The last item to consider is the numerals, which usually appear in both the Didascaliae and the praefationes, and purport to give the chronological order of the several plays. But before any comparison of Donatus with the manuscript sources is possible, certain differences in the latter require explanation. Are these due to accidental or arbitrary changes by scribes, or do they represent independent traditions, and therefore possess ancient authority? Much depends upon the answer to this question, for any error on this point will lead to other and more serious mistakes. The differences involve directly only the Hautontimorumenos, for which not only the Bembinus, but also the best representatives of the y family, and at least one important & codex, Paris. 10304, give III (or TERTIA). But the numeral II (or SECVNDA) is the reading of the Lipsiensis (L), pronounced by Kauer the best of the d family,1 the Paris. 7903, a tenth century y codex, the Riccardianus (E), a so-called mixed manuscript of the eleventh century, and a considerable number of later and inferior codices. Lastly the Victorianus (D), a & manuscript of the tenth century and one of the best of its class, has the numeral IV. On the origin of the numerals II and IV in the Didascalia of the Hautontimorumenos scholars are not agreed. Dziatzko argued strongly in support of the ancient origin of both.2 Karsten and Rabbow, apparently without further investigation, accept his conclusions. Hence one or the other of the two numerals is assigned to the Hautontimorumenos in three of the five chronologies which Karsten believed to have ancient authority,3 and the numeral IV is the basis of an argument by Rabbow in support of his theory of the twofold origin of the praefationes. Most editors of this play give the numeral II in the Didascalia, usually justifying this act by an appeal to the reading of E, L, Paris. 7903. Fabia and Torchiana are about the only scholars who have rejected the independent origin of both numerals. Torchiana seems to regard them as merely variants of the read 1 Zeitschr. f. d. österreich. Gymn., LII, p. 988. 2 Rhein. Mus., XXXIX, p. 339 ff. 8 Mnemosyne, XXII, p. 179 f. ing given by the great majority of better codices.1 Fabia holds the same opinion about the numeral IV, but follows Dziatzko's earlier suggestion that II (or SECVNDA) is a correction more or less ancient designed to harmonize the numerical order with the order of consulships.2 Such divergent views show that the arguments thus far adduced are not decisive. After reviewing these arguments, therefore, I propose to bring forward new evidence in the hope of ending the controversy. Dziatzko, the first to write on this subject, held that II (or SECVNDA) was read in the archetype of the & family, its source being some ancient manuscript independent of our codices. Neither of these positions can be proved correct. Evidence, not conclusive, it is true, but still very strong, indicates that both are incorrect. The first seems to be supported by L, which is alone of the better & codices in this respect. It is opposed by Paris. 10304, which has III, and by D, which has IV, a reading which, if corrupt, came almost certainly from III, not from II. Dziatzko did indeed appeal to the evidence of E, but this was due to his erroneous belief in the close relationship of this manuscript with the & family. In the order of plays, the division of scenes, and usually in its text, E clearly betrays its origin from the y family. To a limited extent it has been under the influence of 8 codices, but this does not warrant the conclusion that it is essentially a & codex itself. Since it can throw no light on the archetype of the family, Dziatzko's conclusion rests upon L alone, and this is overbalanced by other valuable manuscripts of the same family. But even if the archetype of the & family, in common with the other classes of Mss, did make the Hautontimorumenos the third of the plays, is it not still possible that the readings II and IV have come down from ancient times through independent recensions? Such a theory is extremely improbable. Dziatzko was not disturbed by the fact that these numerals are assigned to other plays in the same manuscripts, but 1 Quo tempore P. Terenti fabulae primum actae sint, p. 35, adn. 1. 2 Les Prologues de Térence, p. 35, n. I. Fabia and Torchiana justly regard this as a serious objection to the theory. Why, they might have asked, have these ancient codices altered the numeral of the Hautontimorumenos alone? Why did they not change those also of the Eunuchus and Phormio? Why did they not supply the missing numerals for the Hecyra and Adelphoe, or the missing Didascalia for the Andria? These are points on which a scribe might well have consulted other codices, but there was no reasonable ground for his doing so in the case of the Hautontimorumenos alone, for which a numeral was already given. Dziatzko's theory, despite its evident weakness, is still generally accepted. Fabia and Torchiana have commanded little or no support. This is doubtless due to their failure to show how the variant readings originated. In the conviction that they are right and Dziatzko wrong, I shall try to account for the variants. To make it easier for the reader to follow the argument, part of the Didascalia of the Hautontimorumenos is given below exactly as it appears in Paris. 