Obrázky na stránke

one-it is a matter of indifference what the vulgate readings may be. The habit of giving such readings is a part of the old theory of a textus receptus, handed down from one editor to another and changed here and there by the comparison of new manuscripts or the talent of an emendator. The logic of the doctrine set forth in the preface is that the textus receptus is to be discarded; that a printed edition has no authority as evidence save that of the manuscripts upon which it is based; that it a given printed edition preserves readings of a MS not now accessible it ought to be cited by itself, and its evidence as a representative ought to be brought into relation with the two types of text that can be shown to have existed in the tenth century. The point of my criticism, then, is that the preface embodies the modern theory of determining the history of a text as a historical problem, while the commentary is not free from the old leaven of a textus receptus. To mark a reading “vulg." is not the citation of evidence: it is rather a bushel under which all sorts of things may be hidden.

Apart from the theory involved, this notation has practical disadvantages. The readings of a manuscript like L are worth knowing even when they are palpably wrong, for they may contain a hint of the truth. They should stand out clearly and not be left to inference. A few examples are here given to show how easily one may draw a wrong inference from the commentary as it is arranged. In IV 1538, L and G give the aor. åtetekuńparto. Seaton cites this fact and adds, “ÅreTeKuapovto vulg." A grammarian might be desirous of learning what authority there is, if any, for the imperfect. The fact that ought to have been stated is that the imperfect is a reading of Stephanus. On the same page, IV 1564 "Ardida vulg." should be "Ardida codd.” This is, to be sure, the reading of the printed editions up to Wellauer's time, but the thing worth knowing is not what has been printed but that the true reading is preserved as a varia lectio in the scholia as against all the manuscripts. A fuller statement of the evidence would have been instructive in IV 324 where the genuine reading rests upon the testimony of L’, a varia lectio in the scholia and a note of Stephanus Byzantius, s. V. Kavalkoi, as against L and G.

The important contributions of Rzach to our knowledge of Apollonius Rhodius (Grammatische Studien zu Apollonius Rhodius, Wien, 1878: Wiener Studien, 1881) have been valued and used by Seaton. Rzach proposes to read in IV 618, instead of μετ' ανδράσι κεκλήισται of the MSS μετ' ανδράσιν εκλήισται. Τhis is in the interest of uniformity. A perfect without the k of the reduplication is found in IV 267, 990, 1202, and is in each case the only possible metrical form. The fourth and remaining passage is the one in question, and the slight change proposed here would leave ékhńcotai as the one form used by Apollonius. Seaton's adherence to the manuscript reading as against Rzach seems justifiable, since kekhúcotai is a familiar form and there is here no greater compelling power than the law of unformity. Sometimes, however, Seaton has been too conservative. We read on p. 5 of the preface: "Rzachius inter alia koupat pro te kópai (I 811) et Apeos pro "Apews (II 404) scribenda esse judicavit, recte, ut opinor; neutrum tamen horum contra codices mutare ausus sum.” And yet "Apeos has been taken into the text and koupa. ought to have been. The case of veós, IV 208, for which Rzach proposes to write veós is somewhat more difficult because the final syllable of the word stands in the arsis and must be long. In defence of such lengthening Rzach cites Odyssey X 172 and Argon. I 289, where a syllable with a vowel naturally short is lengthened under the accent in the arsis of the fourth foot. The strength of Rzach's contention against veás, 'Apews, kópai, is not that they are isolated forms but that they are contrary to the law of epic usage. They are Attic, not epic. The same holds true of the dative pl. alot, which is nowhere allowed to appear in the Oxford edition, and of Bapeia, IV 1339, which long stood in the printed texts on very slender authority and which Wellauer rightly interpreted as Bapely. The same objection obtains against apopav, I 372, which Seaton has adopted on the testimony of L 16. The epic genitive πρώρης (II 556) might conceivably have πρωραν Or πρώρην, not πρώραν, as its accusative. Ifa dissyllabic word is to be retained a popnu is the only tolerable form, and that is the form adopted by Lehrs. But there are signs which point to a trisyllabic word as the desideratum. L and G have ήδε κατά πρώραν έσω αλός όσσάτιόν περ. This is a faulty verse, metrically.

The question then is, where is the fault? Brunck, without knowing the reading of L 16, pronounced čow corrupt and proposed ciow. But this conjecture builds upon the unepic form ap pay. Now assuming that cow is sound and that -pav is to be interpreted as short according to the law of epic speech, we reach the conclusion that the difficulty is with the first part of the pápay of the MSS. The emendation of Bergk, TT poetpav, accepted by Merkel, satisfies the conditions of the problem.' The soundness of this reasoning rests upon the two facts, that čow has the weight of authority and that apopav is for the epic speech an incredible form. Then the reading of L 16 is to be interpreted not as a good tradition but as a conjecture by some anonymous scholar who anticipated Brunck.

