M. Tullii Ciceronis De Natura Deorum Libri Tres

ON THE RELATIONS OF ORELLI'S MSS TO THE
ARCHETYPE AND TO EACH OTHER.

The archetype from which all the existing Mss are derived must have been in a very bad condition, having lost considerable portions of the 3rd book, and having suffered much from dislocation in the 2nd book and, as I believe, in the 3rd also. Of the four sections, into which the 3rd book is divided, the whole of C (dealing with the Stoic argument in proof of the Providential Government of the universe) and a considerable portion of D (dealing with the argument in proof of a special Providence ordering the affairs of men) have been lost; and smaller losses occur in the same book at the end of § 13 and 64, and in 1 §§ 25, 65. The great instance of dislocation is in the 2nd book, where $$ 16-86 are placed after § 156, but in my note on III 43 ut jam docebo, I have attempted to show that we have there also a transposition in the мss of 53-60, which ought to come before § 43. For examples of what appears to me dislocation on a smaller scale, see the notes on 1 6 qua quidem in causa, 1 97 an quicquam-vidimus (which, following Bake, I have transferred to the end of § 88), 1 110 atque ita dimetata―appareat, transferred to § 104, also II 167 magnis copiis, III 29 cumque omne animal, § 34 etenim, § 70 n. on Da (3). The cause of the dislocation in Bk. II was doubtless that the archetype had come to pieces and was bound up again in wrong order, without any regard to sense. So B (Cod. Leid. 86) interposes part of the De Divinatione after N. D. 1 64, giving the following $ (64-91) after De Fato § 4, see note on the latter passage in Orelli's ed., and Bake's ed. of the De Legibus p. 104 foll., where B is thus described "constat quaternionibus XXIV. quorum singula octo folia habent. Ceterum descriptus esse videtur e libro cujus quaterniones misere disjecti essent; locis quippe plurimis, continuata scriptura, alienissima interponuntur, omniumque ordo et series turbantur, quae tamen ne legenti fraudem faciunt, vetus manus in margine subinde solet indicare." In Bk. III I have suggested that it may have been Cicero's own intention to omit the mythological section, and that this was afterwards inserted. by the original editor in a wrong place. As to the smaller transpositions they are probably owing to the wrong insertion of marginal notes or of additions written on slips of parchment which got fastened to the

wrong leaves. As to actual losses, some have attempted to account for these by supposing that the Christians destroyed the argument against Providence as impious, others relying on the statement of Arnobius to the effect that some of the treatises of Cicero caused much scandal among the Pagans, have retorted the charge on the latter*; but there is no reason to have recourse to either supposition. The condition of the archetype, as proved by the transpositions in Bk. II, is quite sufficient to account for the loss of other portions, and we meet with similar phenomena in the Leges, the De Fato, the Respublica, and other treatises.

These being the undoubted facts, as shown by all the мss, we have next to inquire which of the мss best represent the archetype and what are their individual characteristics and their relations to each other. The oldest of the Mss of the N. D. is the Vienna codex (V) of the 10th cent., but portions of the treatise have been preserved in a мs of the 9th century described by Narducci, of which an account is given in the Jahresbericht über d. Fortschritte d. classischen Alterthumswissenschaft for 1883, vol. 35, p. 75. The chief fact of importance noted in regard to this мs is that in book II it exhibits the same dislocation as the other Mss.

There is much resemblance between V and the Leyden codex (A) of the 11th century, which comes nearest to it in age. In both we find careless mistakes, but very rarely any signs of an attempt to alter the reading in order to make sense, such as we shall see reason to suspect in BCE. The Palatine codex (P), which is called perantiquust, is I presume of about the same date as A. It is not only

* See Olivet's French translation of the Natura Deorum vol. 1 p. 87 n., citing Arnob. III 7 ante omnes Tullius Romani disertissimus generis, nullam veritus impietatis invidiam, ingenue constanter et libere, quid super tali opinionatione (the distinction of sex in deity) sentiret, pietate cum majore monstravit, a quo si res sumere judicii veritate conscriptas, non verborum luculentias pergeretis, perorata esset haec causa....Sed quid aucupia verborum splendoremque sermonis peti ab hoc dicam, cum sciam esse non paucos, qui aversentur et fugiant libros de hoc ejus... cumque alios audiam mussitare indignanter et dicere, oportere statui per senatum, aboleantur ut haec scripta quibus Christiana religio comprobetur et vetustatis opprimatur auctoritas? Quinimmo si fiditis exploratum vos dicere quidquam de dis vestris, erroris convincite Ciceronem...Nam intercipere scripta et publicatam velle submergere lectionem, non est deos defendere, sed veritatis testificationem timere.

