^{Note 1 in page 930} For Italian this has been very well pointed out by McKenzie in Italica, xiii, 69 et seqq., who writes (p. 70): “So obvious is the inappropriateness of interpreting se in the formula of adjuration as ‘if,’ that some authorities distinguish it from se ‘if,’ and derive it from Lat. sic, in spite of the fact that the phonetic obstacle is as great as in the case of si. Thus Cappuccini, Vocabolario della lingua italiana, asserts that it is an obsolete adverb equivalent to cosí, ‘augurando e deprecando,’ and derived from sic; he quotes Inf., xx, 19. On the other hand, Zingarelli in his Vocabolario, after stating that se ”esprime sempre una condizione, sia in una frase schiettamente dubitativa o ipotetica, sia nella concessione, sia nella limitazione, sia nel desiderio,’ gives as one definition ‘o cosi fosse,’ adding: ‘si fa risalire a sic, ma anche si aveva questo uso.’“

Cf. anche Meyer-Lübke, Gramm, rom., 1, 539 et seq., §613, 3, 632, §562 (“sī avec une voyelle surprenante [!]”).

^{Note 2 in page 930} I can attribute no importance, of course, to the two passages from Italian humanists of the 14th and 15th centuries quoted by Gaspary, ZRPh. xi (1887), 137 and by G.-G. Nicholson, Rom., lxi, 11.

^{Note 3 in page 931} I was fortunate enough to find, in the library of Princeton University, the Zeitschrift für die Altertumswissenschaft, xii (1854), 236, which Spitzer was not able to see (Rom., lxv, 300, n. 1). The sentence si me Deus adiuuet, non habeo nihil is found in W. Grimm, Altdeutsche Gespräche, publ, in Abh. der Akademie der Wiss. (Berlin, 1851), p. 238: this text is the Latin translation of some Old High German phrases of the ninth century (see Abh. 1849, 415). The German text reads: sam mir got helfe, ne habēm ne tropfon, so that the correction of si to sic, proposed by Pott, Zeitschr. f. Altert., xii, 236, is quite obvious.

O.H.G. sam (or sama) is translated ‘so,’ ‘ebenso,’ ‘sic,’ ‘aeque,’ ‘similiter,’ by O. Schade, Altdeutsches Wörterbuch, 2d ed., 740.

I may also remark that the passage se m'aiti Iddio (io il vi credo) is not in Brunetto Latini, as Spitzer says (p. 309), but in Boccaccio's Decamerone (Giorn. iv, nov. ix, 8).

^{Note 4 in page 931} The original desiderative meaning of the (conditional) Lat. sī is denied by Blase, Glotta, xi (1921), 145 et. seqq.—o si (o mihi praeteritos referat si Iuppiter annos Virgil) of course is comparatively frequent (see Blase, Hist. Gramm., p. 134) and has given Italian oh se; but this is something quite different.

^{Note 5 in page 931} I only wish to reinforce Spitzer's criticism on one point.

I do not see great semantic difficulty, as Prof. Nicholson does (Rom., lxi, 6 et seqq.), in drawing from the conditional Latin sī the Rumanian să with the sense of purpose (Lat. ut), in sentences like that quoted by Meyer-Lübke, Grammatik, iii, 641 and by Nicholson himself, Rom., lxi, 10: Duce-mă-voĪu ĪarărĪ Īntr' altă celate, să propovoduesc cuvântul lui Hs ‘je veux m'en aller de nouveau dans une autre cité, pour annoncer la parole de Dieu.’ I completely agree with the explanation which Tommaseo-Bellini, Dizionario, s.u. se, p. 735, n. 3 gives for quite similar Italian cases: ‘Talora innanzi al se si lascia per sottinteso il verbo necessario. Ell[issi] evidente e efficace. Fior. s. Franc. 112 (M): Viene il demonio […] per sospingerlo quindi giuso. Di che Santo Francesco, non avendo dove fuggire […] di subito si rivolse con le mani e col viso e con tutto il corpo al sasso […] brancolando colle mani, se a cosa nessuna si potesse appigliare (cioè cercando se … ); e 147: Corse […] per tutta la città, se per ventura la potesse trovare; Boccaccio, Giorn., ii, Nov. 4, 2: A quella (tavola) s'appiccò, se ‘orse Iddio, indugiando egli l'affogare, gli mandasse qualche aiuto allo scampo suo; […] Vit. S. M. Madd. 77: Pensomi che tornassono un poco dentro alla porta […] in luogo piú onesto, che si poteva, tuttavia se si potesse vedere o udire alcuna cosa; Vit. S. Gir., 4: Andavasi (il leone) discorrendo per la foresta forse se per ventura potesse rinvenire lo compagno suo.’ Cf. also the Dizionario degli Acc. della Crusca, s. u. se, §1, vol. 5, 117.

