Establishment and Extension of the Law of Thurneysen and Havet 9781463221850

Citation preview

Establishment and Extension of the Law of Thurneysen and Havet

A n a l e c t a Gorgiana

347 Series Editor George Anton Kiraz

Analecta Gorgiana is a collection of long essays and

short

monographs which are consistently cited by modern scholars but previously difficult to find because of their original appearance in obscure publications. Carefully selected by a team of scholars based on their relevance to modern scholarship, these essays can now be fully utili2ed by scholars and proudly owned by libraries.

Establishment and Extension of the Law of Thurneysen and Havet

Lionel Horton-Smith

gorgias press 2009

Gorgias Press LLC, 180 Centennial Ave., Piscataway, NJ, 08854, USA www.gorgiaspress.com Copyright © 2009 by Gorgias Press LLC Originally published in All rights reserved under International and Pan-American Copyright Conventions. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise without the prior written permission of Gorgias Press LLC. 2009

1

ISBN 978-1-60724-601-5

ISSN 1935-6854

Extract from The A^merican Journal of Philology 16 (1895)

Printed in the LTnited States of America

III.—ESTABLISHMENT AND EXTENSION OF THE L A W OF THURNEYSEN AND HAVET. I. §i. Introduction. The Law of Thurneysen and Havet, that Prim. Lat. ov- (preserving Idg. B) became av-,1 is one which has not met with the entire approval of all philologists, hence it will be of service to the cause of Italic philology to bring together all the evidence, and examine it in detail. In the present essay I hope not merely to establish this Law, but also to extend it. Just as Pr. Lat. dv(preserving Idg. d) became av-, so too it seems that Pr. Lat. ov(preserving Idg. o) became av-, and in the same way also the Pr. Lat. diphthong ou- (preserving Idg. o) became the diphthong au on its way to the later u : d. §2. The Literature of the subject, together with a discussion of von Planta's arguments for regarding the Law as Prim. Italic. The literature concerning this Latin 2 phonetic change up to the present time is as follows: G. Lowe, Prodrottius Corporis Glossariorum Latinorum (1876), p. 348; F. de Saussure, Syst. 1 It must be carefully observed that this Law deals only with Pr. Lat. Sv-, preserving Idg. 0, and does not affect Lat. ov-, from earlier Lat. ev-, preserving Idg. e [cf. Brugmann, Gr. I, §65, p. 52 (Engl, ed.), and especially Lindsay, T h e Latin Language, ch. I V , §19]. 2 Von Planta, Gramm. der Osk.-Umbr. Spr.,p. 115, has endeavoured to prove that this law is not confined to Latin, but is to be considered Prim. Italic, thinking that he sees traces of it in Umbrian and Oscan. That the Law under consideration should have operated in Italic, as well as in Latin, would in no way surprise us, but for the present, pending, further evidence from these dialects than is yet available, I prefer to regard the change in question as confined to Latin, for the following reasons:

(1) In the Umbro-Samn. dialects Idg. 5 seems, on the whole, to have had a closer pronunciation than in Latin (cf. Brugmann, Gr. I, §§81, 89 ; see also von Planta, ib., §§42-44, pp. 108-115, and §46, p. 116), and hence would have been less likely in these dialects to undergo a change to a before u, as the change in this combination was due to a very open pronunciation of the 5.

LAW

OF THURNEYSEN

HAVET.

445

Prim, des V o y e l l e s (1879), pp. 104 sqq.; R. Thurneysen

in K u h n ' s

Zeitschrift, X X V I I I

AND

(1887), pp. 1 5 4 - 6 2 (dated

December, 1884) ; L. Havet\n

November

and

M6m. de la Soc. de Lingu., vol. V I ,

Part I (1885), pp. 1 7 - 2 1 ; K. Brugmann,

Grundr. I (1886), § 8 1 ;

J.

