(4) feri porod Duvau; so all, except that Conway suggested the possibility of potrod, since the top of the r runs across the vertical as in a ligature. (5) made (for madet) mire cie or mi recie, perhaps madent recte Duvau; made mi recie or made mire, cie Lindsay LL; made (or madet) mi recie Lindsay Hdb.; made mirecie von Planta; made mire cie or maden isecie or madent regie Conway; made mi recte Fröhner; made mire cie Ernout PP; madent recte Diehl, Ernout RT, Lommatzsch. (6) misc sane Duval; so all. (7) asom fero Duval; so all. 2 The variants may be quickly disposed of. Coenalia 'dinner preparations' has an unetymological oe which could be due only to confusion with Greek Kovós, and such a confusion is hardly thinkable at the date of this cista, which must with Duvau be placed a little before 200 B.C. Further, there are just seven men and seven legends; the other legends are obviously utterances of the men, and it is hardly likely that one man alone would lack his speech, and that a general title for the whole scene would be placed just where No. 2 should have his speech engraved. Legends Nos. 2, 3, 4, and 6 run from right to left, and Nos. 1, 2, and 7 run almost vertically downwards; but no certain letters are engraved upside down with reference to the rest of the same legend. Now in No. 2 the second letter is much like a Greek, and may be taken as an opened o or an R with an irregularly drawn hasta; the fourth letter has the third stroke of the N separate from the first two, and may be read pi, but not vi unless the v be read upside down. Thus coepi is possible, but not crevi. Pia for -lia also involves reading a letter upside down. In legend No. 5, the only simple reading is made mi recie, unless we are to accept uncrossed t's, for which I see no real reason. I accept therefore the following text: CONFICE PISCIM; COEPI ALIA; COFECI; FERI POROD; MADE MI RECIE; MISC SANE; ASOM FERO. 4 Now as even Duvau recognized, there is nothing in the scene to make F. Sommer, Handbuch d. Lat. Laut- und Formenlehre2 77 (1914), expresses doubt that the reading coenalia is correct. 3 Professor J. C. Rolfe, in about 1904, told me that some scholar (I do not know whom, nor if the view was ever published) read this legend backward as ait laneos 'the butcher speaks'. Such a view involves reading reversible letters backward, runs the inscription toward and not from the mouth of the nearest man, and does not make it an utterance of the man, but a label; though the e of laneos (= Latin lanius) is justified by Praen. fileia and fileai, CIL I2. 2. 60 and 561. 4 Pages 303-4. us think that it is a sacrifice; the mention of a fish is enough to dispose of such an idea. There is equally nothing to indicate that the man who is departing on the right is carrying an offering to the Lares: the whole proceeding suggests an ordinary tavern scene, where food is being prepared for guests. One may compare the scenes of similar nature used to adorn the walls of a wine-shop at Pompeii." The meaning of the scene is then as follows: No. 1, holding in his hand a pan of some kind, says to No. 2 confice piscim 'finish up the fish'. No. 2, taking down the carcass of a calf or the like, replies coepi alia 'I've begun something else'. No. 4, returning from the actual cooking with an empty tray, says to No. 3 feri porro 'cut up some more', to which No. 3, with a full tray, replies confeci 'I've finished doing it already'. No. 5, addressing the brazier over the fire, remarks made mi regie 'boil for me royally', and No. 6 admonishes him misce sane 'stir it then!' No. 7, hastening to the table where the guests are waiting, calls out assum, fero 'here I am, I'm bringing it!" The linguistic features are as follows: (1) The c in recie has the old value g. (2) The -c of misc has the value of the name of the letter, ce. (3) The -im of piscim has not yet been replaced by -em, after consonant stems. (4) The n before ƒ is written in confice, but not in cofeci. (5) The single consonant is still written for the double (or long) consonant, in porod and asom. (6) Vowel weakening is complete in confice, but not in asom. (7) Final d after a long vowel is written in porod, but not in regie and sane. (8) The final diphthong is written not ei, but i, in cofeci and coepi. Of the last two items I must speak more at length. 9 The variation in the finals of porod and regie sane is matched in the well-known inscription from Spoleto, which contains the following forms: violatod, licetod, datod bis, suntod, exvehito, exferto; eod, quo, dolo bis, malo; bovid bis; die bis, causa. The final d is here lost after a and e, and six times after o, as against six forms with -od and two with -id. There is no reason to assert that the d was lost after all long vowels at precisely the same time; the loss may have started after a and e, and then passed to the position after o, and lastly have affected -id. This Despite Lommatzsch, 1. c. • Described in Mau-Kelsey, Pompeii, 394–6 (1899); cf. also the scenes of ordinary life in a private house, ib. 54-6. "I follow Miss Turnbull's interpretation of asom in the preceding article. 8 Ernout, Parler de Préneste 17 = MSL 13. 309. • CIL XI. 4766 = 12. 2. 366; the final word estod now lacks its last two letters. would explain the phenomena in the Spoleto inscription; and our Praenestine cista, with -od but -e, stands in perfect agreement. The final monophthong in cofeci and coepi is more astonishing.10 In the Latin of Rome we do not find this writing for the earlier diphthong until 150 B.C.; but our Praenestine cista is at least fifty years earlier than that date. Moreover, this diphthong appears rather as e than as i in other Praenestine inscriptions: dat. Hercole, Hercule; nom. pl. magistres, coques, fabres, Pontanes; abl. sueque 'suisque' ede 'iisdem'.12 But the readings with -i in the perfect are confirmed by dedi in another Praenestine inscription,18 alongside donom with unweakened o, even as in our cista cofeci and coepi are accompanied by asom. It would appear that the -ei of the first singular perfect, from earlier -ai, became a close e like any other earlier -ai or -oi and then followed not the normal development in Praenestine, but the influence of the paradigm in the perfect tense: -i-stei, -i-t, -i-mus, -i-stis, which, whatever their exact forms in Praenestine,1 almost certainly had i in the position here set off by hyphens. 