow can generally be explained like κράζω: that is, πτήσσω = πτάκ-ψ-ω; compare the compound aor. kатÉTTтакоV. There are however exceptions to this rule; thus we find ny-vv-μι, perf. Téπηуа, where there is no vowel difference: in other cases, e.g. pnyvvu, where n is found in the present, we have a further increase in the perfect, as eppwya. On the whole then it seems allowable to regard ā and its dialectic equivalent ʼn as occupying the highest step in the a-scale.

1 Corssen however, I. 395, separates the two words.

CH. VI.

CH. VI.

Different methods of

distinguishing them in Greek; especially the employ

ment of the existing division of

A into a,

E, 0.

Then how did they represent the first step and keep it distinct from the second? Sometimes, as we have seen, by strengthening the verbal stem in different ways; either by reduplication, as ἵστημι for σι-στα-μι, or by nasalisation, as in à(v)d-áv-w, or by suffixes, as in λá(k)-0kw, ἅπτω, φαίνω for φαν-ψω, ἄγ-νν-μι, &c. But very frequently they employed a vowel-variation ready to hand, one originally phonetic only, but capable of being applied to distinguish different shades of meaning1; that division of the a sound, so often mentioned, into a, e, o, which will be fully described as soon as we come to phonetic change. Now o is a heavier sound than e; so that while e is employed for the present stem, the greater intensity of idea implied in the completed action can be expressed by o; as e.g. in πέρθω, πέπορθα; στρέφω, ἔστροφα, &c. But this more frequently is the mechanism employed in the formation of nominal bases. Thus by πép0-w we have πόρθ-ο in πτολίπορθος; by στρέφω, στροφή, and στρέφo-s; by ex-w, ox-o-s, and ox-n, and innumerable others. Indeed this may fairly be called the most important of all the methods of forming nouns in the language.

A slight difficulty arises here from the fact that the ascent from e to o is not always (indeed not generally) the entire process in the verb-formations on this method.

1 This faculty of language may be well illustrated from Prof. Curtius, note 21 on page 31 of his Essay on Comp. Philology and Classical Scholarship. He there says (Engl. trans.): "The distinction between èxóμela and exóμelov was surely at first purely phonic, but it subsequently got to be employed to separate the plural from the dual. And the Sanskrit termination of the first person dual vahé is most likely but a variation of the 1st pl. mahé; and scarcely any one would maintain that in the v there is really a significant mark of the dual relation (cf. vayam, plur."we"). Thus too I consider Tév0os as a by-form of πálos, one which the phonic tendency alone has brought forward. [See however, p. 163]. There was never any difference between e.g. ßévéos and ßálos, although a more refined feeling of the language introduced one between Tévoos and Tábos. This is in some degree also the case with the German "Ablaut [i. e. the change of a vowel to another vowel of a different class according to certain laws], more especially in its application to the formation of words. It can be shewn that the change of i, a, and u, in the verb trinken was there before, and that it arose from very different reasons than the difference in the meaning of Trank and Trunk."

2 See the full list in Leo Meyer, Vergleichende Grammatik, 1. 110, &c.

Besides the perfect stem in o and the present stem in e CH. VI. we commonly find another stem in a: thus by TÉTρоpа and τρέφω we have ἔτραφον; by ἔστροφα and στρέφω, Eστρáþην. Now this stem, as expressing the simple momentary action, ought undoubtedly to be expressed by the weakest vowel; and so we find it in the other vowelscales; e. g. from λιπ we have ἔ-λιπ-ον, λείπ-ω, λέλοιπα. Clearly we have this relation; as τέτροφα is to λέλοιπα, so is τρέφω to λείπω, and ἔτραφον to ἔλιπον. But a undoubtedly, as a rule, passes into e; and therefore we seem in this particular case to have a weakening and not a strengthening in the first step of the scale, though the second step is an increase of sound above the first. It may be that the "Sprachgefuhl" of the Greek was here for once at fault; and that the three separate forms being all to hand, e and o were taken, not unnaturally, upon the analogy of the e, ot, and ev, ov, of the other scales, in order to gain that distinctness which, as we have seen before, the Greeks prized above all other people1.

The Latin is not without traces of the same change. Though few and far between, compared with the abundance of the Greek, they are sufficient to shew that it inherited the same method as the sister language, though not the same power of developing it. Thus we find among the verbs the increase from ĕ to ŏ, in mõneo, the causal (compare Sanskrit mân-ayâ-mi) by me-min-i (weakened from me-men-i), men-(ti)s. If man-e-o belong (as has been already suggested) to the same root, we have here an example, I believe the only one in Latin, of the triple form in actual use, but with the distinction practically forgotten. Just like moneo is noceo, the causal of něc (in nex, něcis), Indo-European NAK. Passing to nouns

1 This difficulty is ignored by Schleicher in his Compendium, p. 62, where, in treating of the different intensifications of a, he makes e to o one of the first steps. To me it appears certainly a second step in the verbs mentioned above; and in nouns oxos stands to exw just as oroîxos (second step) is to στείχω (στιχ) and σπουδή to σπεύδω (σπυδ).