(2) Not a single certain instance of dv- (preserving Idg. 8) changed to avcan be produced from Oscan or Umbrian, as von Planta himself (1. c.) admits. (3) Umbr.-Osc. bSv- is not so well referred to Idg. *aeu- as to Idg. *idu- (v. infra, p. 459 and p. 460, note I), so that if our Law were to be extended to Italic, we should expect Umbr.-Osc. *bav- from earlier bdv- (preserving Idg. 0). Lat. 60s, (gen.) bovis, is certainly not a genuine Latin word (v. infra, pp. 458 sqq.), and is indeed most probably due to the influence of Samnitic Oscan (v. infra, pp. 459, 460). (4) Lat. dvis ' s h e e p ' from Idg. *dui- (: Gk. *bpi-, Goth, avi-) ' sheep,' beside Umbr.-Osc.-Pelign. *dvi-, is best explained as due to the influence of these other Italic dialects (v. infra, p. 461 sqq.), aided possibly by the influence of Rustic Latin (v. infra, p. 462, note 4). If we regard our Law as not confined to Latin, but common to the other Italic dialects as well, then, unless we allow that Rustic Latin (in this case the only Italic source left to which Lat. Svis may be ascribed) alone resisted a Law to which all the neighbouring dialects were bowing, we must refer Italic, i. e. (so far as our records enable us to see) Lat.-Umbr.-Osc.-Pelign., ovi- 'sheep,' as also Umbr.-Osc.*^»- [No. (3) supra], to a Greek origin, and that of no very early date ; for had they been borrowed very early, we might have expected them to have become sufficiently 'naturalised' to undergo the change which befell genuine Pr. Lat. dv- (preserving Idg. d), a change which did not commence in Latin till the third century B. C. at the earliest (v. infra, pp. 456,457). [Compare the fate of Greek i/.ai(f )a, which, adopted by the Romans in the period of the Tarquins (Pliny, Nat. Hist. X V I), became subject to the Latin change of e to 0 before /, whence *dlaiva, which became *dleiva and finally oliva (Lindsay, T h e Lat. Lang., ch. I V , §10, p. 228).] It seems, however, unlikely that Umbrian or Pelignian should have had much to do with Greek, hence it is better to regard at least the Umbr.Pelign., and therefore most probably also the Oscan, ovi- as genuine Italic representatives of Idg. *oui-, and Lat. ovis as due to their influence. I would hardly accept von Planta's suggested Idg. *e%i- as the original form whence Lat.-Umbr.-Osc.-Pelign. dvi- are all to be derived, nor do I put faith in his alternative suggestion that, supposing Idg. *o%i-s to be the original form, the preservation of the i f f - may be due to those forms in which i followed (cf. Skr. avya avyai); v. infra, p. 462, note 2. (5) Umbr.-Marruc. avi- ' b i r d ' can quite easily be explained as from Idg. *iui [(: Gk. *afi-) v. infra, pp. 454, 455], whereas Lat. dvom and Vulg. Lat. *dvom ' e g g ' (on which v. infra, pp. 455, 456) prove clearly that Lat. avis ' b i r d ' must come from earlier Latin *dvis from Idg. *oui-s (: Gk. *oft-) ' bird.' (6) In every instance cited, where a consonant standing between the 8 and v is said not to have hindered the influence of the v, the ' dazwischenstehende Consonant' is a liquid, namely and it seems more probable that the al- of alv- in these instances is merely the Latin and Italic representation of Idg. / (v. infra, §5).

446

AMERICAN

JOURNAL

OF

PHILOLOGY.

