niam, tibi dedissem, the antithesis would be, sed non habui pecuniam, (ergo tibo non dedi).

In these conditional sentences of the third class, the imperfect subjunctive never expresses past time, but is merely an imperfect as to its form. In reality, it has the force of the present, as the conditional pluperfect subjunctive has that of the perfect.

It frequently happens, however, that the imperfect subjunctive in conditional sentences, has not the force of the present, but of a proper imperfect, which implies that an act was continued during another past act. The antithesis is, in such cases, always sed non with an imperfect indicative. If the conditional clause is introduced with nisi the antithesis is, of course, formed with sed without non.

Here is to be explained the peculiarity to which the author refers in section 525.

Let us examine the first example adduced by our author, taken from Cicero pro Milone XVII. 45: Quos clamores (Clodius), nisi ad cogitatum facinus appropararet, nunquam reliquisset. The antithesis here is, sed approperabat facinus (namely, quum clamores reliquit). We can, indeed, express this by the pluperfect subjunctive in English; but then the two parts of the sentence would stand in no immediate connection with each other, whereas the Latin imperfect expresses simultaneousness with that which is expressed by the pluperfect in the following clause. We can hereby perceive how much more precise the Latin is in such expressions, than the English.

For the rest, the expression of the author is either obscure and equivocal, or incorrect, viz. that " completed actions of the past times are often transferred, at least partly, to the present, by using the imperfect instead of the pluperfect." The imperfect has nothing in common with the (real) present; it designates only a present, which was such when a PAST act was taking place.

As in the protasis, so also in the apodosis the imperfect subjunctive is very frequently used instead of the pluperfect. But this is to be explained precisely in the same way as that mentioned in the foregoing paragraph.

In the view here given, we have omitted the consideration of the clause following after the conditional clause. These invariaably form sentences by themselves, and have no direct grammatical dependence on the foregoing clause. It is, however, natural that an indicative in the one should be followed by an indicative in the other, etc.; but it is not necessary. It is the simplest way

1847.]

Metuo and timeo-neve, etc.

431

to supply, where such a dissimilarity occurs, a corresponding clause. But we cannot here enlarge upon this subject.

528. At the close of this section, it is said, that Quis putaret, quis arbitraretur, etc. are more rarely used in the sense of," who would have thought, who would have believed; and it would seem from the connection as if the examples there taken from Cicero were the only examples which occur with this writer. Our author did not probably mean to assert this, because the construction is very frequent indeed. Cf. Cic. ad Famm. II. 13. 13: quis putaret?-Ibid. XV. 15, med.: quis putaret?-pro Sext. XLL 89: quid ageret?-pro Sext. IX. 20: quis-arbitraretur? and very of ten elsewhere. The words of our author: The third person is more rarely used in this manner, should be changed to: Also the third person is very often used thus.

§ 533. Our author is not quite correct in making no difference between metuo and timeo with the infinitive and vereor with the infinitive, although the former is very rare with Cicero. Madvig, in the remark § 376 of his grammar, maintains that in good prose only vereor is found with the infinitive, and Freund, in his lexicon on the word, says expressly that timeo with the infinitive is not Ciceronian. But cf. Cic. pro Rosc. Comaed. L 4: quo nomen referre in tabulas timeat. Metuo with the infinitive and with the accusative before the infinitive is found only with the poets.

§ 535. Neve cannot stand after timeo, but either et or aut must follow this word. Timeo ne legat et scribat, or aut scribat. In the former, it is indicated that we fear both; in the latter, either one or the other.

§ 536—7. Klotz, ad Cic. Tusc. IL 26. 64, explains the distinction between non quo and non quod, by saying, non quo means always, with the intention, non quod, in the view (opinion) that. As all the passages have not been critically examined upon this point, we pass it by with adducing a few examples. Cic. ad Famm. XVI. 6. 1, quia precedes quo. The words are: Tertiam ad te hanc epistolam scripsi eadem die magis instituti mei tenendi caussa, quia nactus eram, cui darem, quam quo haberem, quid scriberem; Cic. pro Sext. XLIII 93: quo fortissimum ac summum civem in invidiam homo castus ac non cupidus vocaret, without a comparative; Ibid. XXVIIL 61: non quo periculum suum non videret, sed - putabat, without any causal particle, and with a change of construction; Cic. de R. P. p. 22 (ed. Heinrich): qui-cordatus fuit, et ab Ennio dictus est non quod ea quaerebat, sed quod ea respondebat, where the reason for the indicative is clear.