In matters of orthography, the following points may be noted. In obedience to evidence from various sources and in keeping with the best usage of the present day, Avýokw and Opbouw appear with iota subscript, the former without any manuscript authority in Apollonius, and the latter with L's testimony in III 957, IV 42, 603. The derivative noun Opwouós, has w in L, II 823, although not in III 199. Seaton has preferred Opwouós, although the other form is known to the grammarians and is found in Ven. A of the Iliad as well as in Laurentianus. The scrupulousness of L in these lesser points is one of the characteristic features of the manuscript. It has 8xń IV 289. (Seaton dyn), just as it stands alone in giving trávy (so Seaton) in 7 of the 13 cases of its occur

rence in the poem. The general evidence for iota subscript in Meunvorw is substantially the same as that for n and w in the two verbs above given, (see Fleckeisen's Jahrb. 1865, 245 ff.) but Seaton has not introduced the form in II 1140, the only passage where the word occurs.

The adoption of ómitteuw rather than ómineuw is in the face of a strong array of testimony. In II 406, III 1137, L and G agree in oπιπτεύω; in IV 469 L has oπηπεύω with correction oπιπεύω ; IV 799, L has ómunteúw. has, then, in two instances preserved the form órieuw, and stands alone in this save for two Vatican MSS. That this is the true orthography is clear from epic usage and the tradition of lexicographers. Since Bekker ómiTTEÚw has been banished from Homer, and that, too, upon evidence. The epic compound Tapdevotinns is in point. Paley and Rzach, in Hesiod's Works, 29 and 806, editómieuw, following codex Laur. XXXII 16 of Hesiod. Apollonius Sophistes, Photius, Suidas and Hesychius give omiteúw. Accordingly Kinkel in Lycophron's Alexandra 45 gives órieuw as against the MSS. The article in Liddell and Scott's Lexicon s. v. Ómetteuw needs revising as to orthography, for not one of the examples there cited has a firm foundation.

In conclusion I will mention one more matter of editorial detail, which may serve incidentally to justify and render intelligible Merkel's robust faith in L. This manuscript has a poßaðńs, IV 283; G, and presumably the other manuscripts have a poßadús. Merkel adopted posadís, Seaton poßabús. If the former is correct, we have an adjective in -us carried over to the class in -rs upon becoming a compound word. This is the only occurrence, to my knowledge, of this particular word, but the principle is a wellestablished one. kyxıBadís, IV 1572, is as old as Homer. Medaußadéos, IV 516, is to be referred to a nom. Medaußaðńs, found in Aesch. Prom. 219. poßadís is the opposite of Strabo's tpooßpaxńs. Besides, Apollonius has nodvdapońs, II 912, as Homeric form, and nepıdapońs, I 152, 195, a form peculiar to himself. Todókns is a familiar epic example of the same formation. In later times theBatńs, tolußaons, άμετροβαθής, ισοβαθής, αβαθής occur; but no compound with the end

The form a posadís is therefore, in keeping with the habit of the language and is a significant token of a good manuscript, whereas tipoßadús is easily understood as a blunder. The question has been decided in principle in the text of Aeschylus. The form Medaußabús, Prom. 219, found its way into the earlier printed texts from inferior manuscripts, but has long since been banished and forgotten. It is safe to say that a poßaons will eventually stand in the Oxford text of Apollonius and find its way thence into Liddell and Scott's Lexicon. HAMILTON College,


ing -us.