E. King and C. Cookson, Sounds and Inflexions in Greek and Latin (1888), ch. V, pp. 85-9, ch. IX, pp. 187-9; H. Schweizer-Sidler, Gramm. der Lat. Spr., Part V (1888), §11, 7); F. Stolz, Lat. Gr.2 (1890), §io, p. 258, §15, Rem., p. 264; H. D. Darbishire in The Classical Review, vol. IV, p. 273 b (June, 1890), and in The Transactions of the Cambridge Philolog. Soc., vol. I l l , Part IV (1892), p. 189; R. von Planta, Gramm. der Osk.-Umbr. Dial., vol. I (1892), §§45, 50, 79 (and cf. §96); R. S. Conway in Idg. Forsch., vol. IV (1894), pp. 216, 217 (dated Oct. 1893); F. Stolz in the Historische Gramm. der Lat. Spr., vol. I, Part I (1894), §§101 d), 105 c), 148 f ) ; W. M. Lindsay, The Latin Language (1894), ch. IV, §§17, 19, 41; and P. Giles, Short Manual of Comparative Philology (1895), §180, p. 141. §3. Examples of the change of Pr. Lat. dv- (preserving Idg. o) to dv-; the date of this change; and an exact statement of the Law for this change. It will be well for the purposes of the present essay to quote and examine all the examples cited of this Latin change of dv(7) T h e word-group vacare, etc., in L a t i n , v a k a z e , etc., in U m b r i a n , the only instance cited from the I t a l i c dialects of the change of o (preserving I d g . d) to a after v (that is, u), probably does not come from earlier vdc- at all, but shews original a (v. infra, §6), and even if this w e r e not so, e v e n if this w o r d group w e r e to come from an earlier vdc-, p r e s e r v i n g I d g . 0, this could hardly be of much use to prove that Italic ov- (preserving I d g . 0) b e c a m e dv- in the I t a l i c dialects. (8) T h e explanation of L a t . octavos, Osc. U h t a v i s is at present a matter of so much doubt that it is hardly safe to base any conclusions on this form (v. infra, §7). N o r does von Planta's suggested derivation of U m b r . k l a v l a f k l a v l e s , L a t . cldva clavola seem at all a c c e p t a b l e (v. infra, §7). (9) O s c a n and U m b r i a n shew no trace of the c h a n g e of Prim. Ital. tautosyllabic du (from I d g . tautosyllabic du) to tautosyllabic au (v. B r u g m a n n , G r . I , §81), such as I hope to establish for L a t i n (v. infra, §8). A s it is more than p r o b a b l e that the c h a n g e of the Prim. L a t . diphthong ou (from I d g . tautosyllabic du) to the diphthong au was contemporaneous with the change of Pr. L a t . dv- (preserving I d g . d) to av-, and as both are obviously due to the same cause, viz. very open pronunciation of the 0, the fact that O s c a n and U m b r i a n shew no traces of the former is evidence also against the occurrence of the latter in these dialects. I n conclusion it must be pointed out that even if fresh e v i d e n c e were to establish the L a w for I t a l i c (as w e l l as for Latin), nevertheless e v e n so the c h a n g e cannot b e regarded as Prim. Italic (as von Planta suggests), but can only have arisen later in the different individual developements of Italic, for, as w i l l be seen b e l o w (pp. 456 sqq., and §8 ad fin. and §q), the L a w did not b e g i n to operate in L a t i n itself till the third century B . C . at the earliest.

LAW

OF THURNEYSEN

( p r e s e r v i n g I d g . o) to dv-,1

AND

HAVET.

447

w h i c h , as S t o l z r i g h t l y s a y s in the

H i s t . G r a m m . d e r L a t . S p r . , v o l . I, p a r t I, § 1 0 1 d~), m a n i f e s t l y m u s t h a v e t a k e n p l a c e in c o n s e q u e n c e of v e r y o p e n p r o n u n c i a t i o n of the o in this c o m b i n a t i o n : — cdvos

f r o m P r i m . L a t i n cdvos [ w h i c h h a s b e e n p r o v e d also f o r

V u l g a r L a t i n at 2 0 1 B . C., v . infra, p. 4 5 7 , and is p r e s e r v e d to us a l s o in the c o u n t r y term covum,