+ Ebeling in Philologus XLIII 4 p. 703, which has reached me since the above was put in type, assigns it to the 10th cent. He also gives a short account of the Laurentian Cod. 257 at Florence, which he considers to be of the 11th cent. It agrees with B (against AC Oxf.) in making the dislocation of Bk II commence

careless but capricious and for the most part quite indifferent to making sense. The first two I regard as honest but somewhat incompetent witnesses, the writer of P has no intention to deceive, but his inaccuracy almost amounts to dishonesty, while the writer of B, a far more competent witness, as far as ability goes, is not satisfied merely to report, but deliberately sets himself to improve the readings of the archetype. It is perhaps unnecessary for me to adduce any evidence of the carelessness of APV, as that will be apparent to any one who will glance over the readings of a chapter or two in Orelli's edition, but I shall endeavour to show (1) that A and V are independent of each other, but approach more nearly to one another than to any of the other мss, (2) that B is connected with the archetype by a different line of descent from the others, and has frequently preserved the true reading which has been lost by them, but that it also alters the mistaken readings of the archetype so as to make sense. The varying disarrangements in BPV seem to show that these are mutually independent and are not copied either from A or from the immediate ancestor of A; and this conclusion is confirmed by the subjoined comparison of particular readings. I have compared the readings for the 3rd book, as the мss are less imperfect for this than for the earlier books. I give first those readings in which the unaltered A and V agree against BC and P, taking no account of Orelli's 6th codex E, belonging to the 15th century. In making this comparison it must be remembered that it is only of APV that we have the complete readings in Orelli's edition. Where the true reading is not that of AV it is given first in round brackets. Readings conjectured ex silentio in Orelli's apparatus criticus are put in square brackets, as in the critical notes.

with § 15 after vicissitudines, instead of § 16 after quam deum, and in reading

requiro after rationes at the end of 1 13.

XXX

§ 15. (placari) placeri AV'.

(equo) aequo AV.

§ 18. (omniaque quae a te) omnia quae ad te A'V'.

§ 21.

(nudius) nudus A'V1.

(velis) vellis AV1.

§ 22. (dilatavit) dilata lavit V' (and by corr. fr. dilatavit) A,

§ 37.

(solstitiali) solistitiali AV'.

§ 39. (inscitiam) inscitam A'V'.

§ 45. maritumae AV, maritum hae P, maritimae BCE.

§ 46. lucus [AV], lucis B', locus C, lucos P.

§ 53. (filii) fili AV'.

§ 56.

(Penelopa) poenelopa AV.

(nefas) nefans A'V'.

(Aegyptiis) Aegyptis A'V1.

§ 58. (confici) confeci A'V'.

§ 59.

(Elide delubrum ?) elidelubrum AV.

(Apollinis) Apollonis AV'.

§ 62.

(qui) quiqui A'V1.

§ 65.

(istac ibit) ista ibit AV1.

§ 68.

(coinquinari) quoinquinari AV'.

$73. (cedo) caedo AV.

$74. (conducto) conduto AV.

$76. quom V1 and probably A'.

$78. (reprehendenda) repraendenda A, reprendenda V.

§ 80. (Reguli) reguilis V', regiulis A' probably.

§ 81. (supplicioque Q. Varius) supplicio quae que Varius AV'. si AV (al. sic &c.)

§ 82. (soleo Platonem legens) soleo l. platonem legens AV (probably beginning legens out of order. B. has superscr.)

§ 83. (Syracusas) seracusas AV'.
§ 84. (fulmine) flumine A'V'.
§ 85. (ratione) rationem A'V1.
$86. (fructuum) fructum A'V'.
§ 91. (Critolaus) critolauus AV'.
(Corinthum) corhintum AV.

$94. (cingitis) cincitis A'V',

This gives 55 instances in which AV agree against the other MSS, while the agreements between either of them and any other single мs do not amount to more than 13, as may be seen from the following lists (1) of agreements between V and one other мs, (2) between A and one other MS.

(1) Agreements between V and any other single мs.