Lexique Etymologique des termes les plus usuels du Breton

Moderne. Par Victor HENRY. Rennes, Plihon et Hervé,

1900. Victor Henry's Lexique Etymologique du Breton Moderne (published as fascicle Ill of the Bibliothèque Bretonne Armoricaine, Rennes, 1900) is a book that well deserves the attention of Celtic as well as of English scholars. The author gives us in concise form a clear view of what so far has been done by various scholars towards elucidating the etymological connections of the most usual terms of modern Breton and we seem to be safe in following his guidance, as he exercises great caution and generally puts the reader on his guard, whenever the connection would seem either not to be well established or altogether doubtful. I have noticed only a few instances where the apodictic statement of fact does not seem to be in accord with the author's usual prudence: Under darn "piece' Henry confidently pronounces upon English ‘darn = to mend stockings' as a loan from Welsh darn 'piece,' while Rhỹs (in Murray's NED) considers the idea as absolutely inadmissible. Under ler 'leather' we are told that the corresponding Germanic words, English leather and German ‘leder' are loans from Celtic, while Kluge tells us that the Celtic words are generally considered as loans from the Norse. Under houarn 'iron' we learn that Germanic *eisarn (whence English iron and German eisen) is a direct loan from ancient Celtic, while Kluge admits this only for ON. jarn (from Olr. iarn). Under gwalc'ha 'to sate' Henry brings together Latin volgus with English folk, German Volk, while Kluge pronounces upon the connection as doubtful, it being very questionable whether the Germanic words are conformable to a base *quelgos, *quoigos, nor does the latter mention any connection between the Germanic words and Olr. folc (according to Henry, from Celtic *wolg-o), which connection would seem possible only under the supposition that the former are loans from the latter. Under houc'h 'pig' Henry is confident that English hog is a loan from Welsh (Cornish) hoch, but the idea is rejected on phonetic grounds by Rhys in Murray's NED. Nor do I think that Germanists will take kindly to the proposition, advanced under oaled 'hearth that OE. éled 'fire' is a loan from Celtic *ägileita, or that OE. swin like Breton souin is from Lat. suinus. Ludu ‘ashes' Henry brings together (though doubtingly) with German lodern' to which he assigns the meaning of 'smouldering under the ashes.' I always thought the German word was rather expressive of a blazing up of the fire. As in the instances given the author seems to have deviated from his usual course of prudent caution, he also occasionally presents views now rather antiquated. So under gwell 'better' Greek Boúlopai is quoted as representative of the wwel with which it has nothing to do according to the opinion now prevalent. Hirin, W. eirinen, Olr. airne 'sloe' which



Henry brings from Celtic *arinio- and compares with Skr. arani ‘wooden drill for producing fire' is now with Zimmer considered as cognate of Goth. akran, OE. æcern 'fruit.' Under oad 'age' Henry still brings Olr. des from Celtic *aiwestu cognate with alf-cw following a former suggestion of Stokes who now with Thurneysen posits a Celtic *ait-tu cognate with de-aitão bai and Latin utor (from *oitor).

Under skant "scale of fish' we are rather surprised to see Henry consider English skin as sprung from a true OE. scinn, while it is a loan from Icel. skinn. 'And, surely, the Celtic skant-o is not so isolated as Henry would have us believe. There can not be any doubt about OIcel. skinn being directly related to Breton skant. In fact, the correspondence between them is, as Zupitza points out, as close as it can be (see Zupitza, die germ. Gutturale, p. 156) and there may be a connection with kenn 'skin' Olr, ceinn, Olcel. hinna, which connection is admitted by Henry himself under kenn. I wonder why under koan 'supper,' from Latin cena, Henry does not mention the Irish loan from the same source, cene; see O'Mulconry's Glossary 427; cf. also cen ibid. 217 (cen mo mair. i. cen a cena, mair uita) and Todd Lect. V 55 cãe leis ic a fur. Under klân 'buttock' we miss reference to Olcel. hlaun, under kavel'cradle' (from Low Lat. cavellum) to OE. cawel 'basket' from the same

In regard to ant 'trench'= W. nant 'valley,' I wish to draw attention to C. G. L. V 339, I anes uallis = Corpus Glossary (ed. Hessels) A 570 which seems rather to stand for [n]an[ + ]es ualles than ancrae uallis as Goetz would have it (Thes. Gloss. Emendat., p. 68a); also antea uallis (C. G. L. II 566, 30) seems rather to favor an antes than ancrae cf. nante ualle in the Endlicher Glossary. Under talm, OJr. tailm 'sling' w. telm snare' mention might have been made of cogn. Olcel. pialme (pialfe) 'snare' (Noreen Altisl. Gr. $196, note 2) with which is evidently connected the OE. þelma glossing tendiculum in the Aldhelm-gloss printed in Zts. f. d. A. vol. IX. Worthy of attention seems to me Henry's suggestion that English crumpet is a fashioning of Celtic *cramm-poeth, whence W. cramm-wyth, Breton crampoez 'pasta cocta.'' The word must have been taken over already in Anglo-Saxon times, for Ahd. Gl. II 325, I we read placente fiunt ex farina et simila et melle uel ferro (=farre?). Saxonice dicuntur cron pech (= cronpeth?) with which Steinmeyer, 1. l. compares cronphetas (= cronpethas ?) ex farina, simila, melle in Cod. S. Galli 299, p. 280. Steinmeyer expresses, because of the latter passage, his disbelief in the genuineness of an Anglo-Saxon cronpech (cronpeth?), but granted that cronphetas is Latin, there is nothing to hinder us from supposing this Latin word to be a coinage from Celtic-English crompeth.

Of the greatest interest to the English student are, of course, the Breton loans from Old English and Modern English. So

puzé 'bitch' is conjecturally traced back to a loan from OE. bicce.

« PredošláPokračovať »