coum

or cohum,2

n. ' t h e h o l l o w in

the p l o u g h ' ( w r o n g l y e x p l a i n e d in P a u l , e x F e s t . , p. 3 9 , 4 ' I o r u m , q u o t e m o b u r i s c u m i u g o colligatur, a c o h i b e n d o d i c t u m ' ) , a ' v e s t i g i u m r u r i s ' a d o p t e d b y E n n i u s 3 with t h e m e a n i n g ' t h e h o l l o w 1 1 would at the outset expressly state that, in endeavouring to establish the truth of the ' L a w of Thurneysen and Havet,' viz. that Prim. Lat. dv (preserving Idg. 0) became Lat. dv, I do not imply disagreement from the excellent exposition of the Laws of Ablaut, so ably drawn up by Bartholomae with the aid of the fresh evidence which he has adduced from Armenian (v. Bartholomae in Bezz. Beitr., vol. XVII, pp. 91 sqq.). That »appears beside a in the strong grade (Hochstufen) of the a- ( = Bartholomae's a"-) series is, I think, beyond all doubt, e. g. Gk. by-jioQ : Gk. ay-u Lat. &g-o, Lat. Scris : Gk. a/cpof, Gk. ndafiog : Gk. Kaarufi, Gk. tpoirau : Osc. b a i t e i s Lat. baetere (the latter of which is probably borrowed from Oscan, cf. Brugmann, Gr. I, §432, Rem. 1). But it should be observed that, the change which I am discussing does not in the least affect our views on this Ablaut-scale of Bartholomae ; for the change of Prim. Lat. dv- (preserving Idg. o) to Lat. dv- is, so to speak, quite a ' private' Latin change, and one, moreover, which, as we shall see later, did not begin to operate in Latin before the third century B. C. at the earliest, and, even if common to the other Italic dialects, at any rate could not have been so early as Prim. Ital. (v. supra, p. 444, note 2 ad fin.). Some scholars may perhaps consider that one or two of the instances which I cite here as examples of Latin dv- from Prim. Lat. dv- (preserving Idg. d) should more correctly be referred to Idg. a«, shewing the strong grade a of the a- ( = Bartholomae's a'-) series. Thus, e. g., they may possibly consider that Lat. Idvere (3d conjugation) : Gk. (whence AOEGJ) = Gk. ay-o Lat. &g-o : Gk. hy-finr — Gk. Kdarcjp : Gk. ndafiog, etc., and thus shews Idg. a; but, on the other hand, if we compare Lat. Idvere with Gk. *XoFe (whence Ade), imperf. of Gk. *"AoFa (whence Aow), we must regard the a of Idvere as from Idg. d, which seems to me the better explanation of the two (v. infra, p. 451). In most of the other instances cited the evidence is very strongly in favour of my view [e. g. the formation of faveo (infra, p. 451) and Uvea (infra, p. 451) certainly points to *fdved * (from *XoFa), \oia>, \oéa> from *\oFéa>. T h i s view is supported also by Thurneysen, ib., p. 1 5 6 ; K i n g and Cookson, ib., ch. V , p. 8 6 ; Havet, ib., p. 1 8 ; Schweizer-Sidler, ib., § 1 1 (7), p. 1 2 ; Stolz, L a t . Gr. 2 , §10, p. 258, and in Hist. G r a m m . d. Lat. Spr., § 1 0 1 d), p. 1 1 4 ; Lindsay, ib., ch. I V , § 1 9 ; Giles, ib., §180, p. 1 4 1 . 1 fâveô from earlier L a t . %Jôveô, which is either (a) from earlier *fôveiô 'cause to be,' the causative of V bheu- ; cf. the phrase di faveant ut. . . (so Thurneysen, ib., p. 1 5 4 ; Giles, ib., §180, p. 1 4 1 ; cf. also K i n g and Cookson, ib., ch. I X , p. 188) ; or (F) = 'donner de la rapidité,' being based on an adjective *fovos — G k . Boos from *6oFos (so Havet, ib., p. 1 8 ; cf. also K i n g and Cookson, ib., ch. V , p. 86), or perhaps direct from earlier *fôveiô (causative), which, together with L a t . */ôvos, G k . 6oàs Ôéa (from *6éFcù), would come from I d g . hj dheu- ' run ' (: S k r . dhâvati 'runs,' cf. F i c k , V e r g l . Wôrterb. I*, p. 465, s. v . dhëvo-~). T h e view here taken, that *fôveô was the earlier form of Lat. fâveô, is supported (but without discussion) also b y SchweizerSidler, ib., § 1 1 (7), p. 1 2 . âveô from earlier L a t . *ôveô, from I d g . *dy.-êid, being a causative of the same t y p e as Lat. mônëô, G k . 0péu>. T h i s is the view 1 In unaccented syllables this -âv- became -ù-, e. g. ablûô from *db!âvô (: lâvô from *livo) from *dblâvô, ëlûâcrus from *ilâvëcrus (: lâvâcrum from *lôvâcrum ; cf. Gk. Xosrpév) from *ilôvàcrus ; v. Stolz, Lat. Gr. ! , §14 B, 4, p. 262, and Lindsay, ib., ch. IV, §41.