In the example taken from Livy XXX. 27, the author is doubtful whether non quia with the indicative in the protasis, is according to good usage. Compare Cic. pro Planc. XXXII. 78: non quia multis debeo, sed quia saepe concurrant; Horat. Sat. IL. 2. 89: non quia erat, sed

§ 541. Our author mentions the example in Cic. ad Att. VII. 1. which is corrected in punctuation by Bremi. But there is another passage in Cicero pro Flacco, XXXIII. extr. (where it is to be found in the ed. of Orell.); quid? nos non videbamus habitare una? quis hoc nescit? tabulas in Laelii potestate fuisse, num dubium est? Here also the punctuation presents the means of making the correction. Here it is to be thus punctuated, quis hoc nescit, tabulas in Laelii potestate? num dubium est? so that the accusative before the infinitive is dependent on the clause, quis hoc nescit, not on num dubium est; Cic. ad Famm. XVI 21: Gratos tibi optatosque (rumores) esse, non dubito, writes Cicero the son. In the words: Yet after dubito and non dubito at the beginning of the second paragraph;" the first dubito must be stricken out; for what classic author ever uses dubito thus without a negative particle?

§ 551. The indicative is found, Cic. pro Planc. XXX. 73: quod ejus in me meritum tibi etiam ipsi gratum esse dicebas. Quod is construed with negare as well as with dicere. Cic. ad Famm. VII. 16: quod — negant. Cf. Cic. pro Arch. XII. 31: quod expetitum esse videatis.

§ 553. Add after nescio quomodo, nescio quo pacto. Cic. de Amicit. XXVI. 100: nescio quo pacto deflexit.

561. The different significations of the indefinite and general expressions and the constructions appropriate to them are not pointed out with sufficient clearness in this paragraph. Quis est qui may be the paraphrase of the question with quis. In this case the subjunctive is used only when other reasons make it necessary. Cf. Cic. ad Famm. VII. 12. 21: Quis enim est, qui facit nihil sua caussa ?—ad Attic. XVI. 1. 2: sed quid est, quaeso, quod agripetas Buthrosi concisos audio?-pro Cluent. LXIV. ext. quid est, quod minus probabile proferre potuistis?-Acadd. postt. L. 4. 13: quid est, quod audio? This is rendered manifest by the addition of the pronoun illud. E. g. Cic. pro Sext. LVI. 120: quid fuit illud, quod — summus artifex egit? This use is very frequent with Plautus and Terentius. Quid est, quod has two other significations. It serves, first, for a (negative) exclamation, usually but improperly marked as an interrogation. In this case qui

1847.]

Construction of qui, quamquam, etc.

433

is followed by the subjunctive. If, in the second place, it expresses inquiry for the reason or occasion of a thing, in which case it is often changed into quid est cur, or quid est quamobrem, the subjunctive likewise follows. It would be unnecessary to cite examples which everywhere occur.

563. The example, sunt enim permulti optimi viri, qui valetudinis caussa in his locis conveniunt, where the author, by a slip of the memory, has substituted in his locis conveniunt for in haec loca veniunt, Cic. ad Famm. IX. 14. 1, is in direct contradiction to the teaching of our author respecting the construction of convenire in locis § 489. The statement should therefore be altered.

564. The subjunctive also follows qui when it has the signification although, in which case tamen follows very often. Cf. Cic. de Orat. I. 32. 145: quin etiam, quae maxime propria essent maturae tamen his ipsis artem adhiberi videram; Ibid. I. 18. 82: V. Matthiae ad Cic. pro Rosc. Amer. VIII. 23.