7899 (P), representing the majority of better codices, and in Paris. 7903 and the Lipsiensis (L), representing the variant II (SECVNDA). It is unnecessary to give anything for the Riccardianus (E), since, except for the division into lines, its Didascalia is identical with that in Paris. 7903. DIDASCALIA OF THE HAUTONTIMORUMENOS. INPARIBVS DEINDE DVABVS DEXTRIS GRECA MENANDRI FACTA II Dziatzko believed in the existence of some connection between the version of L and that of E, Paris. 7903, basing this view in part on the presence of the same numeral, in part on the omission of the words ACTA PRIMVM (PRIMA or PRIMO) in both, the omission being somewhat more extensive in E, Paris. 7903. At first glance these parallels seem to contain decisive evidence of some connection, but a closer scrutiny discloses some differences which are rather strange in versions apparently closely related. E, Paris. 7903 have an unusual order of items, while that in L is entirely normal, they have ACTA, not FACTA, as in L, and the numeral is the written word SECVNDA, not the symbol II. Moreover, the lacunae are by no means similar, for while L omits the words Acta PRIMO, the word PRIMA is found in both E and Paris. 7903. In fact, the two versions, apart from the numeral, resemble each other only in points wherein they resemble the version of any other manuscript of this play. The differences suggest that the two are really independent of each other, and this is shown to be true by the very evidence upon which Dziatzko relied. First, it is necessary to point out a striking parallel in the Didascalia of E, Paris. 7903 for the Hautontimorumenos, and that of all classes of later Mss for the Eunuchus. By the oversight of an early scribe the name of the modulator of the Eunuchus lost its usual position next before the tibiae. The same scribe, in all probability, noted the omission of this item only after he had written the numeral. At this point he inserted the reference to the modulator, which appears, therefore, after the numeral and next before the consuls. This is the order in which the several items appear in the Bembinus. The error cannot be ascribed to the copyist of this manuscript, but was made by some predecessor, for the yo codices preserve evidence of the same peculiar disorder. In some common ancestor of these an evident attempt was made to correct the order, for the item on the tibiae was removed to a position following the modulator. That it was so removed can scarcely be doubted. If it is restored to its usual position, the disorder is precisely like that in the Bembinus. Such is Dziatzko's explanation of the unusual order of items in the Didascalia of the Eunuchus.1 But so far is he 1 Rhein. Mus., XX, p. 582, and Anm. 9. from explaining the unusual order in which E, Paris. 7903 give the items for the Hautontimorumenos, that he nowhere refers to it as in any way peculiar. Now this disorder is in every respect identical with that of the Eunuchus Didascalia in the yo codices. In the Eunuchus the disorder must be ascribed to accident, but such an accident a second time in the same items of another play seems extremely improbable. The thought suggests itself that the two may have some connection, and this is confirmed by certain points in which the Didascalia of E, Paris. 7903 for the Hautontimorumenos differs from that of the other Mss for this play, but resembles that given by them for the Eunuchus. These are the omission of EST in the reference to the Greek author, the use of ACTA instead of FACTA, and the form of the numeral SECVNDA instead of II. The preservation of PRIMA and the use of MODOS FECIT instead of MODVLAVIT preclude the thought that E, Paris. 7903 really derived this part of the Hautontimorumenos from the Eunuchus. But the extremely close resemblance renders it practically certain that the scribe of a codex from which E, Paris. 7903 are descended recast this portion of the Didascalia of the Hautontimorumenos on the pattern of that of the Eunuchus. What was the ground which prompted this act and the consequent change in the order of items? The answer is found in the item on the tibiae as given by E, Paris. 7903. The preservation of the word PRIMA shows that the omission of the words INPARIBVS DEINDE was accidental, for no scribe would have kept the one word if he had deliberately dropped the other two. The form of the Didascalia in P, found also in other Mss of the y family not only for the Hautontimorumenos, but also for the Eunuchus,1 shows how easily such an accidental omission came about. The item on the tibiae, including the words ACTA PRIMA, occupies three lines. After some scribe had written TIBIIS, his eye, we must suppose, fell upon DVABVS DEXTRIS in the next line, which he wrote, and then, without noticing his error, proceeded to the item on the Greek author. But the omission of the first set of tibiae left 1 See p. 142. |