Before the se in all of these sentences we may supply a purpose clause (per vedere se); sometimes we might, almost without changing the meaning, substitute a purpose clause for the se-clause: so e.g. in Vit. S. M. Madd., 77: per vedere o udire alcuna cosa (or per tentare di vedere o udire alcuna cosa); or in Fior. S. Franc., 112: per appigliarsi ad alcuna cosa (or per tentare di appigliarsi […]). Similar sentences are also found in Modem Italian, particularly in the familiar speech: cf. e.g. Collodi, Pinocchio, chapt. 24: Allora il burattino […] si pose a guardare di qua e di là se per caso avesse potuto scorgere su quella immensa spianata d'acqua una piccola barchetta con un omino dentro.

I should approach very nearly the meaning of the Rumanian sentence quoted by Prof. Nicholson, If I were to say in colloquial Italian: Voglio andàrmene di nuovo in un'altra città (per vedere) se annuncio la parola di Dio.

See also Bourciez, Eléments de linguistique romane, 279, §254b; 600 et seqq., §§508 et seq.; Meyer-Lübke, Gramm., 3, 667, §590.

Rumanian until the sixteenth century had să (<sĪ<sī) in the conditional sense (Bourciez, 602, §508) and according to Tiktin, Rumän. Elementarbuch, 143, §377 has it now in unreal conditional sentences (he quotes an author of the late nineteenth century, Cosbuc, 178, cf. the note at p. 179).

For the other functions of să (se) in Rumanian, cf. Meyer-Lübke, Gramm., 3, §§18; 117–119; 340; 387, 562; 567; 569–570; 573–575; 582; 643–644; 669; 673; 679.

^{Note 6 in page 932} Italica, xiii (1936), 69 et seqq. I add one example from Diez, Rom. Gramm., 5th ed., iii, 357; one from Spitzer, Rom., lxv, 309; one from Miss Lograsso, Italica, xv, 152, n. 6; two from L. G. Blanc, Grammatik der italiänischen Sprache (Halle, 1844), p. 585; seven from the Vocabolario degli Acc. della Crusca (Verona, 1806), s. u. se, v, 117, four from Tommaseo-Bellini (s. u. u. aitare, ajutare, salvare). The remaining ones which I quote below (and particularly all those of the Filostrato) have been indicated to me by the inexhaustible kindness of Prof. McKenzie.

^{Note 7 in page 932} This distinction between ‘adjuration’ and ‘asseveration’ (or ‘protestation‘) is quite correctly indicated by McKenzie, p. 70; but then his expression on p. 69, line 5 from the bottom, is not quite exact.

^{Note 8 in page 934} This work is attributed by Francesco Tassi (in the Lezione preceding the text, p. 407) to the “celebre e dotto Frate Agostino della Scarperia, che fioriva sul declinare della metà del secolo decimoquarto.” It is an imitation of a Latin treatise De uita contemplatiua composed by “Guigo Quinto Priore della Certosa Maggiore.”

^{Note 9 in page 935} No instance of this use of sīc is found before Catullus; hence it is quite possible that it did not exist in Old Latin. Ch. E. Bennett, Syntax, i, 111 considers the sīc-sentence with future in Terence, H. T., 463: sic me di amabunt, ut me tuarum miseritumst, Menedeme, fortunarum as desiderative, but this appears to me to be very doubtful; cf. Sjögren. Zum Gebrauch des Futurums im Altlateinischen, p. 73.—See, however, what is said below on the passage of Plautus, Miles, 571.

On the other hand, we find no examples after Apuleius (2d cent. a.d.), although the temptation is great to correct si ualeas to sic ualeas in Hist. Apollonii, vii, p. 11, 3 (MS AP): indica [dic other MSS] mihi, si ualeas, quae est haec causa, quod ciuitas ista in luctu moratur? Cf. W. Kroll, Glotta, vii (1916), 80; Hofmann, Lat. Gramm, p. 772, at the beginning of the page. The Historia Apollonii belongs perhaps to the sixth century a.d. Cf. also W. Kroll. La sintaxis científica en la enseñanza del latín (Madrid, 1935), p. 92, n. 1, who quotes, as proof of the existence of the formula of adjuration sic ualeas, the quite vulgar inscription CIL., iv, 6641 (Pompeii): cacator sic ualeas ut tu hoc locum trasea[s]. We must never forget how incomplete is our knowledge of Vulgar Latin!

It is also possible that in the sixth century the pronunciation si instead of sic was so common, that it crept into the text of Hist. Apoll.; si ualeas would then represent the spoken form of sic ualeas and should not be touched. This pronunciation is attested for the Vth century by the grammarian Consentius before following c-(people used to say si custodit) instead of sic custodit G. L. Keil, V, 395, 11 ff.; but hi for hĩc is already found in Pompeii (CIL. iv, 2048) before 79 A.D.; cf. also E. Richter, Beih. ZRPh., vol. 82 [1934], 74 note with more material).