§ 568. The construction which follows dignus and indignus depends entirely on the sense. So quod follows, Cic. pro Rosc. Amer. L. 147: nisi hoc indignum putas, quod vestitum sedere in judicio vides; so the Acc. c. Infin. also in the same, III. 8: sum vel hoc indignissimum est, vos idoneos habitos. Also in Verr. II. 24. 58; Cic. pro leg. Man. XIX. 57, and in other places. Eximius qui is construed like dignus qui in Cic. Div. in Caec. XVI. 52: te illi unum eximum, cui consuleret, fuisse.

574. Quamquam with the subjunctive is very frequent in Cicero if one regard merely the words without searching for the reasons. Cf. de Orat. III. 26. 101: quamquam illa ipsa exclamatio―sit velim crebra; pro Planc. XXII. 53: quamquam ne id quidem suspicionem coitionis habuerit; pro Sext. XXX. 64: quamquam quis audiret?-in Vatin. XIV. 33: quamquam id ipsum esset novum; pro Mil. XXXIII. 90: quamquam esset miserum, and in many other places. As the mood does not depend upon the conjunction, but rather the conjunction upon the mood; quamquam stands with the subjunctive if the sentence requires the subjunctive irrespective of quamquam. But grammarians do best where they make the manner of thinking and of expressing thought prevailing with a people their rule and standard.

§ 575. It should have been remarked, that donec with Cicero is exceeding rare. It is nowhere found in Caesar. Our author should have noticed this distinction according to his usual custom.

579. Rem. The distinction between the conjunctions quum and si appears quite manifest in Cic. pro Rosc. Amer. XXXV.

100: Si prodierit atque adeo quum prodierit (scio enim proditurum esse), audiet.

§ 590. It would seem from this paragraph as if satis est and satis habeo occur with the infinitive perfect only in the silver age. This is however not true. Cf. Cic. de Inven. I. 20. 28: quia satis fuit dixisse, and a little before, si cujus rei satis erit dixisse, and elsewhere. Still it is not frequent in Cicero.

§ 599. Rem. Here it should have been remarked that the historical infinitive of the passive is exceedingly rare. Although Sallust delighted in this construction, as our author rightly observes, yet the passive with him occurs only in the following few places: Cat. XXVII (fatigari); Jug. XXX. (agitari); Ibid. LX. (ferri); Ibid. LXXXIII. (trahi).

§ 607. There are some other interesting examples of the personal construction of several verbs in the passive voice. Cic. pro Sext. LIV. 95: hic accusare eum non est situs.

parva esse

ad

Rem. We may still ask, how dicitur is to be construed when it is not translated by, he is said, but by, it is asserted, or in a similar way. Cf. Cic. de Finn. III. 18. 60: Sed quum ab his omnia proficiscantur officia, non sine caussa dicitur ad ea referri omnes nostras cogitationes, and with a proleptic demonstrative pronoun, Cic. de Finn. V. 24. 72: Atque hoc ut vere dicitur, beate vivendum momenta ista corporis commodorum, sic—; in Verr. IV. 18. 38: De hoc (Diodoro) Verri dicitur, habere eum perbona toreumata. Dicitur must always be followed by an accusative before an infinitive, if a dative is connected with it. De Orat. I. 33. 150: Vere etiam illud dicitur, perverse dicere homines perverse dicendo facillime consequi; pro Mil. V. 12: Sequitur illud, quod a Milonis amicis saepissime dicitur, caedem-senatum judicasse contra rempublicam esse factam, although the accusative before the infinitive is here to be considered as depending on sequitur. The nominative before the infinitive, after dicitur, is also to be found, e. g. in Cic. pro Sext. XVII. 39: C. Caesarinimicissimus esse meae saluti ab eodem quotidianis concionibus dicebatur.

Here two passages may be given containing compound tenses. Cic. Orat. IX. 29: qui-ab Aristophane poëta fulgere dictus esset, and Ibid. IX. 27: ii sunt existimandi Attice dicere.

§ 612. In the sentence, non vales, non audes esse uxor, the unclassical vales should be stricken out. Moreover nescire fre quently occurs thus with Cicero, as we may learn from § 610. Cf. pro Mil. XXII. 75: nescis inimici factum reprehendere. So