452

AMERICAN

JOURNAL

OF

PHILOLOGY.

taken also by Darbishire in the Classical Review, vol. I V , p. 273 b (June, i8yo). W i t h aved from earlier *dved we may compare faveo from earlier *f&veo, just discussed. Favdnius from earlier *Fovonius ' the warming wind, the west w i n d ' from Idg. *dhd$h~- from Idg. dhe%h- ' t o b u r n ' (whence also L a t . f d v e d from Prim. Lat. *feveov) and favilla ' h o t cinders, glowing ashes' from earlier *fovilla, a diminutive of an ancient */dva or *fovos from Idg. *dhoo,h«o-s : Skr. ni-daghd-s ' heat, summer,' Goth, dags, O . H . G . tag, O.Icel. dagr ' d a y , ' Lith. daga-s dagcl 'harvest' (properly 'hot time'), from Idg. Gf. *dhd&ho-s, from Idg. v' dhe^h- ' burn.' 2 In support of this derivation of Favdnius and favilla see especially Thurneysen, ib., p. 159, and Lindsay, ib., ch. I V , §§19, 144 (on favilla); Havet, ib., p. 19, and K i n g and Cookson, ib., ch. I X , p. 189 (on Favdnius and favilla'). Favi from earlier Fdvi, cf. Paul. Festi 62 T h . d. P.: 'Fovi, qui nunc Favi appellantur' ('Fovii .. . Fabii,' Miiller). avilla' lamb,' dimin. of Latin *avis' sheep,' from Prim. Latin *ovis ' s h e e p ' (Lat. dvis ' s h e e p ' itself being due to dialectal influence, v. infra, p. 461 sqq.) from Idg. *dui-s [: G k . 01? from * o f « 'sheep,' and Goth, avi- ' s h e e p ' seen in Goth, avistr 'sheep-fold' (from earlier Goth. *avi-vistr, v. Brugmann, Gr. I, §643)]. In support 1 T h e Latin verb fdved (from Idg. ti/dheiji- ' to burn') has been derived by Brugmann, Gr. I I (Eng. ed., vol. IV), §794 (and cf. §790), from Idg. *dholh-eio (or rather *dhoihi,-e'io): Skr. dahaya-ti ' lets burn,' but if this derivation were correct, we should have expected Lat. fdved to become *faveo, hence Lat. foveo is preferably to be derived (with Thurneysen, ib., p. 159; SchweizerSidler, ib„ §13, 6), p. 15 ; K i n g and Cookson, ib., ch. I X , p. 189, and Havet, ib., p. ig) from earlier Lat. * f l v e o , shewing the vocalism of tened [beside which there was also an Idg. ^ttin-eid, from which comes Skr. tanaya-ti, cf.

Brugmann, Gr. II ( = IV), §794, p. 1150, Germ, ed.], teped, sedeo, fended, tnedlor, vereor, vegeo (v. esp. Havet, 1. c.). In the same way Lat. mdveo is to be derived from earlier Lat. *meveo : Gk. a-fiev-aaadai (Schweizer-Sidler, ib., §13, 6), p. 15 ; Havet, ib., p. 17 ; K i n g and Cookson, ib., ch. I X , p. 186, and Stolz, Lat. Gr.», §15, p. 264) from Idg. \/meu- (cf. Fick, Vergl. WOrterb. I 4 , pp. 103, 286, 511). Lastly, Lat. voveo, whether derived ultimately from ^¡eu- (v. Brugmann, Gr. I, §428 c), and Gr. Gr. 2 , §35, p. 56) or from a y u e j i - (v. Osthoff in Morph. Unters., vol. V , p. 82, note 1), must come from earlier Lat. *veveo, preserving the /-grade of the original root. 1 For these words cf. Brugmann, Gr. I, §376 ; Fick, Vergl. Worterb. I 4 , p. 74. With *fdva (in the text above) beside *fev- Havet, ib., p. 19, compares Lat.

idg