^{Note 10 in page 938} I cannot well understand why Blase, Hist. Gr., p. 134, considers this passage of Apuleius as ‘merkwürdig’; nor why he says that ‘es fehlt der zu erwartende Imperativ und der Wunschsatz ist einem imperativischen Satze koordiniert.’ Of course, we must put a comma after Genios—as e.g. the edition of Giarratano (Paravia, 1929) and the Loeb do—and join directly per fortunas uestrosque Genios with decepto seni subsistite; then ‘der zu erwartende Imperativ’ is subsistite, and the whole sentence is identical with the others, and by no means ‘merkwürdig.‘

^{Note 11 in page 939} As Prof. McKenzie (p. 71) correctly observes, se with the present subjunctive is extremely rare, we could almost say irregular, in Italian, outside of these formulas of adjuration. The se mai continga che il poema sacro […] of Par., xxv, 1 et seqq. is interpreted as being half desiderative by Prof. Spitzer (p. 303, n. 1) which is probably right (cf. the use of contingat in CIL., viii, 1070, 4–5 [c]: ita tibi contingat quod uis, ut hoc sacrum non uioles; Carm. Epigr. 1067 (Rome): síc tibí contingat felíciter ire, uiator/immatura meo perlege fata loco; 1287: opto meae caste contingat uiuere natae; Itala, Genesis, 44, 7 [Lugd.]: non contingat pueris tuis facere hoc uerbum [lxx]).

^{Note 12 in page 940} For Old Italian, this has been pointed out with reason by Prof. McKenzie, Italica, xiii, 71 et seqq. Some of the Old French examples which I quote in the text are of the same kind (Huon de Bordeaux, p. 106; Am. et Am., 3345; Percev., 4028; 1490; 596), as are also one Portuguese and one Spanish example (Appoll., 173).

^{Note 13 in page 940} The connection between the Italian and the French “desiderative” se, denied by Foulet, is strongly asserted by Spitzer, Rom., lxv, 309: “une interprétation intra-française est donc exclue et il ne reste que l'interprétation ‘interromane’ de Diez et de Gaspary.”

^{Note 14 in page 940} For Spanish, more material (also of the “asseverative” type) may be found in Menéndez Pidal, Cantar de mío Cid (Madrid, 1908), i, 372.

^{Note 15 in page 941} I do not include in my enumeration the instances of correlative sīc-ut, because the type is somewhat different; hence they may not be so well compared with the Italian passages. The examples most nearly resembling the “imperative” type is evidently CIL., iv, 6641: cacator sic ualeas, ut tu hoc locum trasea[s] (cf. W. Kroll, La sintaxis científica en la enseñanza del latín [Madrid, 1935], p. 92, n. 1; compare also, with ita, CIL., viii, 1070: ila tibi contingat quod uis ut hoc sacrum non uioles).

Instances of the “asseverative” type are:

Catulus, xlv, 13 et seqq.:

Here the character of “asseveration” is clearly indicated by the translation of F. E. Cormish in the Loeb Classical Library (New York, 1914) 55, who considers it necessary to add the words “I swear”:

“So, my life, my darling Septimius, so may we ever serve this one master as (I swear) more strongly and fiercely burns in me the flame deep in my melting marrow.”

Propertius has (i, xviii, 11):

This usage of sīc is very frequent in Ovid, whose style is relatively popular: Epist., xv, 25:

Metam., viii, 866 et seqq.: quoque minus dubites, sic has deus aequoris artes

In the Ibis there is an abundance of these clauses: vv. 64 et seq.; 273 et seq.; 315 et seq.; 331–337; 339 et seq.; 341; 343–348; 421–424; 235 et seq. (with tamquam instead of ut); 469–476; 521 et seq.; 549 et seq.; 569 et seq.; 583 et seq.; 591; 601 et seq.; 603–606; 621–626 (double). Total: 21, in a little poem of 644 verses! More material in the other works of Ovid can easily be found, I believe, with the help of the excellent Ovidian concordance published by Deferrari, Barry, and McGuire (Washington, 1939).

Ital. cosí, Spanish así, Old French si, French ainsi also are used exactly the same way for Spanish I may quote Cervantes, Quijote, ii, 1:

More material is in R. J. Cuervo, Diccionario de construcción y rgimen de la lengua castellanga, i (Paris, 1896) s. u. así [6], 698 et seq.

