Modern French déchirer L’origine de afr. eschirer, la source de afr. deschirer déchirer, est contestée. Selon Wartburg, l’étymon avancé par Gamillscheg, abfr. *skîran ‘nettoyer, gratter’ = anor. skîra, ags. scîran ‘nettoyer, purifier’, efrison schiren ‘nettoyer, gratter’, est peu convenable du point de vue sémantique. D’autre part, Gamillscheg affirme que celui de Wartburg, abfr. *skerjan ‘diviser’, est insuffisant à l’égard de la phonétique. On propose ici que l’explication la plus probable d’eschirer est l’emprunt au vieux-haut-allemand ou au vieux-bas-francique de ahd. skeran ‘déchirer’= abfr. *skeran, dont la voyelle radicale était i dans la moitié des formes du présent de l’indicatif. An etymology concerning which there has been disagreement is that of OF eschirer, deschirer NF déchirer ‘tear, rend’. That deschirer was produced through the replacement by the Romance prefix desof esin eschirer is obvious and uncontested. However, the origin of the earlier Old French form has been the subject of dispute, though all apparently agree that it was borrowed from Germanic. Diez 1869: 291 proposed OHG skerran ‘scratch’ as the etymon. Meyer-Lübke 1935: 659 rejected skerran, which he labelled Low Franconian, on the ground that its open e would not have become OF i but offered nothing to take its place. Gamillscheg 1969: 302 and 1970: 336 expresses agreement with Meyer-Lübke concerning skerran and in its stead suggests LFrk. *skîran ‘clean, scratch off’, reconstructed on the basis of ON skîra, OE scîran ‘clean, purify’ and EFris. schiren, which means ‘scratch off’ as well as ‘clean’. However, as Wartburg 1966: 107-08 states, although *skîran would be phonologically acceptable, it seems unsatisfactory semantically. Gamillscheg apparently assumes that, whichever of its two meanings was original, a verb which in Germanic could mean both ‘clean’ and ‘scratch’ became restricted to ‘clean’ in Old Norse and Old English but retained both meanings in Low Franconian and much later in Frisian. However, there is no Germanic evidence to support this view. Although the root is attested in every branch of Germanic, none of the older German dialects contains a form which suggests a connection between *skîrand scratching. All of the words containing it which are attested prior to the contemporary Frisian form, including the verb OFris. skëria ‘clean, purify’, show only a consistent association with meanings such as ‘clear’, ‘bright’, ‘clean’ and ‘pure’, one which had been inherited from Proto- Indo-European 1 . Although no verbs occur in Continental West Germanic, the adjectives OS skîr(i), MLG skîr and MHG skîr ‘clean, clear’ are found there. Although Wartburg’s criticism of Gamillscheg’s explanation appears justified, his own proposal seems clearly unacceptable. It is LFrk. *skerian ‘divide’, based on the occurrence in Old High German of skerian (skerjan) ‘decide, determine, assign, allot’, a spelling which Wartburg 1966: 108 holds to have represented two originally distinct but semantically similar verbs, one of which contained a close e and 1 Pokorny 1959: 918; Kluge/ Seebold 2002: 802. 129 Modern French déchirer the other an e which was open. The etymon is proposed to have corresponded to the verb possessing open e. In addition to being a quite uncertain reconstruction, such a verb would, if it indeed existed, be unsuitable both phonologically and morphologically. Old High German janverbs containing e in the root syllable typically possessed a close e, which had arisen as the result of the umlaut of a e toward the end of the eighth century. Prior to the occurrence of umlaut, e was extremely rare under such circumstances, since before i or j of the following syllable Gmc. e had become i in Proto-Germanic 2 . In reconstructing *sk rjan Wartburg thus must assume that it had been derived during the interval between the origin of separate Germanic dialects and the separation of Low Franconian and Old High German. Although such a form was possible, it was much less likely than not in the absence of any evidence offered in its support. On the other hand, it seems certain that *skerjan would have occurred on at least some occasions with a close e, for skerjan is attested as rhyming with nerjan ( *narjan) in Old High German and in its earlier Low Franconian form of *skarjan is held by both Wartburg 1966: 108, 100 and Gamillscheg 1970: 261 to have produced OF escharir ‘assure, affirm’. It was for this reason that Wartburg found it necessary to hold that two different verbs were involved, a view which is supported only by the moderately wide range of meaning of OHG skerjan. Why Wartburg 1966: 107 went to such lengths to reconstruct *sk rjan is unclear, since he explicitly agreed with Meyer-Lübke and Gamillscheg concerning skerran. Although one would expect some clarification as to why the of skerran was unacceptable but that of *sk rjan was not, none is provided. Gamillscheg 1969: 302 therefore is apparently correct in rejecting *sk rjan as phonologically inadequate. He could have added that it is unsatisfactory morphologically as well, since he was certainly aware that the minority of janverbs which did not enter the French fourth conjugation in -ire became first-conjugation verbs with infinitive in -iare 3 . Since he dealt with *skerjan in an entry in his dictionary, he presumably felt that his phonological objection was sufficient and that further discussion could thus be omitted for the sake of brevity. The controversy concerning *skîran and *skerjan was unnecessary. Old High German strong verbs possessing e of the root in the infinitive contained i instead of e in half the forms of the present indicative (and neither i nor e in the past) 4 . Therefore an Old French form produced by the borrowing of such a verb was as likely to have contained original i as . Although skerran, the third-class strong verb originally proposed by Diez, might thus be considered minimally acceptable as the source of eschirer, a more plausible etymon is provided by the fourth-class verb OHG skeran ‘shear, cut’ 5 . In addition to the semantic superiority of ‘shear’ to ‘scratch’, skeran could have produced OF eschirer in a loan from either Old High 2 Braune/ Eggers 1987: 26-28, 31. 3 Gamillscheg 1970: 403. 4 Braune/ Eggers 1987: 278-81. 5 Braune/ Eggers 1987: 278, 279. 130 Paul W. Brosman German or Low Franconian through adaptation of its i as close e in time to have undergone the French change of é i following a palatal. On the other hand, since the e subject to such a change had to be free as well as close, skerran could have yielded eschirer only in a late loan from Old High German after Gmc. i had begun to be adapted as OF i, a development which was also possible in the case of skeran. It thus appears that the most likely explanation of OF eschirer and deschirer is borrowing of OHG skeran or its Low Franconian cognate. Since those two forms would have been phonologically identical, it is at present not possible to choose between them with confidence. New Orleans Paul W. Brosman Bibliography Braune, W./ Eggers, H. 14 1987: Althochdeutsche Grammatik, Tübingen Diez, F. 3 1869: Etymologisches Wörterbuch der romanischen Sprachen, vol. 2, Bonn Gamillscheg, E. 2 1969: Etymologisches Wörterbuch der französischen Sprache, Heidelberg Gamillscheg, E. 2 1970: Romania Germanica, Berlin Kluge, F./ Seebold, E. 24 2002: Etymologisches Wörterbuch der deutschen Sprache, Berlin/ New York Meyer-Lübke, W. 3 1935: Romanisches etymologisches Wörterbuch, Heidelberg Pokorny, J. 1959: Indogermanisches etymologisches Wörterbuch, vol. 1, Bern Wartburg, W. von 1966: Französisches etymologisches Wörterbuch, vol. 17, Basel