Obrazy na stronie

example of the comparatival suffix dispensing with its sibilant.

In treating pos of the Latin, and πpos of the Greek, as having a comparatival suffix, I have assumed that the of the ordinary comparatival suffix (ios, ior) has been lost, just as in the cases of αμφις, περισ-, επισ-, πριν, and pris- of pristinus, priscus, and prius as pronounced, I assumed the loss of the second vowel. To these may be added the Attic Tλewv for Teov, a form which at once accounts for the ordinary superlative πOTOS; and I have already appealed to the evidence of several parallel cases as magis. I have now on the other hand to back my theory by quoting, as others before me have done, clear instances of comparatives which have lost the i, and for this purpose I call as witnesses Tλeov by the side of πλe-cov, minor and minus (for minior), secus (=sequius) so often followed by a quam, plus or plous, which represents πλεον (πλεος) rather than πλειον, primores for primiores, and, if I am right in my theory as to this word, yepwv. Perhaps too, after what has been said of the tendency of prepositions to take comparatival suffixes, I may include with the adverb ενδ-ον from ev, as used in ενδον γεγραπται, infra, which on this theory would stand for evd-cov with an excrescent S.



But I have yet other examples of this loss in existing Greek words of the prepositional class: I mean Teρav (πeρην) and avτηv. Of the two suffixes employed in Sanskrit for the formation of comparatives, viz. tara (I should say ara) and iyān or iyāns, the former corresponds to what is seen in βελτερο-, the latter to what is seen in βελτ-ιον-, so that it may be assumed that the o of the latter form has superseded a long vowel. This is parallel to the case of pηTopes compared as to suffix with tutores, so that the av and ŋ of the Greek words just quoted has its explanation on the theory that here again we have comparatival suffixes. This argument is perhaps strengthened by the consideration

1 These are commonly regarded as cases of nouns, as is also avτ; but I am still inclined to adhere to the view in the text, because I do not see what part such case-endings would have to play.


that reparos, which is so closely connected with Teρā, has itself a suffix identical with that of παλαιος and γεραιος, which seem to have originated in comparatival forms. It may be noted that the Latin possesses a representative of the Greek Teρav in peren-die. But in dealing with πeρāv (περην) 'beyond,' I am brought to that second preposition Tapa, signifying ‘over,' which I hold to be wholly unconnected with the ordinary Tapa, signifying 'nearness.' The παρα over I find in many compounds, as παρα-πηδαω leap over, παρα-φημι talk over, παρα-βαινω transgress; but for full details I must refer to the essay itself, especially p. 113. It will be there seen that I hold this waρa to be decapitated from ὑπ-αρα, and so akin to ύπερ, and ultimately deduced from vπ of vπo, just as I hold the German equivalent ver to have superseded a fuller über, which is but a comparative of auf, as our own over is of up; and again the Latin per is but a shortened form of super. The prepositions which in the old Latin language had a final ad, but afterwards dropped this d, leaving the final vowel long, as extrad, intrad, suprad, infrad, ultrad, citrad, contrad, also claim our attention. Ritschl, in his recent essay on the final d, holds them to be ablatives, but he does not show how this case-ending is suited to the idea. I would suggest then that the suffix may have grown out of a preceding an, and so be a comparatival suffix like that in Tep-av, as here assumed. Ultrad, for example, may have originated in ultran, throwing out an excrescent d; or the letter n may have passed directly into a d, for the two letters are convertible, as in merced- (nom. merces) by the side of mercen-arius or mercennarius. This theory would account for the adjectives extran-eus and interan-eus. So much for disyllabic prepositions, for pros and pro and dis seem to have been disyllabic in origin.

But there still remain, especially in the Latin vocabulary, several monosyllabic forms of greater length, standing by simpler forms, as abs (or as of as-port-o), obs (or os of os-tendo for obs-tendo), subs (or sus of sus-tollo, sus-pendo, sus-cito, for subs-tollo, etc.), ex, trans, by the side of ab, ob, sub, ec (ecfodio, ec-fero), tran (trado, traicio, tram-es). The last of these,

tran, is in fact a corruption from pran (=πeρav), where the letterchange is, as I have elsewhere said, precisely parallel to what is seen in tranquillus, that is pranquillus, for planquillus, a double diminutive from plancus, which is itself a diminutive from planus. But what is the s which alone distinguishes the longer from the shorter forms first enumerated? In the case of ex, I think it just possible that the s may be excrescent, for there is a close affinity between an s and a guttural. Thus our word frog has a byform frox in A.-Saxon. But the g also interchanges with a sibilant, and then the sibilant might take an excrescent k-sound, so as to give us the A.-Saxon variety frosc. This theory receives support from the Greek coxaTos, which is but a derivative from ЄK. Similarly trans may possibly have an excrescent letter for its final, just as the Greek xnu appears in German as gans. And the same theory may account for es, that is evs, as compared with en. On the other side, the s in ex, trans, and as may also be due to the same cause as in the other little words, abs, obs, subs; for as regards these the outgrowth of an s is altogether inadmissible, and a theory which at once accounts for all the forms has the strongest claim to our support. My own belief then is that all six words contain in the s a comparatival suffix shortened from is. But if es (evs) represents an older and fuller ev-is, I arrive at the very form which was anticipated in the first part of this paper as that from which evɩ was corrupted.