For Italian, we can read, e.g., Orl. Fur., xxviii, st. 13:

But they are used in optative sense also without the correlative como, come. For Spanish I may quote the sapphic ode of Estéban Manuel de Villegas (1596–1669) indicated to me by the kindness of Prof. Américo Castro; see Menéndez Y Pelayo, Las cien mejores poesías de la lengua castellana (Buenos Aires-Madrid, no date), p. 145:

More material of this type for Spanish may be found in R. J. Cuervo, Diccionario, s. u. así (5), pp. 699 of s.q., and in the Diccionario of the Academia Española, I (Madrid, 1933), s.u. así (8) 841, col. 2; for Italian see Tommaseo-Bellini, s.u. cosí, nn. 15 and 18, 1781, col. 3.

For French, see e.g. Godefroi, vol. 1, 181; vol. 7, 415 seq. (si m'aist Deus etc.; ainsi m'aist Dieux) and the Grand dictionnaire universel of P. Larousse, I, s.u. ainsi, 160, col. 1 (Malherbe, Bossuet, Fénelon, etc.): ainsi puisse la France être toujours prospère!

I know of no other cases of desiderative sīc in the whole Latin language, outside of the two types examined in this paper, the adjurative and the asseverative type (both with ut and without ut).

^{Note 16 in page 944} The Italian paraphrase of the ed. Pellegrini, p. 220, reads: “Amor mio, se Dio mi tenga lungi ogni danno e mi conceda gioia in voi, io vi attesto [!], o amore, che nel tempo in cui dapprima fui desideroso di voi, la mia età era sí picciola (ero cosí giovane) che non mi sembra d'aver giammai visto né prima né dopo di voi donna alcuna, che piú suscitasse il mio desiderio di servirla e d'amarla.”

^{Note 17 in page 947} There is little doubt, I think, that both Old Italian desiderative and conditional se are stressless (pretonic). For Old French se <sīc, see Rydberg, Geschichte des französischen (Upsala, 1907), i, 855 et seqq.; also Gilliéron, Abeille, 283 et seqq.

As for the fall of final -c, cf. e.g. Fr. là, It. là, Sp. allá Lat. illāc; It. lí, Sp. allí Lat. illīc. Cf. also note 9, at the end, and note 22.

^{Note 18 in page 947} The oldest Provençal texts show si, not se: see Place, Hispanic Review, v, 262. Did se come later from Northern France?

^{Note 19 in page 947} Old Rumanian has se, exactly as Italian, French, and Portuguese; cf. Meyer-Lübke Gramm, rom., 3, 639, §567. Modern Rumanian has să, derived from se, and this from *sĪ(c).

^{Note 20 in page 948} In La sintaxis científica (Madrid, 1935), p. 92, n. 1 he adds CIL. iv (Pompeii), 6641: cacator sic ualeas, ut tu hoc locum trasea[s], the vulgar character of which no one would deny. Cf. also with the CIL. viii, 1070, 4–5 (c): ita tibi contingat quod uis, ut hoc sacrum non uioles.

^{Note 21 in page 948} E.g. the so-called “poetical plurals” of the type gaudia> French joie; see Bonfante, Emerita, iv (1936), 245; G. Devoto, Storia della lingua di Roma (Bologna, 1940), 173; 238. See also the use of gyrus instead of circulus (Bonfante, Emerita, v [1937], 24 et seq.; Devoto, 334; 337).

^{Note 22 in page 950} The relation between sī and sīc was originally a very close one; cf. e.g. W. Kroll, La sintaxis científica p. 91 et seqq.; Ernout-Meillet, s.u. sï; Hofmann, Lat. Grammatik, p. 771 et seq. (with bibl.). There were once the same word, sīc being nothing more than *sī-ce (*sei-ce).

In any case, sīc and sī confused largely in late Latin: see on this subject Hofmann, Lat. Gramm., p. 772 at the top of the page; Rydberg, Geschichte des französischen , (Upsala, 1907), i, 215 et seq.; 236 et seq., with material and bibliography. About the weakening of sī to sĪ>se, see the same work, i, 224 et seqq.; about sīc>se, ibidem, 855 et seqq. Then Gilliéron, Abeille, p. 283 et seqq. and see above notes 9 and 17.

About sī: sĪ in Latin, see Bourciez, Eléments de ling, romane, 3d ed., 282, §257: “il faut observer qu'en latin vulgaire l'ī s'y était abrégé dans des combinaisons sĪ quis, sĪ quidem (Ov[ide], Fast., 4, 603), d'où une forme romane se (sauf en espagnol où le classique si a été conservé, et en français où il a été rétabli vers le XVe siècle).” Cf. also Leumann, Lat. Gramm., 102. §87 A 4.

^{Note 23 in page 950} I wish to express here all my gratitude to Prof. McKenzie, who not only gave me excellent advice during the redaction of this paper, but also generously placed at my disposal his Old Italian material, fruit of his patient and personal investigation.