Traces of the same

method in

Latin.

CH. VI.

Advantage

of the different symbols for long vowels in the Greek.

Quantitative increase.

we have tog-a from √/těg; proc-u-s, "a wooer," by prěc-ari ; soc-iu-s by seq-ui and ad-sec-la, and others'.

We have seen above that ʼn has gained a place in several present stems, sometimes on phonetic grounds, in other cases perhaps by analogy. This gave the Greeks an advantage which they were not slow to use. Since w: ŋ :: o : e, another method of ascent in the a-scale was gained thereby. Thus from the root (F) pay, present py-vv-μ, they formed the perfect eppwya-in exact analogy, as has been already observed, with the Gothic lat, léta, lailôt. This w makes its way also into nounforms; thus from πτακ (πτήσσω) we get πτώκ-ς “the hare" (the by-form πTάσσw is perhaps a denominative verb formed from it): so also ἀρωγός stands by ἀρήγω, "to help" (root paк): and it occurs regularly in reduplicated nouns, as ȧy-wy-ó-s and dy-wy-n, ax-w-n, and (perhaps formed on analogy with these) even ¿d-wd-ǹ from the base EA. Sometimes the long vowel ē seems to be employed in the same way in Latin. Thus we have |fră(n)go, frēgi, ǎgo, ēgi, păciscor and pă(n)go, pēgi: together from cilium and oc-cul-o, we infer a root KAL to cover (found in c(a)l-am), from which we get cāligo and cēlare. Sătus has the vowel short, which is long in Saturnus, and apparently changed to è in semen. But with respect to the perfects the origin of the ē is not beyond dispute: it may be due to lost reduplication; whilst the other examples, beside being insufficient, shew no clear trace of method.

Short e is raised to long e, in Greek rarely, as μé-μŋλ-a (uéλw), and in such cases as rí-On-μ by the side of Tí-0eμev: but more commonly in Latin, where we find sed-es. (sed-eo), teg-ula (těg-o), lex (lēg-is), and col-lēg-a by leg-o, sērus by series; and among verbs ed-i (ed-o). Here the root-vowel was in every case originally ă, as is shewn by the other languages: but it had changed into è in GraecoItalian times.

1 See Schleicher, p. 87.

ā:

Similarly short is lengthened, again in Greek more rarely than in Latin: od-wda is lengthened from Vod; στώμυλος stands by στόμα: and the δίδωμι class of verbs corresponds to the Tínμ class. In Latin uox (uōc-is) stands to uoc-o in the same relation as lex to lego: perhaps however the ō here may represent a: we have √ in a Sanskrit. We find this change in the verbs often enough, as fodio, fōdi, moueo, mōui; but here the same remark must be made as about the ē: persōna is an increase of sound upon son-u-s, and sōp-i-o above sop-or. In these two last cases the o comes from original u: the IndoEuropean roots are SVAN and SVAP respectively, but this does not affect the principle of the increase of the o. As a general rule it would seem that the Latin language preferred the simpler expedient of a merely quantitative increase of sound (ă to ā, ĕ to ē, Ŏ to ō): while the subtler genius of the Greek chose rather a qualitative strengthening, like that of e to o1.

CH. VI.

I-scale.

We may now pass to the i-scale, which may claim the (ii) The merit of being the most perfect and least corrupted exponent of the scheme, both in Greek and Latin. Thus in

one."

1 Mr Roby (Latin Grammar, Preface, p. xxiii) writes: "I have not followed Schleicher and others in the treatment of Latin vocalisation according to what for brevity I may call Sanskrit principles. This method applied to Latin seems to me to fail both in basis and result. Corssen's elaborate treatment of vowel-intensification in the first volume of his new edition is not more satisfactory: and on this point I can refer to Curtius (Studien, 1. 2, p. 294), who, commenting on Corssen's sanguine view of the result of his medley collection of long vowels in root-syllables suffixes and endings, points out that vowel-intensification is after all only a name for the fact that we often meet with a long vowel when we expect a short I am not greatly concerned to defend Corssen in this matter: and with Mr Roby's pithy statement of the result of Corssen's work I agree: the cause, however, is clearly the extreme disorganization of the Latin vocalism, which prevents us from discovering clear traces of a method undoubtedly found in closely cognate languages, and therefore certainly once operative in Latin also, although in a manner which was early obscured, and is now hardly traceable. I may say that Curtius, in the criticism quoted above, seems to me to be referring to one part of Corssen's work only, the "Einlautige Steigerung," not to vowel-intensification in general and even here specially to one point only, the long a in erām, rās, &c. Corssen compares (inter alia) (o)nola, &c. in Greek: and refers the long vowel to this principle. This explanation is certainly not satisfactory: but I do not know a better. That of Schleicher, that eram (for example)=es-aya-m, seems to rest on no good analogy.

« PreviousContinue »