10 Curiously, I find absolutely no mention of this in any of the handbooks or discussions. Yet cofeci, at least, is certain. Only Meillet, Esquisse d'une Histoire de la Langue Latine 98, remarks upon the -i of Praen. dedi, but does not hint at the forms in the cookery inscription. 11 Second declension forms with added -8; Sommer, op. cit. 346–7. = 12 For these forms, cf. Ernout, Parler de Préneste 34–5 = MSL 13. 326-7. 13 CIL XIV. 2863 I. 2. 60, which contains also two genitives of the third declension in -os, nationu with weakened vowel and diovo with unweakened vowel. 14 CIL XIV. 4112 = I2. 2. 561, apparently of mixed Praenestine-Roman origin, has dedit and fecid. SPANISH ETYMOLOGIES CARLTON C. RICE CATAWBA COLLEGE [1. estragar < *stragare: strages. 2. sesgar < *sesecare (defended against Spitzer and Ullrich). 3. simado, sima: σiós. 4. sosegar < *insulsicare.] 1. Estragar 'to spoil', estrago 'ravage'. It is plain that the noun is not derived from the Latin noun strages, but is a postverbal formation. Meyer-Lübke, in his Romanisches etymologisches Wörterbuch, postulates a Vulgar Latin *stragicare, 'verheeren', evidently connecting this type with strages, as *caballicare is formed from caballus. However, it is questionable whether *stragicare would have developt into Spanish estragar. An analogous word, cogitare, becomes cuidar in Spanish. Meyer-Lübke, Romanische Grammatik 1. 444, indicates the regularity of the development cogito > cuido; likewise R. Menéndez Pidal, Manual elemental de gramática española3 66,110. Since no clear case of the development of Latin -agicare into Spanish -agar has ever been cited, it seems reasonable to hold that the type *stragicare would have developt into *estraigar in Spanish. I postulate the type *stragare, formed from strages as plantare is formed from planta. The phonetic development *stragare > estragar is doubtless regular, being exactly paralleled by that of plagare > llagar. 2. Sesgar 'to cut on the bias', sesgo 'oblique'. The type *sesecare 'to cut apart', proposed by me in the Publications of the Modern Language Association of America 20. 343, after having been accepted by Meyer-Lübke, Romanisches etymologisches Wörterbuch, has been attackt by Leo Spitzer, Revista de filología española 13. 116, on the sole ground that the prefix se- is no longer productive in Romance. Spitzer tries to show a connection between sesgar and *sessicare, a type postulated by Meyer-Lübke to account for Old Spanish sessegar 'sich setzen, sich niederlassen', and by Meyer-Lübke taken as the source of Spanish sosegar 'to quiet'. I shall discuss below both sosegar and *sessicare. On the formal side, we should note in the postulated development *sessicare > OS sessegar > S sesgar the suspicious dropping of a syllable between the old and modern Spanish period. Spitzer cites, to be sure, one similar-looking series, vindicare > vendegar > vengar; but this comparison suggests phonological and morphological queries. Would *sessicare, a late formation, show the same treatment of the second vowel as vindicare, an ancient word? Is not vendegar (the occurrence of which I am unable to verify) a semi-learned rather than a popular development? In Menéndez Pidal's Manual 651, 133, I find vengar given as the popular, vindicar as the learned derivative of vindicare. If vendegar is the regular Old Spanish form of vindicare, as Spitzer seems to think, why is it not to be found in Meyer-Lübke's Romanische Grammatik or Romanisches etymologisches Wörterbuch? On the semantic side, the suggested etymology leaves even more to be explained. Spitzer supposes that the original sense of sesgar was 'to set', mentions Spanish sesga 'gore', compares the German phrase 'einen Lappen (Flicken) einsetzen', but fails to attack seriously the problem of showing that the meaning 'to set' might naturally shift to that of 'to cut across'. To be sure, sesga 'gore' is to be associated with the verb sesgar or with the adjective sesgo, but it is readily interpreted as meaning etymologically 'a cutting' or 'an oblique piece'. Spitzer suggests that the type *sesecare, which has always displeased him, might be eliminated in the interest of reducing the number of etyma in the Romance dictionary. However, it is phonetically perfect, and gives a meaning quite close to the meanings of the Spanish derivatives assigned to it. Its formation being paralleled by that of secedere, secernere, secludere, semovere, seponere and various other Latin words, the linguistic facts shown seem to me to indicate that the word *sesecare probably existed in ancient Latin. The type *sexicare, formed from an assumed participle *sexus for sectus was suggested by Ullrich in the Zeitschrift für romanische Philologie 4. 383. Meyer-Lübke, REW 592, calls the formation of *sexus improbable. However, Ullrich cited no less than 13 variants like fixus beside fictus, etc.1 While the formation *sexicare seems to me justifiable, its short vowel makes it phonetically a less desirable type than *sesecare; and the meaning of the latter word is also somewhat closer to that of the Spanish verb. 1 Fartus farsus, fixus fictus, fluxus fluctus, frictus frixus, indultus indulsus, mersus mertus, emulsus emulctus, pulsus pultus, sartus sarsus, scriptus scripsus, tensus tentus, tertus tersus, tortus torsus. |