The prepositions cis and uls have also a sibilant, which calls for explanation. I would suggest then with some confidence, that they are contractions of citis and ultis or ulis, two words which I regard as byforms of the citer and ulter, which have given origin to citra, ultra, etc.

But I have yet other words in which the comparatival suffix is has lost its vowel, as, first, oye 'late,' which may well have been compressed from a fuller omσ-e, so as to be akin, as indeed its meaning asserts, to oπio- and οπισθεν. So too ay back,' probably represents a fuller aπ-is, so as to be one with aπо, which on the preceding theory grew out of an older aπ-os. Again abundant evidence

to the same effect is supplied by modern languages. Thus the French gives us moins in place of the Latin minus, the Gothic mins or minz 'minus,' vairs 'pejus,' seips 'amplius,' another seips, 'tardius posterius,' suns 'statim' or 'ocius,' anaks 'subito' (taken from Grimm's D. G. 3, 589, 590). The Norse again has many examples of the vowel lost before the r of such comparatival adverbs, which in that language corresponds to the Gothic sibilant, as mior minus,' betr 'melius,' verr 'pejus,' heldr 'potius,' fyrr 'prius,' and nine others (ibid. 593). Grimm also treats as comparatives the Latin mox, which he is inclined to connect with ocius, and vix, for which however he finds no satisfactory origin; but here I may safely suggest poyis (see L. and S. under μoyos).

There yet remain a few prepositions which perhaps ought to be brought under the same head with those which have been examined above, as ante and poste from antid and postid. Possibly pōn-e too may have in its first syllable a byform of pōs (whence post), and the e may be the ghost of a second comparatival suffix, as standing for is; cf. magis, mage. The same may be true of the final in sin-e; and the case would be strengthened if we could rely on the French sans as representing a form of the old rustic language of Italy.

VIII-A PARTIAL ATTEMPT TO RECONCILE THE LAWS OF LATIN RHYTHM WITH THOSE OF MODERN LANGUAGES. By T. HEWITT KEY, Esq. SOME fifteen years ago (1855, May 25), I read before the Society a paper on Greek accentuation, in which incidentally I was led to say not a little on the principles of Latin accent. Again, in my Latin Grammar (see ed. of 1862, §§ 22-29; 835, 6; 1138, 9; 1199, 1391, 1404.1, 1465, with addendum of p. 456), I ventured to introduce more or less direct references to this subject, which in most treatises of the kind is wholly ignored. Two English scholars since that period have dealt with the subject of Greek accentuation,-Professor Chandler of Oxford, and less formally the late Public Orator of Cambridge; but the work of the former, however

elaborate, does not discuss the rationale of the theory, though he holds out some promise of treating this side of the question in a future volume. Mr. Clark's views, so far at least as the results are concerned, are stated in a paper entitled “On English Pronunciation of Greek," in the first number of the "Cambridge Journal of Philology." The latter writer through his many-sided scholarship and his well-known accomplishments is likely to conciliate favour; but after a careful perusal of his paper I only the more firmly adhere to the heresies which have long possessed me. The view for which he contends is that the rhythm of Greek and Latin poetry, when these languages were still spoken in their purity, was founded on principles essentially different from the modern accentual system,-the organs of speech and hearing in those happy days being able to distinguish between accent and quantity, so as to combine the due observation of both; but he seems to admit that he, like most of us, is now unable to understand how they could be so distinguished and combined.1 Like Mr. Pennington, he has studied the subject with all the advantages of direct intercourse with modern Greece; and, again, as was the case with Mr. Pennington, his own personal evidence is so far in my favour that he also distinctly states that the actual pronunciation of the modern Greek language fails to combine the two elements, accent alone now governing the rhythm of verse, and that accent falling as often on syllables with an epsilon and omicron as on those with an eta or omega. My own theory, it may be recollected, is, that while the long vowels 7 and w, etc., were originally introduced to mark quantity, and so indirectly served to determine the place of accent, yet, even within the so-called classical period, such long vowels often lost their quantity, but were unhappily retained in writing; and that then, and then only, it became necessary, for the sake of foreigners learning the language,


1 That is, so combined as to contribute to the rhythm of verse. I am assured on good authority that the late Professor Boeckh read the lines of Eschylus so as at once to observe the laws of quantity and the accents as marked; but this is not enough for the present purpose, unless it be also shown that his said observation of the accents tended to improve the rhythm in some slight degree at least. A Bohemian and a Scotchman have the habit of intoning, so to say; but this contributes nothing to the pleasing character of metre.

[merged small][ocr errors]


« PoprzedniaDalej »